A Meta-Analysis of The Effect of TQM On Competitiv

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/245507180
A meta-analysis of the effect of TQM on competitive advantage
Article in International Journal of Quality & Reliability Management · May 2007

DOI: 10.1108/02656710710748349
CITATIONS READS
125 2,136
3 authors, including:
Eman Elshenawy
Suez Canal University
3 PUBLICATIONS 163 CITATIONS
SEE PROFILE
All content following this page was uploaded by Eman Elshenawy on 27 November 2014.
The user has requested enhancement of the downloaded file.

International Journal of Quality & Reliability Management
Volume 24 , Issue 5 , pp. 442-563 , 2007
Articles
A meta-analysis of the effect of TQM on competitive advantage

Eman El Shenawy, Tim Baker, David J. Lemak (pp. 442-471)
Keywords: Competitive advantage, Organizational culture, Process management, Senior
management, Total quality management, Training
An empirical study of the motives and benefits of ISO 9000 certification: the UAE
experience
Walid Zaramdini (pp. 472-491)
Keywords: Companies, ISO 9000 series, United Arab Emirates
Assessing international tourists' perceptions of service quality at Air Mauritius

Girish Prayag (pp. 492-514)
Keywords: Airlines, Customer satisfaction, Customer services quality, Mauritius,
Perception, SERVQUAL
Measuring information quality in the order fulfilment process

Helena Forslund (pp. 515-524)
Keywords: Information, Order processing, Performance measurement (quality), Quality
systems, Sweden
Modeling and analysing system failure behaviour using RCA, FMEA and NHPPP
models
Rajiv Kumar Sharma, Dinesh Kumar, Pradeep Kumar (pp. 525-546)
Keywords: Failure modes and effects analysis, Fuzzy logic, Modelling
Using a reliability capability maturity model to benchmark electronics companies

Sanjay Tiku, Michael Azarian, Michael Pecht (pp. 547-563)
Keywords: Benchmarking, Electronics industry, Product reliability
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0265-671X.htm
IJQRM
24,5 A meta-analysis of the effect of
TQM on competitive advantage
Eman El Shenawy
442 Department of Management and Operations, College of Business,
Washington State University, Vancouver, Washington, USA, and
Received July 2005 Tim Baker and David J. Lemak
Revised April 2006
Department of Management and Operations, College of Business,
Washington State University, Tri-Cities, Richland, Washington, USA
Abstract
Purpose – The purpose of this study is to integrate findings of empirical studies regarding the effect
of total quality management (TQM) on competitive advantage. This purpose is to support building a
theoretical model of TQM and its components. These components are: top management
commitment/leadership, teams, culture, training/education, and process efficiency; they are
grounded in the work of Deming and deduced from three other models offered by Dean and Bowen
and Reed et al.
Design/methodology/approach – This study employed a meta-analysis to synthesize results of a
sample of 51 studies into, effectively, one database. The meta-analysis approach is used to establish
external validity for the theoretical model of TQM used in the paper. The sample includes studies that
were conducted in different countries to provide a comprehensive investigation.
Findings – Each individual component of TQM was associated with competitive advantage, that
these associations each explain roughly 60 percent of the variability in competitive advantage, and
that a 1 point change in an average component score (1-5 Likert scale) results in at least a 0.1 point
change in competitive advantage.
Research limitations/implications – The strong correlations between the five components,
coupled with the limited sample size, made it impossible to fit a competitive advantage explanation
model that included all five components with any statistical significance. Thus, it was not possible to
determine the relative impacts of the five components on competitive advantage. Moreover, these
limitations made the impact of leadership relative to other variables indeterminate, even in two
independent variable models.
Originality/value – Despite the modest findings, this study provides a link between the theory and
practice of TQM efforts and provides direction for future research.
Keywords Total quality management, Competitive advantage, Senior management,
Process management, Training, Organizational culture
Paper type Research paper
Introduction
While the debate over total quality management (TQM) in both the practitioner and
academic audiences has subsided somewhat, decisions with regard to quality and
International Journal of Quality & quality programs are still important. TQM took the practitioner world by storm in the
Reliability Management 1980s with Deming’s “14 Points” serving as the basis for many a “how to” book and
Vol. 24 No. 5, 2007
pp. 442-471 probably just as many consultancies. However, the phenomenon was largely ignored
q Emerald Group Publishing Limited
0265-671X
by the academic world until the mid-1990s (Dean and Bowen, 1994) primarily for a lack
DOI 10.1108/02656710710748349 of theoretical grounding (Reed et al., 1996). Many of the early criticisms of TQM
included its lack of specificity as a concept (e.g. Hackman and Wageman, 1995; A meta-analysis
Zbaracki, 1998) or lack of consensus regarding its effect on competitive advantage (e.g. of the effect of
Young, 1992). Likewise, there was little systematic, empirical evidence to measure
TQM’s impact on firm performance (e.g. Hackman and Wageman, 1995; Reed et al., TQM
1996).
Much headway was made with a special issue of the Academy of Management
Review in 1994 that specifically addressed the lack of TQM theory (e.g. Dean and 443
Bowen, 1994; Sitkin et al., 1994; Spencer, 1994). Likewise, additional empirical work
on the implementation of TQM strongly suggested that formal, systematic efforts
at improving quality really did improve firm financial performance (e.g., Hendricks
and Singhal, 1997; Ittner and Larcker, 1996; Lemak et al., 1997). The one general
question still unanswered about TQM revolves around the “why” of superior
performance, and the most specific answer offered up relates to competitive
advantage.
Despite a few remaining “doubting Thomases” (e.g., Zbaracki, 1998), research on
TQM has generally proven its value to those firms who take its implementation
seriously (Hendricks and Singhal, 2001; Lemak et al., 2002). However, understanding
TQM as a source of competitive advantage still needs much more exploration. Powell
(1995) began the search with his 12 principles for TQM as a basis for achieving
competitive advantage. Reed et al. (1996) argued that TQM is a business-level strategy
that goes beyond process and should address four strategy content options: product
design efficiency, product reliability, process efficiency, and market advantage. They
went on to argue that achieving competitive advantage comes from finding a fit
between the proper strategy content and the firm’s external environment. More
recently, Reed et al. (2000) added to our understanding of TQM as a business strategy
with regard to process, identifying leadership/top management commitment,
employees’ training and education, teams, and culture as the critical processes.
Achieving competitive advantage on the process side revolves around the complexity
and tacitness associated with each. Higher complexity and higher tacitness in one or
more of these key processes means that it is much more difficult for competitors to
copy the process, thus giving the firm a competitive advantage (see Barney, 1991).
However, while the theoretical foundation for TQM is expanding, much more needs
to be done to empirically validate the effect of TQM on firm performance. One
dependent variable that is of interest in this area is competitive advantage (Sitkin et al.,
1994; Powell, 1995). This study is trying to fill this lacuna, first by providing a
theoretical model of TQM that integrates the literature with respect to competitive
advantage, and second, by using a meta-analytic procedure to test the empirical
validation of that proposed model.
We base our theoretical model on three major studies: Dean and Bowen (1994); Reed
et al. (1996, 2000). Each of these studies was grounded in a rich literature review of the
empirical and theoretical research available at the time of the study. Our model
borrows five main constructs from those studies, but then we extend that thinking by
proposing them as five components that lead to generating competitive advantage. The
five components are top management commitment, use of teams, a supportive
organizational culture, training and education, and, finally, process efficiency. To
validate this model, we conducted a meta-analysis to find any main effects in the
relationship between the five components and competitive advantage.
IJQRM This study is striving to answer the following research question: “What is the effect
of the five components of total quality management on achieving competitive
24,5 advantage?”. To answer this question we first provide a literature review to validate
the concepts of TQM and competitive advantage and to generate our hypotheses. Next,
we explain the meta-analysis procedure, including definition of variables and report
the results. Finally, we provide a discussion of those results and their limitations, and
444 offer some suggestions for future research.
TQM and competitive advantage

The main objective of TQM as described by Deming (1982) is to develop and sustain a
competitive advantage through achieving the utmost efficiency. This efficiency is
manifested in cost reduction and improvement of customer satisfaction (Deming, 1982).
Much of what he proposes is right in step with objectives of traditional management
theory and marketing theory (Lemak et al., 2002; Spencer, 1994). A previous
meta-analysis of a generic competitive advantage paradigm shows that “cost and
differentiation are high-level discriminators of any competitive strategy design”
(Campbell-Hunt, 2000, p. 127). Some studies demonstrate that TQM can generate cost-
or differentiated-based competitive advantage (e.g., Reed et al. 1996), which is
sometimes sustainable (e.g., Reed et al., 2000). Many studies demonstrate that TQM
achieves competitive advantage as evidenced by superior financial performance (e.g.,
Lemak et al., 1997), improved customer satisfaction (Mehra and Agrawal, 2003), faster
response to competitors (Spitzer, 1993), and improved product quality (Escrig-Tena,
2004). Product quality achieves high financial performance and market value (Cho and
Pucik, 2005), and eventually an enhanced competitive advantage (Flynn, Schroeder
and Sakakibara, 1995; Kroll et al., 1999). However, the empirical research on this
relationship is equivocal. Some studies demonstrate that TQM leads to a competitive
disadvantage. For example, Powell (1995) contends that TQM does not guarantee
competitive advantage in absence of critical resources. Moreover, many studies
indicate that TQM leads to business failure (Davis, 1997).
To conclude, there is no valid empirical evidence to generalize the fact that TQM
adoption leads to competitive advantage. Some researchers go far to argue that TQM
“is a ubiquitous organizational phenomenon that has given little research attention . . .
Despite thousands of articles in the business and trade press, total quality remains a
hazy, ambiguous concept” (Dean and Bowen, 1994, pp. 393-4). The literature is
controversial regarding the effect of TQM on competitive advantage, and there is still a
need to validate this effect empirically in a generalizable manner. Therefore, we argue
that:
H1. TQM adoption is positively associated with achieving competitive advantage.
Major components of TQM

Research on TQM to date does not provide a clear consensus regarding the major
components (principles) of TQM that represent the competencies TQM provides when
applied (Escrig-Tena and Bou-Llusar, 2005). Since the early trials to theorize TQM and
link it to its practical application (Saraph et al., 1989), it has been hard to capture the
wild phenomenon of TQM into one comprehensive theory. Much of the literature that
validates TQM components theoretically and empirically is based on the criteria of the
Malcolm Baldrige National Quality Award (e.g. Wilson and Collier, 2000; Davig et al.,
2003), or other suggested components (e.g. Powell, 1995). The former provides six A meta-analysis
criteria: leadership, information and analysis, strategic quality planning, human of the effect of
resource development and management, management of process quality, and customer
focus and satisfaction (e.g. Wilson and Collier, 2000). Application of TQM is TQM
ambiguous. Firms apply TQM in different approaches and adopt different components
of TQM. The application of TQM in the field is usually dependant on what quality
award or program the firm aims at achieving. For instance, a firm in the USA has the 445
option to follow the criteria of Malcolm Baldrige National Quality Award or the
Arizona State Quality Award, etc. The firm may decide to choose one or another award
based on its goal and the market it is trying to penetrate. A firm from Europe or other
parts of the world will follow the criteria of one or more ISO9000 series to get a kind of
certified quality. And so on, firms follow a kind of quality program to get certificates to
indicate that they adopted TQM. Each of these programs has different elements that
represent the different criteria for TQM available for firms’ adoption. However, as with
some previous studies, these criteria might be misleading because of the conceptual
overlap between many of them.
The ambiguous concept and practical understanding of TQM caused the
controversy regarding the effect of TQM on competitive advantage. A clear,
theoretical definition of the components of TQM is required if we are to test the
relationship between TQM and competitive advantage. Therefore, we searched the
TQM theoretical literature and decided that three studies (Dean and Bowen, 1994; Reed
et al., 1996, 2000) best captured the essence of TQM going back through more than 50
years of research.
Dean and Bowen (1994) conduct a massive literature review back to the seminal
writings of Juran, Deming, and Crosby. The authors indicate that the seminal writings
provide different concepts of TQM. The authors suggest a theory of TQM as a
managerial philosophy incorporates three major principles “teamwork,” “customer
focus,” and “continuous improvement.” The principle of teamwork suggests
facilitating collaboration of group efforts across and within functions to better the
collective decision making, feedback, and problem solving toward satisfying customer
needs (Dean and Bowen, 1994). Customer focus represents the core principle; because
customer satisfaction guaranties organizational success, all organizational efforts
should be directed and focused toward satisfying customers’ needs (Dean and Bowen,
1994). The authors suggest that the organizational customers include internal and
external customers as well. Continuous improvement is directed toward customer
satisfaction that is achieved through continuous revision and problem solving of
production and managerial processes (Dean and Bowen, 1994). These three principles
constitute a continuum because each depends and leads to the other, and they are
achieved through a connected series of techniques (Dean and Bowen, 1994). The
authors provide a comparison between TQM principles stated in the Malcolm Baldrige
National Quality Award and those stated in the literature and conclude that the Award
has a comprehensive prescription for TQM application. However, in their statement of
the three major principles of TQM, they did not include some critical principles such as
leadership commitment. They add TQM does not focus on strategy formation but
rather on implementation, which is necessary for organizational performance (Dean
and Bowen, 1994). This is consistent with our view of TQM as a business strategy that
is based on the seminal literature of TQM.
IJQRM Reed et al. (1996) argue TQM has a major focus on improving market-driven
24,5 performance and hence is a business strategy. As a strategy, TQM has process and
content. The balance between process and content is necessary for success; firms that
won the TQM award and did not have balance were not able to achieve competitive
advantage (Reed et al., 1996). TQM application achieves customer orientation that is
directed through continuous improvement and process efficiency, which are the major
446 factors to accomplish reliable high-quality products or cost efficient products (Reed
et al., 1996). Quality or cost efficiency may lead to better the firm alignment with the
environment and to achieve a sustainable competitive advantage (Reed et al., 1996).
Reed et al. (2000) present a theoretical model to describe the effect of TQM on
competitive advantage. The authors consider TQM a business strategy that
accomplishes a sustained inimitable competitive advantage for the firm. The
authors conducted a massive literature review of the theoretical and empirical research
of TQM. They conclude that TQM focus on the processes of the firm that achieve high
levels of tactiness and complexity. Focus on the process leads to customer satisfaction
and reduced cost, which are the major two sources of competitive advantage (Reed
et al., 2000). The authors suggest four major principles for TQM, leadership and top
management commitment, training and education, teams, and culture. The interaction
between these four principles creates the focus on processes and eventually
competitive advantage (Reed et al., 2000). Leadership and top management
commitment is the most effective factor in the interaction because TQM is a change
strategy that requires transformational leadership skills (Reed et al., 2000). There was
no empirical testing of the model introduced in the paper; rather, a recommendation to
test this model empirically was provided for future research (Reed et al., 2000).
To sum, the three articles can be combined to build a theoretical model for the vital
components of TQM. Those articles are based on the seminal literature of TQM
(including Deming), which suggests that there is a fair amount of consensus about five
components that seem to be the cornerstones of TQM: leadership and top management
commitment, use of teams, organizational culture, training and education, and process
efficiency. These components must be combined in TQM programs to achieve
competitive advantage. More details regarding each component are indicated below.
Leadership and top management commitment

This construct refers to the ability and vision to control internal operations and direct
them toward satisfying customer needs (Deming, 1982). Top management has a vital
role in establishing a system that provides value and satisfaction for customers with
the highest possible performance results (Ahire et al., 1996). Top management must
provide the necessary resources to accomplish quality goals and commit to continuous
improvement. Likewise, the actions of top management impact the other four
components through implementation of teams, providing training for them, as well as
individuals, communicating the values of the organization culture, and providing
resources for improvements in process efficiencies (Reed et al., 2000; Wilson and
Collier, 2000).
Culture
A culture that is quality and customer focused is the glue that connects and unifies all
components of TQM toward the final goal of achieving competitive advantage.
Understanding that “culture is the values, beliefs, and norms that guide behavior in A meta-analysis
organizations” (Reed et al., 2000, p. 15) and then using it to instill and reward behaviors of the effect of
that contribute to quality and customer satisfaction is key. Culture is also necessary for
providing employee empowerment that guides employees to focus on quality and TQM
inspect their own mistakes (Ahire et al., 1996).
Teams 447
Teamwork has essential role in TQM success. Human resource theory advocates using
teams to build on individual differences for creative problem solving and improving
organizational performance (Waldman, 1994). There is a consensus in TQM research
that unless employees are involved through cross-functional teams, TQM cannot be
accomplished and commitment to continuous improvement would fade (e.g. Dean and
Bowen, 1994; Ahire et al., 1996; Whitfield et al., 1995). There is a wide agreement in the
literature on the importance of teams and group work on organizational performance in
terms of increasing productivity, efficiency, and creativity (e.g. Whitfield et al., 1995;
Sitkin et al., 1994). Teams promote employee fulfillment and satisfaction, which in turn
leads to higher productivity and efficiency (e.g. Sitkin et al., 1994). Teamwork is essential
for identifying quality problems and correcting them (Hackman and Wageman, 1995).
Training and education

Training employees is crucial for building the “human capital” of the organization
(Dean and Bowen, 1994; Sitkin et al., 1994; Hackman and Wageman, 1995). Training
and education are essential to provide employees with new techniques and practices
necessary to implement TQM successfully using techniques such as brainstorming
(e.g. Ahire et al., 1996; Hackman and Wageman, 1995). Training and education are also
necessary for teaching the TQM philosophy that requires permanent change in
individual behaviors and attitudes (Reed et al., 2000) and leads to strengthening of
organization’s culture.
Process efficiency
This construct emphasizes continuous improvement of all organizational operations
(Reed et al., 1996). This is the core operational aspect of TQM and it is rooted in the
original control-focus of TQM (Sitkin et al., 1994). It aims at enhancing the reliability
and efficiency of control processes of the firm, in order to identify the core problems
that cause deficiencies and correct them (Sitkin et al., 1994). The final reward of process
efficiency is reducing operating cost (Reed et al., 1996). TQM advocates managing by
facts and make scientific decisions regarding all details of work to reach the level of
zero defects at the organizational level (Hackman and Wageman, 1995). To improve
performance, an efficient control system that is based on quality information must be
applied (Young, 1992). Zbaracki (1998) emphasizes that TQM cannot be fully analyzed
without understanding its technical side. Therefore, we argue that TQM strategy must
have the previous five components to succeed in achieving competitive advantage for
the firm:
H2. TQM can be defined in terms of five major components (top management
commitment, teams, culture, training, and process efficiency), and the
combined application of these components leads to achieving competitive
advantage.
IJQRM The role of top management and leadership
24,5 Empirical research consistently shows that top management has a major role in
guiding the organizational activities toward better performance (Homburg et al., 1999).
For instance, Young et al. (2001) found that top management has the upper hand in
guiding and directing the organization to adopt and implement TQM. They argue that
institutional factors have more effect on the long run; still top management is the one
448 who takes the early decision to adopt any managerial innovations or changes.
Likewise, charismatic leadership is often mentioned as a key for the adoption and
success of TQM (Reed et al., 2000; Young et al., 2001). Empirical research of TQM
consistently finds a strong association between top management commitment and
performance (e.g. Samson and Terziovski, 1999). Therefore, following the upper
echelons theory that top management direct organizational performance (Hambrick
and Mason, 1984), we propose that top management commitment and leadership
should be more strongly associated with competitive advantage than other
components:
H3. The component top management commitment/leadership is more strongly
associated with competitive advantage than the other four components.
Variables and methodology

To test our hypotheses in a manner that maximizes generalizability, we chose a sample
of empirical studies that examined privately-owned firms from a number of countries.
We chose meta-analysis to conduct the investigation since it is the only method for
integrating findings across studies that “can control chance and other artifacts and
provide a foundation for conclusions,” for the sake of theory development (Hunter and
Schmidt, 1990).
The first published meta-analysis article was by Glass and Smith (1979). They
aggregated a number of small-sample studies that examined the relationship between
class size and educational performance. In a meta-analysis, individual studies are often
not statistically significant, typically due to sample size limitations. Thus,
meta-analysis effectively concatenates these studies into one with a very large
sample size. So, if the effects being studied truly are significant, their significance can
be demonstrated. Moreover, researchers sometimes claim to “prove” their conjecture on
the basis of one statistically significant study. If they use a 95 percent significance
level, they often claim that the power of their test is 0.95 and, therefore, 95 percent of all
replications of their study should yield a significant result. This is erroneous, as the
power of their test is usually much less than 0.95. A meta-analysis overcomes the
problem of researchers generalizing from a single study by aggregating studies that do
not have to be identical in construct validity, variable ranges, and/or sample sizes,
among other study characteristics (Hunter and Schmidt, 2004).
Since Glass and Smith (1979), meta-analysis has been used successfully numerous
times (e.g., Conchar et al., 2005; Lewin and Donthu, 2005; Bertua et al., 2005). The
statistical theory which underpins what Glass and Smith (1979) and others have done
was developed by Cochran (1953). Specifically, Cochran developed the Q test for
determining if a random or fixed effects model is appropriate, and Cochran developed
much of the statistical theory for estimating a study-aggregated Pearson correlation.
We use these methods in the Analysis section. Cochran’s seminal basic work and its
subsequent adoption into study aggregation methods is an example of mathematical
theory being later used in unpredictable fashion to bolster the methodology in a variety A meta-analysis
of disciplines. of the effect of
Hunter and Schmidt’s (1990) method of meta-analysis is the most commonly used in
behavioral research (Lipsey and Wilson, 2001) because it controls for statistical TQM
artifacts, and is able to aggregate empirical results across studies and to reach “an
estimate of the true relationship between two variables in a population” (Dalton et al.,
2003, p. 16). The exact procedures for this method are described in Hunter and Schmidt 449
(1990) and Lipsey and Wilson (2001), and the relevant portions are explained in the
Analysis section below.
Procedures of meta-analysis
The first, and the most difficult, stage of any Meta-analysis is to find eligible studies
(Lipsey and Wilson, 2001). We spent several months in this stage alone just identifying
potential studies for our sample. We began by using a number of databases such as
ProQuest, PsychInfo, ElSiver, Social Science Abstracts, etc. searching by terms such as
“TQM, total quality, quality management, quality, firm performance, and competitive
advantage.” We also used the usual search methods, manual research in periodical’s
indices and contents, and traced early articles, using reference lists of the most recent
articles. Cooper (1989) called the last method the “ancestry.” We found more than 7,000
articles in the basic search. Most of these articles were case studies, practitioner articles
or other non-empirical works. We did identify some 200 studies that were empirical in
an academic sense (i.e. they included hypothesis testing using reasonable size samples
and established statistical methodologies). One limitation of meta-analysis is that it can
only be used with studies that use the same or very similar methodologies (Hunter and
Schmidt, 1990). Also, because some studies did not report descriptive statistics (means
and standard deviations) and/or correlation matrices, some excellent studies that used
sophisticated statistical methods had to be eliminated (e.g., factor analysis: Cua et al.
(2001); Forker and Hershauer (2000); path analysis: Ravichandran and Rai (2000);
ANOVA: Easton and Jarrel (1998); regression: Mohran and Tenkasi (1995); or other
methods using archival data: Hendricks and Singhal (1997); Lemak et al. (1997)). Since
survey research was the most commonly used methodology in our sample of 200 TQM
empirical studies that was the first criterion we chose to construct the sample to test
our hypotheses. However, the sample did not include studies that used methods of
collecting data other than mailing surveys such as personal interviews (e.g.
Bayo-Moriones and de Cerio, 2004; Rungtusanatham et al., 1998), and observation (e.g.
Sohal, 1998). The data set also excluded studies that reported using the same sample
used in other studies such as Sun (1999), and Solis et al. (2000).
The other criteria we used was that the study had to include at least two of the
variables of interest, the study sample had to come from one or more countries, and be
of sufficient sample size. Again, we were forced to eliminate a number of excellent
studies (e.g., Appleby and Mavin, 2000; De Toni and Nassimbeni, 2000; Eskildsen and
Dahlgaard, 2000; Glenn et al., 2002; Kanji and Wallace, 2000). Applying these strict
criteria to the remaining studies left us with a sample size of 51.
Note that we did not consider any variance in firm size, type, or industry type as our
intent was to maximize generalizability. This is also consistent with the assumption
that employing TQM should be improving performance for any firm whether small or
large, manufacturing or service or in any industry type (Powell, 1995; Walton, 1986).
IJQRM For example, the results of applying TQM should not vary with the firm size measured
24,5 by capital investments. Hendricks and Singhal (2001) found no significant empirical
evidence to relate capital-intensive to results of applying TQM. We are investigating
whether adopting TQM affects the firm’s performance at all, not to which extent it
would affect firm performance. Finally, we did not control for the period of applying
TQM. Hendricks and Singhal (2001) provide empirical evidence that performance of
450 firms did not vary between earlier and later implementation of TQM.
Data collection methods

Studies included in the meta-analysis used mail surveys as a method to collect data.
Surveys were directed to managers who are responsible for quality, CEO or quality
assurance managers in all studies. The random effect model we are applying allows
some variation in data collection method (Lipsey and Wilson, 2001).
Construct validity and variation of variables

Choosing studies that included our variables of interest was a major criterion. Hunter
and Schmidt (1990) illustrate that variation in variables names does not necessarily
reflect variation in their concepts. Therefore, we closely examined how every study
defined its variables and chose only those that matched the conceptual definitions of
variables in our model. While, Hunter and Schmidt (1990) permit some variation in
defining independent variables that does not violate construct validity we took care to
limit such variation. For example, we limited the measure of process control variables
to only process efficiency. We could have used path analysis to explicitly model
differences between studies in construct formation. This is a meta-analysis technique
for overcoming construct validity differences across studies (Hunter, 1987). This would
have had the advantage of increasing our sample size, as more studies could have been
included. However, such models sometimes are not successful and the results become
muddled. Hunter and Schmidt (1990) permit more variation in the dependent variable
given that their method includes correction for the resulted attrition artefact. Their
method does not require perfect construct validity for independent or dependent
variables providing that any empirical study deviate from such validity by committing
some type of errors. However, we limited variation in dependent variable to the lowest
possible level. More details of how we defined our independent and dependent
variables follow in the next section. Essentially, we adjust for across-study differences
in variable variation by correcting for heteroscedasticity in our regression models, and
using Fisher’s Zr method in our correlation analysis. We could have used path analysis
instead to control for range variability (Hunter and Schmidt (2004)), but a priori it is not
clear that this would have yielded more insightful results.
Measurements of independent and dependent variables

We choose the most closely related measures to our constructs. So, if a study contained
more than one related measure we chose the closest one to our construct as defined in
the study. For instance, we used the following measures for top management
commitment and leadership: “top management leadership” (Kaynak, 2003; Ugboro and
Obeng, 2000), “executive commitment” (Parzinger and Nath, 2000; Powell, 1995),
“management support” (Flynn, Sakakibara and Schroeder, 1995), “committed
leadership” (Curkovic et al., 2000), “TMT involvement” (Douglas and Judge, 2001),
and “visionary leadership” (Anderson et al., 1995). For culture, we used the following A meta-analysis
measures: “employee empowerment” (e.g. Ahire and O’Shaughnessy, 1998; Powell, of the effect of
1995; Ugboro and Obeng, 2000), “plant environment” (Flynn, Sakakibara and
Schroeder, 1995), “closer customer relations” (Curkovic et al., 2000), “employee TQM
relations” (Kaynak, 2003), and “work attitudes” (Flynn, Schroeder and Sakakibara,
1995).
For teams, we used different measures such as: “open organization” (Powell, 1995), 451
“workforce management” (Flynn, Sakakibara and Schroeder, 1995), “cross-functional
quality teams” (Curkovic et al., 2000), “customer driven” (Douglas and Judge, 2001),
“employee involvement” (Ahire and O’Shaughnessy, 1998), and “human resource”
(Choi and Eboch, 1998). For training, we used: “quality training” (Curkovic et al., 2000),
“TQM-oriented training” (Douglas and Judge, 2001), “learning” (Anderson et al., 1995),
“employee quality training” (Ahire and O’Shaughnessy, 1998), and “training” (Kaynak,
2003). For process efficiency, we used the following measures: “process improvement”
(Powell, 1995), “continuous improvement” (e.g. Curkovic et al., 2000), “design quality
management” (Ahire and O’Shaughnessy, 1998), “product design process” (Flynn,
Schroeder and Sakakibara, 1995), and “process quality management” (Ahire and
Dreyfus, 2000).
Competitive advantage
The dependent variable, competitive advantage is not a unitary construct. Rather, it
consists of an index of two major sets of constructs, those relating to improve revenues
(improving market share, or differentiation), and those relating to reduce costs
(decreasing production cost) (Campbell-Hunt, 2000; Reed et al., 1996). Measures of
competitive, advantage as introduced in studies included in our meta analysis are:
“financial performance” or improved revenues (Kaynak, 2003), “industry expert rating
of performance” (Douglas and Judge, 2001), “growth in market share” (Curkovic et al.,
2000), “product quality” (Ahire and O’Shaughnessy, 1998), and “customer satisfaction”
(e.g. Ugboro and Obeg, 2000; Anderson et al., 1995). These measures represent the most
commonly used measures of competitive advantage in TQM literature.
Coding procedures
They were done to give each study one serial number, except for studies that used two
different samples. Coding recorded the studies: author/s, date, publication type,
independent and dependent variables, sample size, data collection methods, and
location of firms. Coding preceded our calculations, it was limited because no analysis
to clusters of groups included in any study was needed, and we allowed variation in
firms’ size, type, and industry to reach generalizable results. A portion of the coding
process and the studies list are illustrated in Tables I and II.
Calculation procedures
We used the mean competitive advantage score (1-5 Likert scale) as the measure of
estimated effect size across studies. Our original model for testing the three hypotheses
was a regression with competitive advantage as the dependent variable and leadership,
culture, process efficiency, training, and teams as the independent variables. All
variables are 1-5 Likert scale responses. The unit of observation is the study, and each
variable is scored as the mean response in the study. The idea of meta-analysis in this
IJQRM
ID Country Study information Publication
24,5
1 AS Quazi and Padibjo (1998) IJQPM
2 AUS Prajogo and Brown (2004) TQMJ
3 AUS Prajogo and Sohal (2003) IJQRM
4 AUS Prajogo and Sohal (2006) EJOR
452 5 BOT Temtime (2003) TQM
6 BR Da Silva et al. (2005) TQMM
7 CAN Gatewood and Riordan (1997) JQM
8 CH Hopkins et al. (2004) IJQRM
9 CH Hua et al. (2000) TQM&BE
10 CH Li et al. (2003) IJQRM
11 CH Roa et al. (1997) JQM
12 CH Raghunathan et al. (1997) IMBDS
13 ETH Temtime and Solomon (2002) TQM
14 HK Ho et al. (1999) HFEM
15 IND Raghunathan et al. (1997) IMBDS
16 IND Roa et al. (1997) JQM
17 JP Daniel and Reitsperger (1994) JIBS
18 JP Da Silva et al. (2005) TQMM
19 JW Poksinska et al. (2002) TQMM
20 MEX Roa et al. (1997) JQM
21 SIN Brah et al. (2002) IJQRM
22 Hopkins et al. (2004) IJQRM
23 Solis et al. (1998) MQS
24 UK Yusof and Aspinwall (2000) TQM
26 US Daniel and Reitsperger (1994) JIBS
27 US Douglas and Judge (2001) AMJ
28 US Flynn et al. (1995) DS
29 US Powell (1995) SMJ
30 US Raghunathan et al. (1997) IMBDS
31 US Ugboro and Obeng (2000) JQM
32 US Wilson and Collier (2000) DS
33 X1 Forza and Filippini (1998) IJPE
34 Kannan and Tan (2005) Omega
35 X2 Flynn and Saladin (2001) JOM
Notes: Publication name abbreviation, EJOR European Journal of Operational Research, IJQRM
International Journal of Quality & Reliability Management, AMJ Academy of Management Journal, DS
Decision Sciences, TQM Total Quality Management and Business Excellence, SMJ Strategic
Management Journal, JQM Journal of Quality Management, JOM Journal of Operations Management,
DISS doctoral dissertation. TQMJ The Quality Management Journal, TQMM The TQM Magazine;
Table I. Country abbreviation is the first two or three letters of the country name. The following examples
Coding, and list of studies show some abbreviations. SIN Singapore, NOR Norway, MEX Mexico, IND India, SW Sweden, and so
included in the regression on. X1 holds for a sample that contained mix of N. American and European countries, X2 is the same
meta-analyses as X1 plus some Asian countries. AS holds for a number of Asian countries
context is to approximate the regression database and resulting estimated regression

equation that would have resulted had we concatenated each individual survey
response across the given study and then across studies.
A linear regression with no interaction terms is sufficient for our hypothesis testing.
Provided that regression model assumptions are met, H1 and H2 would be
A meta-analysis
ID Country Study information Publication
of the effect of
1 AS Quazi and Padibjo (1998) IJQRM TQM
2 CH Sun (2000) IJQRM
3 HK Antony et al. (2002) IJQRM
4 HK Ho et al. (1999) HFEM
5 HK Lee and Palmer (1999) TQM&BE 453
6 IND Motawani et al. (1994) IJQRM
7 NG Osuagwu (2002) BPMJ
8 NOR Sun (2000) IJQRM
9 UK Prabhu and Robson (2000) TQM&BE
10 UK Yusof and Aspinwall (2000) TQM
11 US Ahire and Dreyfus (2000) JOM
12 US Ahire and O’Shaughnessy (1998) IJQ
13 US Ahire et al. (1996) DS
14 US Anderson et al. (1995) DS
15 US Cameron and Barnett (1997) DISS
16 US Choi and Eboch (1998) JOM
17 US Curkovic et al. (2000) DS
18 US Daniel and Reitsperger (1994) JIBS
19 US Douglas and Judge (2001) AMJ
20 US Flynn et al. (1995) DS
21 US Parzinger and Nath (2000) TQM&BE
22 US Powell (1995) SMJ
23 US Wilson and Collier (2000) DS
24 X2 Flynn and Saladin (2001) JOM
25 Calvelage and Humble (2002) DS Proceedings
26 Fuentes-Fuentes et al. (2004) Omega
27 Prajogo and Brown (2004) TQMJ
Notes: Publication name abbreviation, EJOR European Journal of Operational Research, IJQRM
International Journal of Quality & Reliability Management, AMJ Academy of Management Journal, DS
Decision Sciences, TQM Total Quality Management and Business Excellence, SMJ Strategic
Management Journal, JQM Journal of Quality Management, JOM Journal of Operations Management,
DISS doctoral dissertation; TQMJ The Quality Management Journal; Country abbreviation is the first
two or three letters of the country name. The following examples show some abbreviations. SIN Table II.
Singapore, NOR Norway, MEX Mexico, IND India, SW Sweden, and so on. X1 holds for a sample that Coding, and list of studies
contained mix of N. American and European countries, X2 is the same as X1 plus some Asian included in the
countries. AS holds for a number of Asian countries correlation meta-analyses
demonstrated if all independent variables are statistically significant. H3 would be

demonstrated if joint confidence intervals for bleadership 2 bi for all non-leadership
variables i had lower limits greater than zero.
Correction for heteroscedasticity is always required in a regression meta analysis
(Cooper and Hedges, 1994), as the dependent variable’s standard error can vary widely
from one study to the next due to different sample sizes, and other study differences
that alter sampling errors. Our analysis used the method of moments (Cooper and
Hedges, 1994, pp. 310-11) to derive the study weights for the weighted least squares
estimation procedure that corrects heteroscedasticity. We used ordinary least squares,
as opposed to weighted least squares, to derive the two constants that determine the
weights (see Cooper and Hedges, 1994, p. 319). Either approach is equally good.
IJQRM Before doing statistical inference with the final models, we also did graphical checks
24,5 of the residuals. These revealed no bias, and the residuals were close to normal. This,
combined with the heteroscedasticity correction, ensure that our confidence intervals
are valid and as tight as the sample size and inherent variability permits. See the
Analysis section for more details.
An observation was omitted if it contained at least one missing variable value
454 across the variable set for the analysis. The full five independent variable model thus
had only 14 valid observations (39 studies used the five-point Likert scale; this was our
original database listed in Table I. Note that some of these studies had multiple
versions; thus, only 35 are listed. We dropped the 12 seven-point studies as it is
impossible to convert responses from one scale to the next; they do not scale
proportionately as the respondent may react to the two scales in differing manners).
The ridge regression in the Analysis section was performed on SAS version 9.1.3.
The heteroscedasticity correction was performed in Microsoft Visual Cþ þ using the
IMSL Stat and Math libraries for assorted statistical functions. Visual Numerics (www.
vni.com) is the vendor for IMSL. The remainder of the analysis was performed in
Microsoft Excel.
Analysis, results, and discussion

We used the “random effect meta-analysis” method, since it is a method that includes
appropriate studies while allowing for the missing or unreachable studies such as
unknown conference papers or dissertations (Hunter and Schmidt, 1990; Lipsey and
Wilson, 2001). For this method, studies are considered units of a random sample that
represent wider population of studies (Hunter and Schmidt, 1990; Lipsey and Wilson,
2001). Results of the random effect meta-analytic method show the individual effect
size per construct per number of studies used to calculate that effect size.
For our full model (all five independent variables included), all five confidence
intervals for the regression coefficients covered zero. Also, all four and three variable
models were tried, and none had all coefficients as significant. Thus, the desired
simultaneous analysis of a large number of independent variable effects could not be
performed. Ideally, we would like to fit a five variable model with all variables
significant; this would verify H2. These insignificant results are due in part to
multi-collinearity. In the full model, the independent variable correlation matrix has
entries from 0.42 to 0.90 in Table III, with most toward the upper end of this range.
Also, the small sample sizes make it difficult to ascertain significance as the number of
parameters increases.
We tried ridge regression to isolate the effects of each independent variable in
the complete five variable model. For sets of regression coefficients whose
Culture Leadership Teams Training Process
Table III. Culture 0.61 0.60 0.88 0.67

Correlations between Leadership 0.45 0.69 0.42
independent variables in Teams 0.70 0.90
five independent variable Training 0.73
model Process
corresponding data vectors are highly correlated with one another, often their A meta-analysis
standard error estimates are inflated. This causes their t-statistics to be depressed, of the effect of
making the variables look statistically insignificant when they very well may be
significant. The goal of ridge regression is to reduce these standard errors by TQM
introducing a small bias in the coefficient estimates. Thus, ridge regression
generates different coefficient estimates from the usual ordinary least squares
estimates, but hopefully these new estimates are statistically significant. The idea 455
is to select the minimum amount of bias introduced to eliminate standard error
inflation. This inflation is measured by each variable’s variance inflation factor, or
VIF. Approximately, if the VIF exceeds 10, then inflation is likely present (see
Belsey et al., 1980).
We thus varied the bias from 0 to 0.016 in increments of 0.002 and estimated the
coefficients via ridge regression along with their corresponding VIFs in Table IV. Since
all VIFs are less than 10, the ridge regression procedure will not help determine true
statistical significance within this set of independent variables. Because most of our
two variable non-ridge estimated models were completely significant but all more
inclusive models were not, collinearity is clearly present. Ridge regression is not
guaranteed to ferret out multicollinearity; it is an approximate method.
Thus, our analysis is limited to two independent variable models. These are
summarized in Table V. Note that each of the five components appears somewhere in the
significant models. This provides support for H1 – all five components do associate
with competitive advantage if one considers their association in isolation from the rest of
the TQM environment. However, because models with more independent variables are
insignificant, we cannot determine how essential each component is to the maintenance
of competitive advantage in the full TQM context – with all five components
simultaneously impacting competitive advantage. Thus, although there is support for
H2 in a loose sense, a larger number of studies is needed to fully evaluate this hypothesis.
Bias Coefficient/VIF Culture Leadership Process Teams Training
0 Coef 1.44 0.64 1.61 2 0.37 2 0.45

Vif 4.51 1.99 6.08 5.44 6.59
0.002 Coef 1.44 0.64 1.61 2 0.37 2 0.45
Vif 4.36 1.97 5.85 5.23 6.35
0.004 Coef 1.43 0.64 1.59 2 0.36 2 0.42
Vif 4.23 1.94 5.63 5.04 6.13
0.006 Coef 1.42 0.63 1.57 2 0.34 2 0.40
Vif 4.10 1.92 5.42 4.86 5.91
0.008 Coef 1.41 0.63 1.55 2 0.33 2 0.38
Vif 3.98 1.89 5.23 4.69 5.71
0.010 Coef 1.40 0.62 1.53 2 0.31 2 0.36
Vif 3.87 1.87 5.04 4.53 5.52
0.012 Coef 1.39 0.62 1.52 2 0.30 2 0.34
Vif 3.76 1.85 4.87 4.38 5.34
0.014 Coef 1.38 0.61 1.50 2 0.29 2 0.32
Vif 3.66 1.82 4.71 4.24 5.17
0.016 Coef 1.37 0.61 1.48 2 0.27 2 0.31 Table IV.
Vif 3.56 1.80 4.55 4.10 5.00 Ridge regression results
IJQRM
Confidence interval on
24,5 Independent Confidence interval Confidence interval difference between coefficients
variables on coefficient (1) on coefficient (2) (1) and (2) if (1) is leadership R2
(1) culture,
(2) process (0.09, 0.86) (0.10, 0.84) NA 0.65
456 (1) culture,
(2) training (0.12, 0.85) (0.12, 0.78) NA 0.67
(1) leadership,
(2) training (0.08, 0.83) (0.15, 0.81) (2 1.0, 0.60) 0.66
(1) leadership,
(2) culture (0.14, 0.88) (0.12, 0.84) (2 0.51, 1.47) 0.64
(1) leadership,
Table V. (2) process (0.20, 0.77) (0.25, 0.80) (2 1.37, 0.25) 0.68
Summary I of all two (1) process,
independent variable (2) training (0.10, 0.80) (0.15, 0.79) NA 0.68
regression models where (1) leadership,
all variables are (2) teams (0.07, 0.49) (0.25, 0.87) (2 1.02, 0.86) 0.65
significant (95 percent (1) teams,
confidence intervals) (2) culture (0.06, 0.63) (0.09, 0.85) NA 0.61
H3 is indeterminate since all confidence intervals involving leadership cover zero.

Again, a larger sample size might reveal the dominance of leadership over the other
components in association with competitive advantage.
A careful examination of the residual plots in Figures 1-16 reveal that the
assumptions necessary for the regression estimates to be unbiased and of minimum
variance are met: form of regression equation (linear, in this case) is appropriate, errors
are normally distributed, errors have constant variance across all combinations of
independent variables, and errors are uncorrelated over time (the last assumption is
irrelevant in our study since the data are not time ordered) (Neter et al., 1996). Thus, our
confidence intervals from the regression analysis are as tight as the sample size
permits. For the regression equation form to be correct, one would expect all
scatterplots to have residuals centered about zero. This is the case everywhere. This
Figure 1.
Culture-process regression
A meta-analysis
of the effect of
TQM
457
Figure 2.
Culture-process regression
Figure 3.
Culture-training
regression
Figure 4.
Culture-training
regression
purely visual approach for checking for is supported by Pedhazur (1997), Cohen and
Cohen (1983), and Berry and Feldman (1985). Without this assurance, both the type I
and II errors are larger than they would be with the correct equation form.
To obtain assurance that the coefficient estimators have the smallest possible
variance, one also wants to see that these plots exhibit constant variance about zero
residual within each graph across all levels of the independent variable, and the swarm
of points about the residual zero line roughly follow a normal distribution. Both of
IJQRM
24,5
458
Figure 5.
Leadership-training
regression
Figure 6.
Leadership-training
regression
Figure 7.
Culture-leadership
regression
these properties are exhibited. Berry and Feldman (1985) provide justification that our
purely graphical approach for testing the constant variance assumption is sufficient.
One could use a variance homogeneity test, such as the Goldfield-Quandt test
(Nunnally, 1978) or the Glesjer test (Berry and Feldman, 1985) when the error variances
are tighter in the central part of the independent variable set, but these tests are
unnecessary due to the robustness of the minimum variance result in the presence of
A meta-analysis
of the effect of
TQM
459
Figure 8.
Culture-leadership
regression
Figure 9.
Leadership-process
regression
Figure 10.
Leadership-process
regression
modest violations of the constant error variance assumption. If these violations are
pronounced, then the Type I error probabilities are much higher than one would expect
if this assumption was not violated (Berry and Feldman (1985)).
Because the residual plots also appear to follow a normal distribution at a given
independent variable level in the plots in Figures 1-16, we have no reason to believe
that the normality assumption is violated. Draper and Smith (1998) provide
justification for this purely graphical test of this assumption. Other researchers use
leverage indices to determine which outliers are influential, and delete these outliers if
IJQRM
24,5
460
Figure 11.
Process-training
regression
Figure 12.
Process-training
regression
Figure 13.
Leadership-teams
regression
justified (Tabachnick and Fidell, 2001). But there we could not justify the deletion of
any data points, outside of omitting all seven-point scale responses. These responses
were deleted since they cannot be reconciled with the five-point responses, and the
five-point responses were far more numerous. Some researchers use the Shapiro-Wilk
test (Shapiro and Wilk, 1965) to formally test if there are significant deviations from
normality in the residuals, but Draper and Smith (1998) believe this is unnecessary due
to the robustness of regression estimators to violations of the normality assumption.
A meta-analysis
of the effect of
TQM
461
Figure 14.
Leadership-teams
regression
Figure 15.
Culture-teams regression
Figure 16.
Culture-teams regression
Looney (1995) recommends examining the skewness and kurtosis of the residuals, but
Wilkinson and The Task Force on Statistical Inference (1999) disagree, stating that
such examinations do not catch assumption violations. Hair et al. (1995) suggest
viewing the normal probability plot in checking residual normality, but this is
equivalent to examining the residual plots for a normal shape.
Because unbiased and minimum variance estimators are evident, no further data
transformations on the models in Table V would be warranted. Table VI provides
further evidence about the indeterminant nature of our results. The difficulty in
IJQRM predicting competitive advantage is borne out in the prediction intervals in Table VI.
They are all extremely wide, ranging from a very low score of around 1 to an extremely
24,5 high score of 5.
We also utilized Fisher’s Zr transformation meta analysis method for generalizing a
correlation for each independent variable against the dependent variable across studies
(Lipsey and Wilson, 2001). This is another way to test H3 – if the joint confidence
462 intervals about each correlation are such that leadership’s lower limit lies above the
upper limit of all other variables, then we will have demonstrated H3. Our database
consists of the studies in Table II. Fisher’s transformation is the only analysis that
permits synthesis of correlations across studies; it is a meta-analysis standard (Wolf
(1986)). The Pearson correlation has a complex standard error, and Fisher’s Zr is the
only way to make the generalization across studies tractable. This transformation
achieves rough normality so that the adjustments for sampling variability differences
across studies can occur. Let r be the sample Pearson correlation coefficient. We define
the effect size ESZr to be:

1þr
0:5 ln : ð1Þ
12r
Doing analysis with this effect size permits the development of a cross-study
correlation coefficient. The key to estimating the error in generating this coefficient is
the proper generation of study weights, denoted wi, where the subscript i refers to a
particular study. We first need to determine if the studies can be shown to be truly
heterogeneous – the correlations do not measure the same population correlation. If so,
then a random effects model is best for producing the weights. Let the candidate
weight for a study be n 2 3, where nP is the sample sizePof the study. Then, Pm if the studies
are homogenous, the statistic Q ¼ m i¼1 ðwi ES Zri
Þ 2 m
i¼1 w i ES Zri
Þ2
= i¼1 wi follows a
chi-squared distribution with m 2 1 degrees of freedom, where m is the number of
studies. If the homogeneity hypothesis is rejected, then a random effects model is
warranted for determining the wi. This was the case, as demonstrated in the first row of
Table VII. So, the appropriate wi is 1=se 2 þ v^ u , where:
1
se 2 ¼
X
m
ðni 2 3Þ
i¼1
Independent variables Observations Prediction interval at means
(1) culture, (2) process 18 (1.39, 5.71)

Table VI. (1) culture, (2) training 29 (1.54, 5.73)
Summary II of all two (1) leadership, (2) training 32 (1.54, 5.71)
independent variable (1) leadership, (2) culture 31 (1.59, 5.70)
regression models where (1) leadership, (2) process 30 (1.58, 5.73)
all variables are (1) process, (2) training 30 (1.61, 5.69)
significant (95 percent (1) leadership, (2) teams 30 (1.58, 5.68)
confidence intervals) (1) teams, (2) culture 16 (1.35, 5.75)
and A meta-analysis
Q2m21
of the effect of
v^ u ¼ 0 m 1: TQM
X
2
B ðni 2 3Þ C
Xm B i¼1 C
B
ðni 2 3ÞB m C
X C 463
i¼1 @ A
ðni 2 3Þ
i¼1
With these Padjusted

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi weights, aP1 2 a=2 confidence interval for ES Zr is
^z12a=2 1= m w i , where E Zr ¼ m wi ESZr =Pm wi . One can then insert this
S
i¼1 i¼1 i i¼1
ESZr into equation (1) and solve for the point and interval estimates for the cross-study
correlation coefficient.
Table VII contains the confidence limits (at the joint 95 percent level) for all five
independent-dependent variable correlation coefficients, along with the associated
homogeneity test results. Note that a Bonferroni correction factor was applied to the
significance level so that all five confidence intervals would be correct at the desired
significance level. Since all of the confidence intervals for the Pearson correlation
coefficient r overlap almost completely, there is no evidence that any independent
variable – teams, process, culture, training, leadership – has a greater impact on
competitive advantage than any other variable. Thus, Fisher’s Zr transformation Meta
analysis method completely supports in a cross-checking manner the results obtained
using linear regression.
While the results here are modest, this study has helped understand the important
relationship between TQM and competitive advantage. Given that all five components
are usually associated with each other theoretically, the multi-collinearity between
them as independent variables is not surprising.
Conclusion
TQM’s effect on performance had been disputed for decades. Research did not clearly
specify the components of TQM and practitioners were confused which model to follow
to apply TQM effectively and avoid its high failure rates. This paper followed the
seminal literature assumption that TQM is a business strategy that leads to achieving
Culture Leadership Process Teams Training
Q 343.26 373.59 353.68 332.94 413.10

p-value 0 0 0 0 0
v^ u 0.06 0.07 0.05 0.06 0.08
ESZr 0.48 0.50 0.51 0.50 0.48
z-value 2.57 2.57 2.57 2.57 2.57
se 0.05 0.05 0.05 0.06 0.06 Table VII.
ESZr lcl (joint 95 percent) 0.33 0.34 0.38 0.35 0.32 Correlation analysis of
ESZr ucl (joint 95 percent) 0.63 0.65 0.64 0.66 0.64 each independent
r lcl (joint 95 percent) 0.32 0.33 0.36 0.33 0.31 variable against
r ucl (joint 95 percent) 0.55 0.57 0.57 0.58 0.57 competitive advantage
IJQRM competitive advantage when properly implemented. The study suggested a model for
24,5 TQM that incorporates five major components: top management commitment and
leadership, teamwork, culture, training and education, and process efficiency. Each of
these leads to achieve competitive advantage.
This study contributes to the theoretical development of TQM. It helps to generate
the path forward for a worldwide applicable theory of TQM by identifying and
464 empirically testing a collective of viable components of effective TQM programs. While
the relative individual impact of each of these components remains indeterminate;
collectively, they are associated with competitive advantage.
The study has significant implications for practice. It guides managers to what are
the most effective components they should solicit in any quality program they intend
to adopt. Cost of failure of quality programs is high and hard to recover. The recent
dynamic business environment requires the utmost caution regarding any business
strategy such as TQM. Managers can have a clear indication of the successful
component of a TQM program that together should be applied to achieve competitive
advantage.
In terms of future research efforts, perhaps each component could be validated
individually using other methodologies. For example, we hypothesized that leadership
would be most strongly associated with competitive advantage. However, while some
empirical studies support this argument, others argue that the relationship between
leadership and competitive advantage is mediated by interaction between leadership
and other TQM components (e.g. Eskildsen and Dahlgaard, 2000; Kanji and Wallace,
2000). This might be true given that top management guides all other activities of the
organization according to the upper echelons theory of Hambrick and Mason (1984).
This would be an interesting issue for future research.
Limitations
Clearly, this study was limited to those research efforts using survey research. This
severely restricted the sample size and, perhaps, kept us from making more definitive
statements as to the validity of this five-component model of TQM. More research
using other methodologies is necessary to examine the meta-analytic results of other
forms of empirical investigation. Also, while we tried to make the results here as
generalizable as possible, there may be some significant industry and country effects.
These effects also need to be examined. While this study is step in the right direction
there is still much left to do, both theoretically and empirically.
References
Ahire, S. and Dreyfus, P. (2000), “The impact of design management and process management on
quality”, Journal of Operations Management, Vol. 18 No. 5, pp. 549-75.
Ahire, S.L. and O’Shaughnessy, K.C. (1998), “The role of top management commitment in quality
management: an empirical analysis of the auto parts industry”, International Journal of
Quality Service, Vol. 3 No. 1, pp. 5-37.
Ahire, S., Golhar, D. and Waller, M. (1996), “Development and validation of TQM implementation
constructs”, Decision Sciences, Vol. 27 No. 1, pp. 23-56.
Anderson, J., Rungtusanatham, M., Schroeder, R. and Devaraj, S. (1995), “A path analytic model
of a theory of quality management method”, Decision Sciences, Vol. 26 No. 5, pp. 637-58.
Antony, J., Leung, K., Knowles, G. and Gosh, S. (2002), “Critical success factors of TQM A meta-analysis
implementation in Hong Kong industries”, International Journal of Quality & Reliability
Management, Vol. 19 No. 5, pp. 551-66. of the effect of
Appleby, A. and Mavin, S. (2000), “Innovation not imitation”, Total Quality Management and TQM
Business Excellence, Vol. 11 Nos 4/6, pp. S554-61.
Barney, J. (1991), “Firm resources and sustained competitive advantage”, Journal of
Management, Vol. 17 No. 1, pp. 99-120. 465
Bayo-Moriones, A. and de Cerio, J. (2004), “Employee involvement”, Human Factors and
Ergonomics in Manufacturing, Vol. 14 No. 2, pp. 117-34.
Belsey, D., Kuh, E. and Welsh, R. (1980), Regression Diagnostics, Wiley, New York, NY.
Berry, W. and Feldman, S. (1985), “Multiple regression in practice”, Sage University Paper Series
on Quantitative Applications in the Social Sciences, Series No. 07-050, Sage, Thousand
Oaks, CA.
Bertua, C., Anderson, N. and Salgado, J. (2005), “The predictive ability of cognitive ability tests:
a UK meta-analysis”, Journal of Occupational and Organizational Psychology, Vol. 78 No. 3,
pp. 387-409.
Brah, S.A., Tee, S.S.L. and Rao, B.M. (2002), “Relationship between TQM and performance of
Singapore companies”, International Journal of Quality & Reliability Management, Vol. 19
No. 4, pp. 356-79.
Calvelage, D. and Humble, J. (2002), “Investigation of quality management programs for small
businesses”, Decision Sciences 2002 Annual Meeting Proceedings.
Campbell-Hunt, C. (2000), “What we learned about generic competitive strategy: a meta-analysis”,
Strategic Management Journal, Vol. 21 No. 2, pp. 127-54.
Cho, H.-J. and Pucik, V. (2005), “Relationship between innovativeness, quality, growth,
profitability, and market value”, Strategic Management Journal, Vol. 26 No. 6, pp. 555-75.
Choi, T. and Eboch, K. (1998), “The TQM paradox”, Journal of Operations Management, Vol. 17
No. 1, pp. 59-75.
Cochran, W. (1953), Sampling Techniques, 1st ed., Wiley, New York, NY.
Cohen, J. and Cohen, P. (1983), Applied Multiple Regression/Correlation Analysis for the
Behavioral Sciences, Lawrence Erlbaum Associates, Hillsdale, NJ.
Conchar, M., Crask, M. and Zinkham, G. (2005), “Market valuation models of the effect of
advertising and promotional spending: a review and meta-analysis”, Academy of
Marketing Science, Vol. 33 No. 4, pp. 445-60.
Cooper, H. (1989), Integrating Research: A Guide for Literature Review, Sage Publications,
London.
Cooper, H. and Hedges, L. (1994), The Handbook of Research Synthesis, Russell Sage Foundation,
New York, NY.
Cua, K., McKone, K. and Schroeder, R. (2001), “Relationships between implementing of TQM, JIT,
and TPM and manufacturing performance”, Journal of Operations Management, Vol. 19
No. 6, pp. 675-94.
Curkovic, S., Vickery, S. and Drog, C. (2000), “Quality-related action programs”, Decision
Sciences, Vol. 31 No. 4, pp. 885-905.
Da Silva, G., Tadashi, O. and Kikuo, N. (2005), “Looking through and beyond the TQM horizon:
lessons learned from world-class companies”, The TQM Magazine, Vol. 17 No. 1, pp. 67-85.
Dalton, D., Certo, S. and Roengpitya, R. (2003), “Meta-analysis of financial performance and
equity”, Academy of Management Journal, Vol. 40 No. 1, pp. 13-26.
IJQRM Daniel, S.J. and Reitsperger, W.D. (1994), “Strategic control systems for quality: an empirical
comparison on the Japanese and US electronics industry”, Journal of International
24,5 Business Studies, Vol. 25 No. 2, pp. 275-94.
Davig, W., Brown, S., Friel, T. and Tabibzadeh, K. (2003), “Quality management in small
manufacturing”, Industrial Management & Data Systems, Vol. 2 Nos 1/2, pp. 68-77.
Davis, T. (1997), “Breakdowns in total quality management”, International Journal of
466 Management, Vol. 14 No. 1, pp. 13-23.
De Toni, A. and Nassimbeni, G. (2000), “Just-in-time purchasing”, Omega, Vol. 28 No. 6,
pp. 631-51.
Dean, J. and Bowen, D. (1994), “Management theory and total quality: improving research and
practice through theory development”, Academy of Management Review, Vol. 19 No. 3,
pp. 392-418.
Deming, W. (1982), Quality, Productivity, and Competitive Position, Center for Advanced
Engineering Studies, Massachusetts Institute of Technology, Cambridge, MA.
Douglas, T. and Judge, W. (2001), “Total quality management implications and competitive
advantage”, Academy of Management Journal, Vol. 44 No. 1, pp. 158-69.
Draper, N. and Smith, H. (1998), Applied Regression Analysis, Wiley, New York, NY.
Easton, G. and Jarrel, S. (1998), “The effect of total quality management on corporate
performance”, Journal of Business, Vol. 71 No. 2, pp. 253-307.
Escrig-Tena, A.B. (2004), “TQM as a competitive factor: a theoretical and empirical analysis”,
International Journal of Quality & Reliability Management, Vol. 21 Nos 6/7, pp. 612-37.
Escrig-Tena, A.B. and Bou-Llusar, J.C. (2005), “A model for evaluating organizational
competencies: an application in the context of a quality management initiative”, Decision
Sciences, Vol. 36 No. 2, pp. 221-57.
Eskildsen, J. and Dahlgaard, J. (2000), “A causal model for employee satisfaction”, Total Quality
Management and Business Excellence, Vol. 11 No. 8, pp. 1081-93.
Flynn, B.B. and Saladin, B. (2001), “Further evidence on the validity of the theoretical models
underlying the Baldrige criteria”, Journal of Operations Management, Vol. 19, pp. 617-52.
Flynn, B., Sakakibara, S. and Schroeder, R. (1995), “Relationship between JIT and TQM”,
Academy of Management Journal, Vol. 38 No. 5, pp. 1325-60.
Flynn, B., Schroeder, R. and Sakakibara, S. (1995), “The impact of total quality management
practices on performance and competitive advantage”, Decision Sciences, Vol. 26 No. 5,
pp. 659-91.
Forker, L. and Hershauer, J. (2000), “Some determinants of satisfaction and quality performance
in the electronic components industry”, Production and Inventory Management, Vol. 41
No. 2, pp. 14-20.
Forza, C. and Filippini, R. (1998), “TQM impact on quality conformance and customer
satisfaction”, International Journal of Production Economics, Vol. 55 No. 1, pp. 1-20.
Fuentes-Fuentes, M., Albacete-Saez, C. and Llorens-Montes, F. (2004), “The impact of
environmental characteristics on TQM principles and organizational performance”,
Omega, Vol. 32 No. 6, p. 425.
Gatewood, R.D. and Riordan, C.M. (1997), “The development and test of a model of total quality:
organizational practices, TQ principles, employee attitudes and customer satisfaction”,
Journal of Quality Management, Vol. 2 No. 1, pp. 41-65.
Glass, G. and Smith, M. (1979), “Meta-analysis of research on the relationship of class-size and
achievement”, Educational Evaluation and Policy Analysis, Vol. 1, pp. 2-16.
Glenn, A., Solis, L., Raghunathan, T. and Rao, S. (2002), “An empirical investigation of the A meta-analysis
relationships among supplier-based relationships, quality practices, and outcomes”,
Decision Sciences Institute 2002 Annual Meeting Proceedings. of the effect of
Hackman, J. and Wageman, R. (1995), “Total quality management: empirical, practical and TQM
conceptual issues”, Administrative Science Quarterly, Vol. 40 No. 2, pp. 309-42.
Hair, J., Anderson, R., Tatham, R. and Black, W. (1995), Multivariate Data Analysis, 4th ed.,
Prentice-Hall, Upper Saddle River, NJ. 467
Hambrick, D.C. and Mason, P.A. (1984), “Upper echelons: the organization as a reflection of its
top managers”, Academy of Management Review, Vol. 9 No. 2, pp. 193-206.
Hendricks, K. and Singhal, V. (1997), “Does implementing an effective TQM program actually
improve operating performance?”, Management Science, Vol. 43 No. 9, pp. 1259-74.
Hendricks, K. and Singhal, V. (2001), “Firm characteristics, total quality management, and
financial performance”, Journal of Operations Management, Vol. 19 No. 3, pp. 269-85.
Ho, D.C.K., Duffy, V.G. and Shih, H.M. (1999), “An empirical analysis of effective TQM
implementation in the Hong Kong electronics manufacturing industry”, Human Factors
and Ergonomics in Manufacturing, Vol. 9, pp. 1-25.
Homburg, C., Krohmer, D. and Workman, J.P. (1999), “Strategic consensus and performance:
the role of strategy type and market related dynamism”, Strategic Management Journal,
Vol. 20 No. 4, pp. 339-57.
Hopkins, S.A., Nie, W. and Hopkins, W.E. (2004), “A comparative study of quality management
in Taiwan’s and China’s electronics industry”, International Journal of Quality & Reliability
Management, Vol. 21 Nos 4/5, pp. 362-77.
Hua, H., Chin, K.S., Sun, H. and Xu, Y. (2000), “An empirical study on quality management
practices in Shanghai manufacturing industries”, Total Quality Management, Vol. 11 No. 8,
pp. 1111-22.
Hunter, J. (1987), “Multiple dependent variables in program evaluation”, in Mark, M. and
Shotland, R. (Eds), Multiple Methods in Program Evaluation, Jossey-Bass, San Francisco,
CA.
Hunter, J. and Schmidt, F. (1990), Methods of Meta-analysis, Correcting Error and Bias in
Research Findings, Sage Publications, London.
Hunter, J. and Schmidt, F. (2004), Methods of Meta-analysis, Correcting Error and Bias in
Research Findings, 2nd ed., Sage Publications, London.
Ittner, C.D. and Larcker, D.F. (1996), “Measuring the impact of quality initiatives on firm
financial performance”, in Fedor, D.D. and Ghost, S. (Eds), Advances in the Management of
Organizational Quality,Vol. 1, JAI Press, Greenwich, CT, pp. 1-37.
Kanji, G. and Wallace, W. (2000), “Business excellence through customer satisfaction”, Total
Quality Management and Business Excellence, Vol. 11 No. 7, pp. S979-98.
Kannan, V.R. and Tan, K.C. (2005), “Just in time, total quality management, and supply chain
management: understanding their linkages and impact on business performance”, Omega,
Vol. 33 No. 2, p. 153.
Kaynak, H. (2003), “The relationship between total quality management practices and their effect
on firm performance”, Journal of Operations Management, Vol. 21 No. 4, pp. 405-35.
Kroll, M., Wright, P. and Heiens, R. (1999), “The contribution of product quality to competitive
advantage: impacts on systematic variance and unexplained variance in returns”,
Strategic Management Journal, Vol. 20 No. 4, pp. 375-84.
Lee, K.S. and Palmer, E. (1999), “An empirical examination of ISO 9000-registered companies in
New Zealand”, Total Quality Management & Business Excellence, Vol. 10 No. 6, pp. 887-9.
IJQRM Lemak, D., Mero, N. and Reed, R. (2002), “When quality works: a premature post-mortem on
TQM”, Journal of Business and Management, Vol. 6 No. 4, pp. 391-409.
24,5
Lemak, D., Reed, R. and Satish, P. (1997), “Commitment to quality management: is there a
relationship with firms’ performance?”, Journal of Quality Management, Vol. 2 No. 1,
pp. 77-86.
Lewin, J. and Donthu, N. (2005), “The influence of purchase situation on buying center structure
468 and involvement: a select meta-analysis of organizational buying behavior”, Journal of
Business Research, Vol. 58 No. 10, pp. 1381-90.
Li, J., Anderson, A.R. and Harrison, R.T. (2003), “Total quality management principles and
practices in China”, International Journal of Quality & Reliability Management, Vol. 20
No. 9, pp. 1026-59.
Lipsey, M. and Wilson, D. (2001), Practical Meta-analysis, Sage Publications, London.
Looney, S. (1995), “How to use tests for univariate normality to assess multivariate normality”,
American Statistician, Vol. 49 No. 1, pp. 64-70.
Mehra, S. and Agrawal, S.P. (2003), “Total quality as a new global competitive strategy”,
Mohran, S. and Tenkasi, R. (1995), “Total quality management”, Employee Relations, Vol. 3 No. 3,
pp. 26-41.
Motawani, J.G., Mahmoud, E. and Rice, G. (1994), “Quality practices of Indian organizations:
an empirical analysis”, International Journal of Quality & Reliability Management, Vol. 11
No. 1, pp. 38-52.
Neter, J., Kutner, M., Nachtsheim, C. and Wasserman, W. (1996), Applied Linear Statistical
Models, Irwin, Chicago, IL.
Nunnally, J. (1978), Psychometric Theory, 2nd ed., McGraw-Hill, New York, NY.
Osuagwu, L. (2002), “TQM strategies in a developing economy”, Business Process Management
Journal, Vol. 8 No. 2, pp. 140-60.
Parzinger, M. and Nath, R. (2000), “A study of the relationships between total quality
management implementation factors and software quality”, Total Quality Management
and Business Excellence, Vol. 11 No. 3, pp. 535-70.
Pedhazur, E. (1997), Multiple Regression in Behavioral Research, 3rd ed., Harcourt-Brace,
Orlando, FL.
Poksinska, B., Dahlgaard, J.J. and Antoni, M. (2002), “The state of ISO 9000 certification: a study
of Swedish organizations”, The TQM Magazine, Vol. 14 No. 5, pp. 297-307.
Powell, T. (1995), “Total quality management as competitive advantage”, Strategic Management
Prabhu, V.B. and Robson, A. (2000), “Impact of leadership and senior management commitment
on business excellence: an empirical study in the North East of England”, Total Quality
Management, Vol. 11 Nos 4/5/6, pp. S399-S409.
Prajogo, D.I. and Brown, A. (2004), “The relationship between TQM practices and quality
performance and the role of formal TQM programs: an Australian empirical study”,
The Quality Management Journal, Vol. 11 No. 4, pp. 31-43.
Prajogo, D.I. and Sohal, A.S. (2003), “The relationship between TQM practices, quality
performance, and innovation performance”, International Journal of Quality & Reliability
Management, Vol. 20 No. 4, pp. 901-18.
Prajogo, D.I. and Sohal, A.S. (2006), “The relationship between organization strategy, total A meta-analysis
quality management (TQM), and organization performance – the mediating role of TQM”,
European Journal of Operational Research, No. 168, pp. 35-50. of the effect of
Quazi, H.A. and Padibjo, S.R. (1998), “A journey toward total quality management through ISO TQM
9000 certification – a study on small- and medium-sized enterprises in Singapore”,
International Journal of Quality & Reliability Management, Vol. 15 No. 5, pp. 489-508.
Raghunathan, T.S., Subba Rao, S. and Solis, L.E. (1997), “A comparative study of quality 469
practices: USA, China and India”, Industrial Management & Data Systems, Vol. 97 No. 5,
p. 192.
Ravichandran, T. and Rai, A. (2000), “Quality management in systems development”, MIS
Quarterly, Vol. 24 No. 3, pp. 381-415.
Reed, R., Lemak, D. and Mero, N. (2000), “Total quality management and sustainable competitive
advantage”, Journal of Quality Management, Vol. 5 Nos 7/8, pp. 5-26.
Reed, R., Lemak, D. and Montgomery, J. (1996), “Beyond process: TQM content and firm
performance”, Academy of Management Review, Vol. 21 No. 1, pp. 173-202.
Roa, S.S., Raghunathan, T.S. and Solis, L.E. (1997), “A comparative study of quality practices and
results in India, China, and Mexico”, Journal of Quality Management, Vol. 2 No. 2,
pp. 235-50.
Rungtusanatham, M., Forza, C., Filippini, R. and Anderson, J. (1998), “A replication study of
quality management under the Deming management method”, Journal of Operations
Samson, D. and Terziovski, M. (1999), “The relationship between total quality management
practices and operational performance”, Journal of Operations Management, Vol. 17 No. 4,
pp. 393-409.
Saraph, J.V., Benson, P.G. and Schroeder, R.G. (1989), “An instrument for measuring the critical
factors of quality”, Decision Sciences, Vol. 20 No. 4, pp. 810-29.
Shapiro, S. and Wilk, M. (1965), “An analysis of variance test for normality”, Biometrika, Vol. 52
No. 4, pp. 591-611.
Sitkin, S., Sutcliffle, K. and Schroeder, R. (1994), “Distinguishing control from learning in total
quality management”, Academy of Management Review, Vol. 19 No. 3, pp. 537-64.
Sohal, A. (1998), “Assessing manufacturing/quality culture and practices in Asian countries”,
Solis, L., Raghu-Nathan, T. and Rao, S. (2000), “A regional study of quality management
infrastructure practices in USA and Mexico”, International Journal of Quality & Reliability
Solis, L.E., Rao, S., Raghu-Nathan, T.S., Chen, C.-Y. and Pan, S. (1998), “Quality management
practices and quality results: a comparison of manufacturing and service sectors in
Taiwan”, Managing Service Quality, Vol. 8 No. 1, pp. 46-54.
Spencer, B. (1994), “Models of organization and total quality management”, Academy of
Management Review, Vol. 19 No. 3, pp. 446-61.
Spitzer, R.D. (1993), “TQM: The only source of sustainable competitive advantage”, Quality
Progress, Vol. 26 No. 6, pp. 59-65.
Sun, H. (1999), “Diffusion and contribution of total quality management”, Total Quality
Management and Business Excellence, Vol. 6 No. 6, pp. 901-14.
Sun, H. (2000), “A comparison of quality management practices in Shanghai and Norwegian
manufacturing companies”, International Journal of Quality & Reliability Management,
Vol. 17 No. 6, pp. 636-60.
IJQRM Tabachnick, B. and Fidell, L. (2001), Multivariate Statistics, 4th ed., Allyn & Bacon, Needham
Heights, MA.
24,5
Temtime, Z.T. (2003), “The moderating impact of business planning and firm size on total
quality management practices”, The TQM Magazine, Vol. 15 No. 1, pp. 52-60.
Temtime, Z.T. and Solomon, G.H. (2002), “Total quality management and the planning behavior
of SME’s in developing economies”, The TQM Magazine, Vol. 14 No. 3, pp. 181-91.
470 Ugboro, I.O. and Obeng, K. (2000), “Top management leadership, employee empowerment, job
satisfaction, and customer satisfaction in TQM organizations: an empirical study”, Journal
of Quality Management, No. 5, pp. 247-72.
Waldman, D. (1994), “The contributions of total quality management to a theory of work
performance”, Academy of Management Review, Vol. 19 No. 3, pp. 510-36.
Walton, M. (1986), The Deming Management Method, Pedigree, New York, NY.
Whitfield, J., Anthony, W. and Kacmar, K. (1995), “Evaluation of team-based management: a case
study”, Journal of Organizational Change Management, Vol. 8 No. 2, pp. 17-28.
Wilkinson, L. and The Task Force on Statistical Inference (1999), “Statistical methods in
psychology journals: guidelines and explanations”, American Psychologist, Vol. 54 No. 8,
pp. 594-604.
Wilson, D. and Collier, D. (2000), “An empirical investigation of the Malcolm Baldrige National
Quality Award”, Decision Sciences, Vol. 31 No. 2, pp. 361-90.
Wolf, F. (1986), Meta-analysis: Quantitative Methods for Research Synthesis, Sage Publications,
Thousand Oaks, CA.
Young, G., Charns, M. and Shortell, S. (2001), “Top management network effects on the adoption
of innovative management practices”, Strategic Management Journal, Vol. 22 No. 10,
pp. 935-51.
Young, S. (1992), “A framework for successful adoption and performance of Japanese
manufacturing practices in the United States”, Academy of Management Review, Vol. 17
No. 4, pp. 677-700.
Yusof, S.M. and Aspinwall, E.M. (2000), “Critical success factor in small and medium enterprises:
survey results”, Total Quality Management, Vol. 11 Nos 4-6, pp. 448-62.
Zbaracki, M. (1998), “The rhetoric and reality of total quality management”, Administrative
Science Quarterly, Vol. 43 No. 3, pp. 602-53.
Further reading
Becker, B. and Schram, C. (1994), “Examining explanatory models through research synthesis”,
in Cooper, C. and Hedges, L. (Eds), The Handbook of Research Synthesis, Russell Sage
Foundation, New York, NY.
Buch, K. and Rivers, D. (2001), “TQM: the role of leadership and culture”, Leadership
& Organization Development Journal, Vol. 22 Nos 7/8, pp. 365-71.
Spector, B. and Beer, M. (1994), “Beyond TQM programs”, Journal of Organizational Change
Westphal, J., Gulati, R. and Shortell, S. (1997), “Customization or conformity? An institutional and
network perspective on the content and consequences of TQM adoption”, Administrative
Science Quarterly, Vol. 42 No. 2, pp. 366-94.
About the authors A meta-analysis
Eman El Shenawy was a Doctoral Student at Washington State University at the time of writing
this paper. She has six conference papers in well-respected academic conferences, and two of the effect of
textbooks in business. Her area of expertise is strategic management and international TQM
negotiations. She had an MBA from Washington State University and a Master of Science in
Management from the Suez Canal University, Egypt. Eman El Shenawy is the corresponding
author and can be contacted at: eman@wsu.edu
Tim Baker is an Associate Professor of Management and Operations at Washington State 471
University at Tri-Cities. His research has appeared in Decision Sciences, Production and
Operations Management, and IEEE Transactions in Engineering Management. Dr Baker has 12
years of operations research analyst experience in industry, encompassing airline and hotel
revenue management, petroleum products distribution system design, weapons system
effectiveness analysis, and manpower planning modeling. Dr Baker received a PhD in operations
management from The Ohio State University in 1994. He has an MS in operations research and
systems analysis from the University of North Carolina at Chapel Hill, and a BA in mathematics
and economics from Claremont McKenna College.
David J. Lemak is Professor of Management at the Tri-Cities campus of Washington State
University. He has a PhD from Arizona State University, a MBA from Indiana University and a
BA from Ohio Wesleyan University. His areas of expertise are strategic management and
organization theory. He has published two books, more than 25 papers in peer-reviewed journals
and given nearly 50 scholarly presentations to a variety of professional organizations.
To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints
IJQRM
24,5 An empirical study of the motives
and benefits of ISO 9000
certification: the UAE experience
472
Walid Zaramdini
College of Business and Economics, UAE University, Al-Ain,
Received April 2005
Revised April 2006 United Arab Emirates
Abstract
Purpose – The aim of the paper is to analyze the different aspects of the motives and benefits of
certified ISO 9001: 2000 companies in the United Arab Emirates (UAE).
Design/methodology/approach – A total of 900 certified UAE companies were surveyed in April
2004 in UAE. The response rate was equal to 23.2 percent.
Findings – The paper finds that certified companies were more concerned by internal reasons like
improving processes or products than by external reasons like pressure from customers or imitation of
competitors. In addition, the certification process had generated more internal benefits than external
ones.
Research limitations/implications – These findings may indicate that the surveyed firms will
sustain their quality efforts as continuous improvement is one of the cornerstones of ISO 9001: 2000
standard. In April 2004, when this survey was conducted, the number of certified companies listed in
Dubai or in Abu Dhabi Stock Exchanges was still small. This limited the analysis of the certification’s
financial benefits.
Originality/value – This is one of the first papers trying to analyze the motives and benefits of ISO
9001: 2000 certified companies in the UAE. The selection of the motives and benefits were validated
through statistical analysis and the relationship between expected and perceived benefits was
discussed.
Keywords ISO 9000 series, Companies, United Arab Emirates
Introduction
The ISO 9000 series of quality standards have been available in the marketplace for
nearly two decades. They serve as guidelines for any organization willing to establish
or improve its quality management system (QMS). Usually, when a company obtains
the ISO 9001: 2000 certificate, its business partners will be more confident in its QMS
(Stevenson and Barnes, 2001). Actually, the ISO 9000 standards are adopted in 154
countries and more than 670 000 ISO 9001: 2000 certificates have been issued
worldwide (ISO, 2004).
Some researchers investigated the “pros” and “cons” of the certification process
International Journal of Quality & (Stevenson and Barnes, 2001) or how the ISO 9000 helped companies in their total
Reliability Management quality management (TQM) journey (Magd and Curry, 2003; Sun, 1999). Other authors
Vol. 24 No. 5, 2007
pp. 472-491
q Emerald Group Publishing Limited
0265-671X
This work was financially supported by the Research Affairs at the UAE University under a
DOI 10.1108/02656710710748358 contract number: 01-03-4-11/4
studied the financial performances of the certified firms (Chow-Chua et al., 2003; Motives and
Häversjö, 2000; Tsekouras et al., 2002; Wayhan et al., 2002). benefits of
van der Wiele et al. (2005) provided an interesting classification of the research on
ISO 9000. They identified six major issues, namely: ISO 9000
(1) the relationship between ISO 9000 and TQM;
(2) perceptions about the benefits of ISO 9000;
473
(3) the relationship between ISO 9000 and organizational improvement;
(4) the usefulness of ISO 9000 for different sizes and types of organizations;
(5) the long-term effects of ISO 9000; and
(6) the motivation to implement ISO 9000.
Certain researchers investigated the state of ISO 9001: 2000 certification in the Arab
world. For example, Mezher (2000) studied the transition of 30 Lebanese firms from
ISO 9000 certification to TQM. A total of 32 Saudi manufacturing firms were surveyed
by Mezher and Ramadan (1999). They analyzed the costs and benefits linked with the
certification process. The main benefits perceived by Saudi firms were associated with
the improvement of customer service and firm’s efficiency. Similar results were found
by Magd et al. (2003). Egyptian companies were studied by Magd and Curry (2003).
The surveyed firms showed more interest towards internal operational benefits like
improvement in documentation or in the efficiency of the quality system. However,
some of these studies are just descriptive and do not include an in depth statistical
analysis.
We have noticed that among the Arab countries, the United Arab Emirates (UAE)
has the highest number of certified companies (about 900 certificates according to “The
ISO survey – 2004” (ISO, 2004)). The remaining top four positions are held by Egypt
(810 certificates), Saudi Arabia (394 certificates), and Morocco (296 certificates). The
UAE economy is heavily dependent on oil and natural gas export (it has about 10
percent of world oil reserves) and recently diversified its economy into the
manufacturing and services sectors. Dubai has become a Middle East hub for tourism
and trade businesses. However, the UAE has not been the subject of important
research regarding the ISO 9000 standards. This gave impetus to the following
research questions:
(1) What are the motivational and beneficial aspects of ISO 9001: 2000 certification
in UAE?
(2) What are the relationships between these two aspects?
Thus, it would be interesting to conduct an empirical study on the motives and benefits
of ISO 9000 certified companies in UAE. The contributions of this study to the body of
knowledge related to ISO 9000 research are as follows:
.
An extensive review of the different motives and benefits that are linked with
ISO 9001: 2000 certification.
.
An empirical study of the motives and benefits of the certified UAE
organizations, where statistical analyses were performed (reliability and
validity tests, t-test of the means, principal components analysis and
correlation analysis).
IJQRM .
Certain motives can be considered as “expected benefits”. Then, the gaps
24,5 between these expected benefits and their associated “perceived” benefits were
analyzed.
The remaining part of the paper will be as follows: first, we review the previous
literature on motives and benefits of ISO 9001: 2000 certification. Next, we present the
474 research methodology followed by the survey results. After discussing the main
findings of the empirical research, we finally conclude by providing useful
recommendations for practitioners.
Motives for seeking certification

A high number of reasons motivated different organizations for obtaining the ISO
9001: 2000 certificate or its predecessors (ISO 9001/9002/9003: 1994) (see Table I).
Jones et al. (1997) studied the reasons of certified Australian companies. They
divided them into three categories: “developmental”, “non-developmental” and
“mixed”. Companies, which belonged to the first category, were motivated by the
internal benefits obtained from the certification process like the improvement of the
“company’s internal processes” or “business performances”, whereas companies
belonging to the “non-developmental” category were pushed towards certification by
the market forces (explicit demand of important customers or necessary condition to
bid for government tenders). The “mixed” category regrouped companies having both
types of reasons. In fact, the developmental reasons are synonymous with the internal
reasons, and the non-developmental reasons are synonymous with the external ones.
Out of the 272 companies that replied to their mail questionnaire, only one company in
every seven pursued the certification for developmental reasons. When examining the
reasons of certified Malaysian companies, Yahya and Goh (2001) adopted the same
approach of Jones et al. (1997). However, they found no dominant category between the
companies studied. Escanciano et al. (2001) adopted the internal-external classification
of Vloeberghs and Bellens (1996) and found that Spanish companies were more
motivated to pursue the certification for internal reasons rather than for external ones.
In this research, and after an extensive review of previous literature, we selected 19
common reasons that could motivate UAE companies in seeking ISO 9001: 2000
certification. We note that the redundant and open to interpretation type of
motivational statements were eliminated from the selection. Furthermore, the selected
motives were classified into two groups: internal group and external group. The
internal motives are synonymous with the internal forces that are pushing the
organization in seeking ISO 9000. These forces could be either associated with
operational performances (like improving the quality system of the organization) or
with managerial performances (like improving communication between the
organization’ employees), whereas the external motives refer to the motives that are
associated with the company’s external environment. We mean by external
environment all those elements outside an organization that have potential to affect
all or part of it (like customers, competitors, government and suppliers).
Benefits of certification
Different authors studied the benefits of ISO 9001: 2000 certification from the
viewpoint of quantitative financial performance (Chow-Chua et al., 2003; Häversjö,
Number Reasons Typea Sources
1 Top management decision I Brown et al. (1998), Carlsson and Carlsson (1996), Escanciano et al. (2001)
2 Improving product and/or service I Buttle (1997), Carlsson and Carlsson (1996), Ebrahimpour et al. (1997), Escanciano et al. (2001), Llopis
quality and Tari (2003), Poksinska et al. (2002), Singels et al. (2001)
3 Improving processes and procedures I Bryde and Slocock (1998), Carlsson and Carlsson (1996), Ebrahimpour et al. (1997), Escanciano et al.
(2001), Jones et al. (1997), Krasachol et al. (1998), Lee (1998)
4 Improving productivity and/or I Brown et al. (1998), Bryde and Slocock (1998) Carlsson and Carlsson (1996), Llopis and Tari (2003)
efficiency
5 Reducing incidents, rejections and I Carlsson and Carlsson (1996), Llopis and Tari (2003)
complaints
6 A step towards total quality I Brown et al. (1998), Bryde and Slocock (1998), Carlsson and Carlsson (1996), Escanciano et al. (2001)
management Krasachol et al. (1998), Llopis and Tari (2003)
7 Use it as a basis for internal costs I Buttle (1997), Carlsson and Carlsson (1996), Escanciano et al. (2001)
reduction
8 Improving communication within the I Ebrahimpour et al. (1997)
organization
9 Improving relationships between I Ebrahimpour et al. (1997)
employees and management
10 Use it as a promotional and/or E Brown et al. (1998), Escanciano et al. (2001), Jones et al. (1997), Singels et al. (2001)
marketing tool
11 Maintaining and/or increasing market E Bryde and Slocock (1998), Carlsson and Carlsson (1996), Escanciano et al. (2001), Krasachol et al.
share (1998), Singels et al. (2001)
12 Demand and/or pressure from E Bryde and Slocock (1998), Carlsson and Carlsson (1996), Ebrahimpour et al. (1997), Escanciano et al.
customers (2001), Jones et al. (1997), Krasachol et al. (1998), Lee (1998), Singels et al. (2001)
13 Competitive advantage E Brown et al. (1998), Bryde and Slocock (1998), Carlsson and Carlsson (1996), Ebrahimpour et al. (1997),
Escanciano et al. (2001), Jones et al. (1997), Krasachol et al. (1998), Lee (1998), Singels et al. (2001)
14 Requisite to compete in the sector E Brown et al. (1998), Escanciano et al. (2001), Jones et al. (1997), Lee (1998), Singels et al. (2001)
15 Our competitors are ISO 9001 certified E Ebrahimpour et al. (1997), Escanciano et al. (2001)
16 Direct entry to new market E Carlsson and Carlsson (1996), Escanciano et al. (2001), Singels et al. (2001)
17 Be role model to suppliers E Brown et al. (1998), Bryde and Slocock (1998)
18 Improving the organization’s public E Ebrahimpour et al. (1997), Escanciano et al. (2001), Jones et al. (1997)
image
19 Requested by the government E Carlsson and Carlsson (1996), Huarng et al. (1999), Singels et al. (2001)
a
Note: E for external and I for internal
certification
Motives of ISO 9000
ISO 9000
Motives and
475
Table I.
benefits of
IJQRM 2000; Tsekouras et al., 2002; Wayhan et al., 2002) or qualitative performances (Arauz
24,5 and Suziki, 2004; Brown et al., 1998; Bryde and Slocock, 1998; Buttle, 1997; Chow-Chua
et al., 2003; Escanciano et al., 2001; Huarng et al., 1999; Jones et al., 1997; Singels et al.,
2001). For example (Buttle, 1997) found that “improving efficiency”, “improving
awareness of procedural problems” and “better management control” were the top
three benefits perceived by the firms studied, followed by “using standard as a
476 promotional tool” and “increasing customer satisfaction”. (Escanciano et al., 2001)
found similar results where “better understanding of processes/responsibilities” and
“company image in the market improved” were considered as the most important
benefits.
Similarly to the motives items list, each benefit item was classified according to its
internal or external most likely effect. Table II lists all the 20 benefits items collected
from different articles.
Research methodology
The collection of data was done through a questionnaire survey. The latter was mailed
to 900 UAE certified companies nationwide, in the beginning of April 2004,
accompanied with a return postage envelope. The companies’ addresses were collected
from different sources namely: Abu Dhabi Chamber of Commerce and Industry, Gulf
Oilfield Directory 2003 published by Arabian Publications (a publisher of professional
directories in UAE) and ISO Master Guide 2004-2005 published by Spear Publishing
(a marketing consulting company in Dubai). The list included companies operating in
different fields such as oil and gas, telecommunications, construction, information
technology, manufacturing and government agencies. Apart from the general
demographic questions that were asked (like size, field of operation, ownership, etc.),
the questionnaire survey included two sets of questions. One set is related to the
motivational aspects of the ISO 9001: 2000 certification and the second is related to the
benefits aspects. These questions were adapted from previous literature on the subject
as detailed in the preceding section.
A study group composed by three quality managers and two quality consultants,
who have an extensive experience in quality management within UAE firms, helped us
review the survey. In the light of their comments, the questionnaire was refined to
better adapt with the local culture. Evidently, the internal-external classification of the
question statements was not made apparent to the respondent. The questionnaire was
directly sent to the quality managers of the companies surveyed because it was an
important recommendation from the study group. Besides, many authors in similar
studies selected the quality manager as the most suitable person to answer their
questionnaires as he/she is usually the key person in coordinating the certification
process and is supposed to have had an extensive training in quality management
(Carlsson and Carlsson, 1996; Escanciano et al., 2001; Llopis and Tari, 2003). The Likert
scale of five points was used for evaluating every question statement. The scale was
assigned values from one to five to indicate agreement or disagreement with the
statement.
Out of 900 sent questionnaires, 220 were received. Among the 220 received
questionnaires, eleven were eliminated from subsequent analysis as they had
incomplete responses. Thus, the research is based on the data provided by 209
respondents, which leads to 23.2 percent response rate.
Number Benefits Typea Sources
1 Improved product and/or service I Arauz and Suziki (2004), Brown et al. (1998), Escanciano et al. (2001), Singels et al. (2001)
quality
2 Reduction in incidents, rejections and I Arauz and Suziki (2004), and Escanciano et al. (2001)
complaints
3 Increased productivity and/or I Buttle (1997), Dick (2000), Gotzamani and Tsiotras (2002), Häversjö (2000), Jones et al. (1997), Lee
efficiency (1998), Santos and Escanciano (2002)
4 Reduced internal costs I Arauz and Suziki (2004), Buttle (1997), Escanciano et al. (2001)
5 Improved profitability I Arauz and Suziki (2004), Buttle (1997), Dick (2000), Gotzamani and Tsiotras (2002), Häversjö (2000),
Jones et al. (1997), Lee (1998), Santos and Escanciano (2002)
6 Increased workforce motivation and I Buttle (1997), Brown et al. (1998), Escanciano et al. (2001), Gotzamani and Tsiotras (2002), McLachlan
retention (1996)
7 Employees become more quality aware I Brown et al. (1998), Chow-Chua et al. (2003), Dick (2000), Quazi and Padibjo (1998), Tsiotras and
Gotzamani (1996)
8 Improved processes and procedures I Arauz and Suziki (2004), Chow-Chua et al. (2003), Poksinska et al. (2002)
9 Elimination of redundancy or I Chow-Chua et al. (2003), McLachlan (1996)
unnecessary work
10 Better working environment I Escanciano et al. (2001)
11 Better customer service I Arauz and Suziki (2004), Buttle (1997)
12 Increased customer satisfaction E Arauz and Suziki (2004), Buttle (1997), Lee (1998), McLachlan (1996), Gotzamani and Tsiotras (2002),
Quazi and Padibjo (1998)
13 Reduction in the number of customer E Buttle (1997), Escanciano et al. (2001)
audits
14 Expansion to international markets E Chow-Chua et al. (2003), Brown et al. (1998), Buttle (1997)
15 Greater competitive advantage E Dick (2000), Gotzamani and Tsiotras (2002), McLachlan (1996), Quazi and Padibjo (1998)
16 Effective promotional and/or E Buttle (1997)
marketing tool
17 Improved market share E Brown et al. (1998), Dick (2000), Jones et al. (1997), McLachlan (1996), Santos and Escanciano (2002)
18 Suppliers’ quality improved E Escanciano et al. (2001)
19 Established and/or improved mutual E Arauz and Suziki (2004), Escanciano et al. (2001)
cooperation with suppliers
20 Organization’s image in the market E Chow-Chua et al. (2003), Tsiotras and Gotzamani (1996), Vloeberghs and Bellens (1996)
improved
Note: a E for external and I for internal
certification
Benefits of ISO 9000
ISO 9000
Motives and
477
Table II.
benefits of
IJQRM Survey results
24,5 Table III presents a descriptive statistical analysis of the companies’ demographic
information. We note that most of the responding companies were from the private
sector and a great number of them have either a small size or a medium size. In
addition, the majority of the surveyed companies are operating in the service sector.
Most of the respondents have a long working experience in the quality field. We note
478 that 108 surveyed companies have acquired the ISO 9001: 1994 or ISO 9002:1994
certificate then acquired the ISO 9001: 2000 certificate, whereas 101 organizations are
newly accredited.
The data analysis uses reliability tests, validity tests, t-test of the means, factor
analysis and correlation analysis. All statistical analyses were performed by using
SPSS (Release 12.0.0).
Reliability and validity tests

The internal consistency of a group of measurement items refers to the degree to which
items in the group are homogenous. In this study, the internal consistency is estimated
by using the Cronbach’s alpha reliability coefficient. The Cronbach’s alpha for the 19
motives items is equal to 0.823 (higher than 0.6), which suggests a satisfactory
reliability (Malhotra, 2004). Moreover, the previous Cronbach’s alpha was recalculated
after dropping a specific item but its value has not increased, except for the “Requested
by the government” item where the reliability coefficient became equal to 0.831.
Description Frequency
Company size
Small (no. of employees , 100) 78
Medium . 100 no. of employees , 500) 77
Large . 500 no. of employees , 2,000) 40
Very large (no. of employees . 2,000) 14
Ownership
Private 177
Government 15
Joint 17
Activities
Services 117
Manufacturing 89
Oil and gas 3
Respondents experience
, 2 years 15
. 2 years . . . , 5 years 47
. 5 years 134
Missing 13
No. of companies that shifted from
Table III. ISO 9001:94 to ISO 9001:2000 83
Description of ISO 9002:94 to ISO 9001:2000 25
respondents Newly accredited 101
The Cronbach’s alpha for the 20 benefits items equals 0.901 (higher than 0.6). An Motives and
internal consistency analysis was performed but has not led to a higher Cronbach’s benefits of
alpha value; hence all benefits items are considered important.
The validity analysis of the questionnaire instrument refers to the extent to which it ISO 9000
measures what it was intended to measure. In this study, two different types of validity
are considered: content validity and construct validity.
The first one, also called face validity, depends on to what extent the researchers 479
developed the measurement items to cover the content domain of the variable being
measured. Usually, it is subjectively judged by the researchers. As earlier presented in
Tables I and II, the motives and benefits items were carefully selected from an
extensive review of literature. Furthermore, the validity of the lists of motives and
benefits statements was confirmed by a panel of quality managers and consultants
during the preliminary testing stage.
The construct validity refers to what extent the questionnaire instrument measures
a theoretical construct or trait it was supposed to measure. Some authors adopted the
internal-external classification prior to their statistical analyses (Jones et al., 1997;
Escanciano et al., 2001; Yahya and Goh, 2001; Vloeberghs and Bellens, 1996). Others
used it when performing a principal component analysis (Arauz and Suziki, 2004;
Brown et al., 1998; Bryde and Slocock, 1998; Buttle, 1997; Huarng et al., 1999; Llopis
and Tari, 2003; Singels et al., 2001).
So, in this research, the classification of motives and benefits items into internal and
external categories is first adopted as a theoretical trait (Tables I and II). Then, a
confirmatory factorial analysis (principal components with varimax rotation) is
conducted. We fix the number of factors to two then we check if each motive or benefit
statement belongs to its initial category. In other words, an internal motive statement
should have a high loading on the internal factor and have a low loading on the
external factor. Tables IV and V confirm the construct-validity of the internal-external
aspects of the motives and benefits items, presented in Tables II and III, respectively.
However, in Table V, the benefit statement (increased customer satisfaction) should
have loaded higher on the external factor. This result could be explained by the fact
that the respondents have considered it as an internal benefit item; nevertheless, the
remaining benefit items followed the predicted classification of Table II.
The low cumulative variance explained in both confirmatory factorial analyses
(40.61 percent and 47.26 percent) suggests an increase in the number of factors as this
will be more suitable from a statistical point of view. This issue will be treated later in
the factorial analysis section.
Analysis of the means

In this section, we are interested in ranking the motives and benefits items according to
their mean value. A t-test was performed on the respondents’ data in order to determine
the significant items. Since a five-point Likert type scale was used (1 ¼ strongly
disagree, 5 ¼ strongly agree), therefore the null hypothesis (H0) is that the average
score of motivation is less or equal to three. The alternative hypothesis (Ha) is that it is
greater than three. The confidence level is set at 95 percent, thus a ¼ 0:05. The results
of the t-test for the question statements corresponding to the motivational aspects are
described in Table VI.
IJQRM
Component
24,5 Motives items Internal External
Improving productivity and/or efficiency 0.782 20.039

Reducing incidents, rejections and complaints 0.701 0.102
Use it as a basis for internal costs reduction 0.694 0.188
480 Improving processes and procedures 0.684 20.046
Improving product and/or service quality 0.683 20.005
Improving communication within the organization 0.655 0.255
Improving relationships between employees and management 0.642 0.136
A step towards total quality management 0.469 20.044
Top management decision 0.213 0.166
Requisite to compete in the sector 0.083 0.772
Our competitors are ISO 9001 certified 0.004 0.732
Demand and/or pressure from customers 20.015 0.696
Direct entry to new market 0.048 0.669
Competitive advantage 0.158 0.652
Use it as a promotional and/or marketing tool 0.074 0.630
Improving the organization’s public image 0.132 0.491
Maintaining and/or increasing market share 0.219 0.480
Be role model to suppliers 0.395 0.448
Table IV. Requested by the UAE government 20.133 0.300
Construct-validity for the Eigenvalue 3.90 3.81
motives items Cumulative variance explained (%) 20.55 40.61
Component
Benefits items Internal External
Increased productivity and/or efficiency 0.767 0.076

Elimination of redundancy or unnecessary work 0.709 0.150
Reduced internal costs 0.692 0.192
Improved product and/or service quality 0.662 0.078
Improved processes and procedures 0.644 0.104
Better customer service 0.643 0.121
Better working environment 0.636 0.171
Increased customer satisfaction 0.615 0.219
Reduction in incidents, rejections and complaints 0.603 0.189
Improved profitability 0.601 0.370
Employees become more quality aware 0.566 0.078
Increased workforce motivation and retention 0.551 0.297
Greater competitive advantage 0.228 0.805
Expansion to international markets 0.087 0.799
Improved market share 0.142 0.709
Effective promotional and/or marketing tool 0.004 0.628
Organization’s image in the market improved 0.142 0.614
Suppliers’ quality improved 0.416 0.605
Established and/or improved mutual cooperation with suppliers 0.468 0.568
Table V. Reduction in the number of customer audits 0.167 0.522
Construct-validity for the Eigenvalue 5.48 3.96
benefits items Cumulative variance explained (%) 27.43 47.26
Motives and
Item name Type Mean Rank Std dev. t-values
benefits of
Top management decision I 4.62 1 0.593 39.54 * ISO 9000
Improving processes and procedures I 4.46 2 0.596 35.49 *
Improving product and/or service quality I 4.39 3 0.657 30.63 *
Improving productivity and/or efficiency I 4.16 4 0.778 21.52 *
Reducing incidents, rejections and complaints I 4.13 5 0.777 21.01 * 481
Improving the organization’s public image E 4.05 6 0.813 18.63 *
A step towards total quality management I 3.99 7 0.882 16.22 *
Use it as a promotional and/or marketing tool E 3.99 8 0.912 15.62 *
Competitive advantage E 3.78 9 0.995 11.33 *
Improving communication within the organization I 3.75 10 0.825 13.08 *
Maintaining and/or increasing market share E 3.65 11 0.919 10.24 *
Requisite to compete in the sector E 3.59 12 1.115 7.69 *
Improving relationships between employees and
management I 3.46 13 0.843 7.95 *
Be role model to suppliers E 3.43 14 1.008 6.17 *
Use it as a basis for internal costs reduction I 3.26 15 0.904 4.13 *
Our competitors are ISO 9001 certified E 3.17 16 1.139 2.18 *
Demand and/or pressure from customers E 3.12 17 1.111 1.619
Direct entry to new market E 2.84 18 1.066 2 2.20 *
Requested by the government E 2.17 19 1.028 2 11.641 *
Table VI.
Note: * Indicates that the corresponding p-value is lower than 0.05 Motives
The motives items are listed in decreasing order by their mean score. All statements
turned out to be significant at 95 percent confidence level, except for the “demand
and/or pressure from customers” item. The most significant ones are “top management
decision”, “improving processes and procedures”, and “improving product/service
quality” showing a t-value of 39.54, 35.49 and 30.63, respectively. As shown in Table VI,
the first five positions were occupied by internal motives. The first external motive,
“improving the organization’s public image”, held the sixth position.
However, for the “Demand and/or pressure from customers” statement, we cannot
reject the null hypothesis, in other words the respondents have not considered it as an
important reason for seeking the ISO 9001: 2000 certification. The “Direct entry to new
markets (like EU community)” presented a significant negative t-value. Hence, the
usually advanced argument that firms seek the certification for accessing the European
markets is not the case in UAE. The “Requested by the government” motive item
scored the lowest mean value. In fact, it was included in the questionnaire for testing
the consistency of the respondents’ answers (the UAE government does not impose on
public or private organizations the possession of ISO 9001: 2000 certificate).
Table VII presents the mean, standard deviation and t-values results for the 20
benefit statements. The first four important benefits are associated with internal
operational issues like processes, procedures and product/service quality. The internal
benefit statements like “improved processes and procedures”, “employees become
more quality aware” and “improved product and/or service quality” showed significant
positive t-values. The “Reduction in the number of customer’s audits” benefit
statement held the last position in Table VII.
IJQRM
Item name Type Mean Rank Std dev. t-value
24,5
Improved processes and procedures I 4.28 1 0.612 30.17 *
Employees become more quality aware I 4.20 2 0.584 29.60 *
Improved product and/or service quality I 4.16 3 0.664 25.19 *
Better customer service I 4.11 4 0.645 24.89 *
482 Organization’s image in the market improved E 4.07 5 0.714 21.71 *
Increased customer satisfaction E 4.06 6 0.712 21.48 *
Effective promotional and/or marketing tool E 4.00 7 0.806 17.85 *
Reduction in incidents, rejections and complaints I 3.99 8 0.693 20.65 *
Increased productivity and/or efficiency I 3.79 9 0.774 14.73 *
Better working environment I 3.73 10 0.717 14.75 *
Greater competitive advantage E 3.71 11 0.907 11.28 *
Elimination of redundancy or unnecessary work I 3.60 12 0.815 10.60 *
Improved market share E 3.47 13 0.809 8.38 *
Established and/or improved mutual cooperation
with suppliers E 3.45 14 0.796 8.25 *
Suppliers’ quality improved E 3.45 15 0.860 7.64 *
Improved profitability I 3.39 16 0.772 7.35 *
Increased workforce motivation and retention I 3.36 17 0.786 6.69 *
Reduced internal costs I 3.34 18 0.835 5.96 *
Expansion to international markets E 3.22 19 1.092 2.91 *
Reduction in the number of customer audits E 3.09 20 0.972 1.28
Table VII.
Benefits Note: * Indicates that the corresponding p-value is lower than 0.05
Expected and perceived benefits analysis

Certain similarities exist between Tables II and III items with respect to particular
aspects of motives and benefits under study in this empirical research. For example, in
Table II we identified “improving product and/or service quality” as a motive item and
in Table III we identified the “improved product and/or service quality” as a benefit
item. As the difference between the two statements is not only in the verbal context but
also in how the respondents perceive the effect of ISO 9000 on the product and/or
service quality, by consequence the two items can be classified as “expected” benefit
and “perceived” benefit, respectively. In Table VIII we gathered nine pairs of expected
and perceived benefits associated with ISO 9001: 2000 certification, then we analyzed
their aggregate mean differences (expected-perceived). If the mean difference, for a
certain pair of statements, had a significant negative value then the perceived benefit
value exceeded the expected benefit value. Hence, the ISO 9001: 2000 certification led to
a good quality surprise. In other words the respondents found an improvement or an
amelioration in the statement studied. On the contrary, if the mean difference had a
significant positive value then the ISO 9001: 2000 certification led to an unsatisfactory
quality result.
In Table VIII, most of the significant mean differences were positive with a
corresponding positive t-value like in pairs: 1, 2, 3, 4 and 6. Apparently, the
respondents did not feel that, after acquiring the ISO 9001: 2000 certificate, their
respective firms improved their productivity, processes, procedures and product
quality. Except for pair 6, where the associated means are nearly equal to three, the
mean values of the internally-oriented pairs (1, 2, 3 and 4) are greater or equal to four.
Hence, the respondents still obtained a good outcome from the certification process.
Motive item – benefit item Mean 1 Mean 2 Mean diff. Std. error mean t-value Sig. (two-tailed)
Pair 1 (I) (Improving – increased) productivity and/or

efficiency 4.16 3.79 0.368 0.051 7.238 0.000
Pair 2 (I) (Improving – improved) product and/or service
quality 4.39 4.16 0.234 0.046 5.110 0.000
Pair 3 (I) (Improving – improved) processes and procedures 4.46 4.28 0.187 0.043 4.357 0.000
Pair 4 (I) (Reducing – reduction in) incidents, rejections and
complaints 4.13 3.99 0.139 0.051 2.746 0.007
Pair 5 (I) Use it as a basis for internal costs reduction –
reduced internal costs 3.26 3.34 20.086 0.053 2 1.636 0.103
Pair 6 (E) Maintaining and/or increasing market share –
improved market share 3.65 3.47 0.182 0.061 2.987 0.003
Pair 7 (E) Competitive advantage 3.78 3.71 0.072 0.066 1.092 0.276
Pair 8 (E) (Use it as a (n) – effective) promotional and/or
marketing tool 3.99 4.00 20.010 0.050 2 0.190 0.849
Pair 9 (E) Improving the organization’s public image –
organization’s image in the market improved 4.05 4.07 20.024 0.056 2 0.429 0.668
(motives – benefits)
Paired samples t-test
ISO 9000
Motives and
Table VIII.
483
benefits of
IJQRM Although the externally oriented pairs (8 and 9) had a non-significant t-value and
24,5 presented a negative mean difference, they seem to show the usefulness of the
certification (their mean value are significantly greater than four). The remaining pairs
5 and 7 are disregarded from this analysis as they showed non-significant mean
differences along with relatively low respective mean values.
484 Factor analysis

The factor analysis is used to reduce the multiple relationships that may exist among
variable statements. It uncovers the common dimensions that link together the
seemingly unrelated variables, and provides insight into the underlying structure of
the data. The principal component extraction method was chosen to analyze the
correlation matrix, and to extract the eigenvalues over one. For easier interpretation of
the data set, the varimax rotation was applied. Only the factor loadings, that had
values greater than 0.5, were considered.
The factor analysis of the motives items was performed for only 16 items because
the significance of the t-test statistics along with the sign of the t-value, carried out in
the previous sub-section, showed no important effect for the last three statements (see
Table VI). We note that the value of the overall Kaise-Meyer-Olkin measure of
sampling adequacy (KMO statistic) for the motives factor analysis was equal to 0.806
(greater than 0.5) hence the usefulness of the performed factor analysis (Malhotra,
2004). The motives items were reduced to four factors. They were named as
“operational improvement”, “following competitors”, “human resources improvement”
and “supplier relationship”. The naming of the factors was done according to the
statements included in every factor. For instance, Factor 1 of Table IX was called
“operational improvement” because the question statements under this grouping
included issues such as “improving productivity/efficiency”, “improving
product/service quality” and “improving processes/procedures”. These motivational
aspects might be regarded as associated with the operational function of the
organization. The second factor was labeled “following competitors” as three out of the
four variables that had higher loading were explicitly associated with competition
issues per se. Lastly, and by applying the same reasoning process for the first two
factors, the third factor was called “human resources improvement” and the fourth was
called “supplier relationship”.
The factor analysis of the benefits items was performed for only 19 items because
the significance of the t-test statistics along with the sign of the t-value, carried out in
the previous sub-section, showed no important effect for the last statement “Reduction
in the number of customer audits” (see Table VIII). The value of the overall
Kaise-Meyer-Olkin measure of sampling adequacy (KMO statistic) for this factor
analysis is 0.875 (greater than 0.5). Five factors were selected and named as:
“profitability”, “human resources”, “product quality”, “marketing” and “suppliers” (see
Table X).
Correlation analysis
Between each motive factor and benefit factor, a bi-variate Pearson correlation was
carried out and the results were reported in Table XI. Let us use 0.4 as a cut-off point
when considering the existence of a correlation between two factors. By examining
Table XI, the following cases emerge:
Motives and
Varimax rotated loading
Operational Following Human Supplier benefits of
Item name improvement competitors resources relationship ISO 9000
Improving productivity and/or
efficiency 0.816
Improving processes and 485
procedures 0.771
Improving product and/or service
quality 0.757
Reducing incidents, rejections and
complaints 0.689
Use it as a basis for internal costs
reduction 0.600
Requisite to compete in the sector 0.807
Our competitors are ISO 9001
certified 0.793
Use it as a promotional and/or
marketing tool 0.690
Competitive advantage 0.565
Improving communication within
the organization 0.772
Improving relationships between
employees and management 0.741
Maintaining and/or increasing
market share 0.565
Be role model to suppliers 0.616
Top management decision 0.576
A step towards total quality
management 0.576
Improving the organization’s public Table IX.
image 0.551 Rotated component
Eigenvalue 2.98 2.61 2.00 1.62 matrix for the motives
Cumulative variance explained (%) 18.68 35.05 47.6 57.76 items
.
The highest correlation coefficient is equal to 0.566 and shows that the motive
factor “following competitors” is significantly correlated with the benefit factor
“marketing”. Both factors have in common certain items like “promotion
effectiveness” and “competitive advantages”.
.
Similar to the above analysis, the motive factor “operational improvement” and
the benefit factor “profitability” have a significant moderate correlation (the
correlation coefficient is equal to 0.435). The above-mentioned factors have three
similar variables in common like: “improving productivity and/or efficiency”
which has a higher loading on the “operational improvement” factor, and
“improved productivity and/or efficiency” which has a higher loading on
“profitability” benefit factor (see Tables IX and X).
The remaining values of Table XI, although significant, do not lead to a particular
interpretation.
IJQRM
Varimax rotated loading
24,5 Human Product
Item name resources quality Marketing Suppliers
Reduced internal costs 0.789

Improved profitability 0.750
486 Increased productivity and/or efficiency 0.683
Employees become more quality aware 0.748
Increased workforce motivation and retention 0.692
Better working environment 0.609
Improved processes and procedures 0.594
Elimination of redundancy or unnecessary
work 0.538
Better customer service 0.750
Increased customer satisfaction 0.696
Improved product and/or service quality 0.687
Reduction in incidents, rejections and
complaints 0.617
Effective promotional and/or marketing tool 0.832
Improved market share 0.701
Organization’s image in the market improved 0.653
Greater competitive advantage 0.610
Suppliers’ quality improved 0.823
Established and/or improved mutual
Table X. cooperation with suppliers 0.766
Rotated component Expansion to international markets 0.591
matrix for the item Eigenvalue 2.76 2.59 2.56 2.50 2.40
benefits Cumulative variance explained (%) 14.52 28.17 41.68 54.87 67.54
Benefits factors
Motives factors Profitability Human resources Product quality Marketing Suppliers
Operational improvement 0.435 * * 0.259 * * 0.347 * * 20.074 0.046

Following competitors 0.077 2 0.096 2 0.070 0.566 * * 0.141 *
Human resources 0.271 * * 0.213 * * 0.049 0.190 * * 0.212 * *
Table XI. Supplier relationship 0.006 0.172 * 0.184 0.146 * 0.241 * *
Correlations between the
motives and benefits Notes: * Correlation is significant at the 0.05 level (two-tailed); * * Correlation is significant at the 0.01
factors level (two-tailed)
Discussion
Five out of the six most important reasons that motivated UAE firms in seeking ISO
9001: 2000 certification are associated with internal issues. So, the certification process
is believed to lead to a better quality in: processes, procedures and products. This result
highly corroborates with the studies of: Buttle (1997), Carlsson and Carlsson (1995),
Llopis and Tari (2003) and Poksinska et al. (2002). The remaining reason is related to
the improvement of the organization’s public image, an outcome found as well by:
Arauz and Suziki (2004), Escanciano et al. (2001), Huarng et al. (1999) and Llopis and
Tari (2003). In fact, acquiring a quality certificate through a third party accreditation
agency can be used by the certified firm as an objective evidence of the high standard Motives and
of its QMS. Therefore, caring about their public image does not necessarily mean that benefits of
the responding firms acquired the ISO 9000 certificate for just impressing their
stakeholders. ISO 9000
The top benefits perceived by the surveyed firms illustrate again the
internal-external duality. Similar results were observed in the studies of: Brown et al.
(1998), Buttle (1997), Chow-Chua et al. 2003), Escanciano et al. (2001), and Magd et al. 487
(2003). It seems that the UAE firms are satisfied with the ISO 9001: 2000 certification as
they have realized their expected benefits.
Comparing 19 motives and 20 benefits items is difficult due to their intrinsic
differences. For example the “top management decision” motive statement has no
counterpart in the benefits list. However, the relationship between the motives and
benefits items can be studied through a principal component analysis. Four motives
factors were determined and named as: “operational improvement”, “following
competitors”, “human resources improvement” and “supplier relationship”. Also five
benefits factors were determined and named as: “profitability”, “human resources”,
“product quality”, “marketing” and “suppliers”. We note that the motives and benefits
factors, determined in this study, are similar to the ones listed in the papers of: Arauz
and Suziki (2004); Buttle (1997); Huarng et al. (1999); Llopis and Tari (2003) and Singels
et al. (2001). After conducting the factorial analysis, we performed a correlation
analysis between the motives and benefits factors. We found that the external motive
factor “following the competitors” presented a significant moderate correlation with
the external benefit factor “marketing”. A similar result was also found between the
internal motive factor “operational improvement” and the external benefit factor
“profitability”.
The relationship between the motives and benefits was further investigated by
analyzing the mean differences between the expected-perceived pairs of motives and
benefits. The analysis revealed two results:
(1) A little “disappointment” was observed by the respondents relative to the
improvement of the internal operational aspects like: productivity or product
quality.
(2) A “satisfactory” performance was noticed by the firms relative to marketing
and public image issues.
In other words, the responding firms met some of their external goals and to a lesser
degree their internal ones. Although the internal goals have been met moderately, they
still represent an important achievement because they scored above four on a five-point
Likert scale. Besides, improving the efficiency of the firm or its product quality cannot
be based solely on ISO 9001: 2000 certification. It seems that the managers have
over-emphasized the outcomes of ISO 9000. This could explain why certain
discrepancies have emerged between certain expected and perceived benefits.
Conclusions
The state of implementation of ISO 9000 in different countries is reported in many
academic articles. However, relative to the Arab world, a paucity of research does exist.
In this paper, we have studied the motives and benefits of ISO 9001: 2000 certified
firms in the UAE through a large-scale questionnaire survey. In fact, the UAE has the
IJQRM highest number of certified firms among the Arab countries and their contact
24,5 addresses are readily available from different databases. Logically, studying the
motives and benefits will not only explain why the number of certifications in UAE is
high, but also can provide the non certified firms with useful information which can
lead them to join the ISO 9000 movement. The response rate is equal to 209 out of 900
contacted organizations. Different manufacturing and service firms have answered our
488 survey and most of them present a medium or small size. More than half of the
responding firms have moved from the ISO 9001/9002/9003: 1994 to the ISO 9001: 2000
version of the standard.
Our findings suggest that the certified firms have adopted the right approach when
seeking ISO 9000 certification because they have focused on the potential internal
benefits. As a consequence of that, the certified firms have improved their business
performances. The UAE present one of the most competitive economic environment in
the region, therefore the UAE firms are keen to implement a reliable quality
management system. The ISO 9000 certification can help them in achieving their
quality targets so this can justify the high number of certified firms in UAE. The link
between the motives and benefits was analyzed by performing different statistical
techniques and revealed a significant moderate correlation between motives and
benefits factors.
This research can be continued in the following two areas: First, the outcomes of
this empirical study can represent an important reference when investigating the
long-term benefits of ISO 9001: 2000 certification in UAE. A similar approach was
adopted by Terziovskiet al. (2003, 1999), who studied the effects of ISO 9000
certification on Australian companies. Second, it will be interesting to study the impact
of ISO 9001: 2000 certification on the financial performances of the certified UAE firms.
Recommendations for practitioners

The practical implications of this study are as follows:
.
The ISO 9001: 2000 certification has delivered important internal and external
benefits to the certified UAE companies regardless of their size, field of operation
and ownership.
.
The responding firms have obtained interesting results because they have
focused on the right motives, which are mainly associated with operational
aspects.
.
The ISO 9001: 2000 certification have been used as a marketing tool and have
thus improved the firms’ public image.
.
It seems that the responding firms are not fully aware of the potential use of ISO
9001: 2000 certification as a passport for international business.
.
In order to successfully integrate the benefits of the certification along the supply
chain, the responding firms should have spent more efforts in convincing their
suppliers about its usefulness.
Finally the ISO 9001: 2000 certification is just the first step in the quality ladder.
Indeed, the UAE certified firms should continue their efforts relentlessly in order to
attain and sustain organizational excellence (Dale and Lascalles, 1997).
References Motives and
Arauz, R. and Suziki, H. (2004), “ISO 9000 performance in Japanese industries”, Total Quality benefits of
ISO 9000
Brown, A., van der Wiele, T. and Loughton, K. (1998), “Smaller enterprises’ experiences with ISO
9000”, International Journal of Quality & Reliability Management, Vol. 15 No. 3, pp. 273-85.
Bryde, D. and Slocock, B. (1998), “Quality management systems certification: a survey”,
International Journal of Quality & Reliability Management, Vol. 15 No. 5, pp. 467-80. 489
Buttle, F. (1997), “ISO 9000: marketing motivations and benefits”, International Journal of Quality
& Reliability Management, Vol. 14 No. 9, pp. 936-47.
Carlsson, M. and Carlsson, D. (1996), “Experiences of implementing ISO 9000 in Swedish
industry”, International Journal of Quality & Reliability Management, Vol. 13 No. 7,
pp. 36-47.
Chow-Chua, C., Goh, M. and Wan, T.B. (2003), “Does ISO 9000 certification improve business
performance?”, International Journal of Quality & Reliability Management, Vol. 20 No. 8,
pp. 936-53.
Dale, B.G. and Lascalles, D.M. (1997), “Total quality management adoption: revisiting the levels”,
The TQM Magazine, Vol. 9 No. 6, pp. 418-28.
Dick, G.P.M. (2000), “ISO 9000 certification benefits: reality or myths?”, The TQM Magazine,
Vol. 12 No. 6, pp. 365-71.
Ebrahimpour, M., Withers, B.E. and Hikmet, N. (1997), “Experiences of US and foreign-owned
firms: a new perspective on ISO 9000 implementation”, International Journal of Production
Research, Vol. 35 No. 2, pp. 569-76.
Escanciano, C., Fernández, E. and Vásquez, C. (2001), “Influence of ISO 9000 certification on the
progress of Spanish industry towards TQM”, International Journal of Quality & Reliability
Gotzamani, K.D. and Tsiotras, G.D. (2002), “The true motives behind ISO 9000 certification –
their effect on the overall certification benefits and long term contribution towards TQM”,
Häversjö, T. (2000), “The financial effects of ISO 9000 registration for Danish companies”,
Managerial Auditing Journal, Vol. 15 Nos 1/2, pp. 47-52.
Huarng, F., Horng, C. and Chen, C. (1999), “A study of ISO 9000 process, motivation and
performance”, Total Quality Management, Vol. 10 No. 7, pp. 1009-25.
International Standards Organization (ISO) (2004), “The ISO survey – 2004”, available at: www.
iso.org/iso/en/prods-services/otherpubs/pdf/survey2004.pdf
Jones, R., Arndt, G. and Kustin, R. (1997), “ISO 9000 among Australian companies: impact of time
and reasons for seeking certification on perceptions of benefits received”, International
Journal of Quality & Reliability Management, Vol. 14 No. 7, pp. 650-60.
Krasachol, L., Willey, P.C.T. and Tannock, J.D.T. (1998), “The progress of quality management in
Thailand”, The TQM Magazine, Vol. 10 No. 1, pp. 40-4.
Lee, T.Y. (1998), “The development of ISO 9000 certification and the future of quality
management: a survey of certified firms in Hong Kong”, International Journal of Quality
Llopis, J. and Tari, J.J. (2003), “The importance of internal aspects in quality improvement”,
McLachlan, V.N. (1996), “In praise of ISO 9000”, The TQM Magazine, Vol. 8 No. 3, pp. 21-3.
IJQRM Magd, H. and Curry, A. (2003), “ISO 9000 and TQM: are they complementary or contradictory to
each other?”, TQM Magazine, Vol. 15 No. 4, pp. 244-56.
24,5
Magd, H., Kadasah, N. and Curry, A. (2003), “ISO 9000 implementation: a study of manufacturing
companies in Saudi Arabia”, Managerial Auditing Journal, Vol. 18 No. 4, pp. 313-22.
Malhotra, N. (2004), Marketing Research, Pearson Education, Englewood Cliffs, NJ.
Mezher, T. (2000), “The transformation of Lebanese firms from ISO 9000 certified to TQM”,
490 Quality Assurance, Vol. 8 No. 1, pp. 37-56.
Mezher, T. and Ramadan, H. (1999), “The costs and benefits of getting the ISO 9000 certification
in the manufacturing sector in Saudi Arabia”, Quality Assurance, Vol. 6 No. 2, pp. 107-22.
Poksinska, B., Dahlgaard, J.J. and Antoni, M. (2002), “The state of ISO 9000 certification: a study
of Swedish organizations”, The TQM Magazine, Vol. 14 No. 5, pp. 297-306.
Quazi, H.A. and Padibjo, S.R. (1998), “A journey toward total quality management through ISO
9000 certification – a study on small- and medium-sized enterprises in Singapore”,
Santos, L. and Escanciano, C. (2002), “Benefits of the ISO 9000:1994 system: some considerations
to reinforce competitive advantage”, International Journal of Quality & Reliability
Singels, J., Ruël, G. and van de Water, H. (2001), “ISO 9000 series certification and performance”,
Stevenson, T.H. and Barnes, F.C. (2001), “Fourteen years of ISO 9000: impact, criticisms, costs,
and benefits”, Business Horizons, May-June, pp. 45-51.
Sun, H. (1999), “The patterns of implementing TQM versus ISO 9000 at the beginning of the
1990s”, International Journal of Quality & Reliability Management, Vol. 16 No. 3, pp. 201-14.
Terziovski, M., Power, D. and Sohal, A.S. (2003), “The longitudinal effects of the ISO 9000
certification process on business performance”, European Journal of Operational Research,
Vol. 146 No. 3, pp. 580-95.
Terziovski, M., Sohal, A. and Moss, S. (1999), “Longitudinal analysis of quality management
practices in Australian organizations”, Total Quality Management, Vol. 10 No. 6,
pp. 915-26.
Tsekouras, K., Dimara, E. and Skuras, D. (2002), “Adoption of a quality assurance scheme and its
effect on firm performance: a study of Greek firms implementing ISO 9000”, Total Quality
Tsiotras, G. and Gotzamani, K. (1996), “ISO 9000 as an entry key to TQM: the case of Greek
industry”, International Journal of Quality & Reliability Management, Vol. 13 No. 4,
pp. 64-76.
van der Wiele, T., van Iwaarden, J., Williams, R. and Dale, B. (2005), “Perceptions about the ISO
9000 (2000) quality system standard revision and its value: the Dutch experience”,
Vloeberghs, D. and Bellens, J. (1996), “Implementing the ISO 9000 standards in Belgium”, Quality
Progress, Vol. 29 No. 6, June, pp. 43-8.
Wayhan, V.B., Kirche, E.T. and Khumawala, B.M. (2002), “ISO 9000 certification: the financial
performance implications”, Total Quality Management, Vol. 13 No. 2, pp. 217-31.
Yahya, S. and Goh, W-K. (2001), “The implementation of an ISO 9000 quality system”,
Further reading Motives and
Casadesús, M. and Giménez, G. (2000), “The benefits of implementation of the ISO 9000 standard: benefits of
empirical research in 288 Spanish companies”, The TQM Magazine, Vol. 12 No. 6,
pp. 432-41. ISO 9000
Conti, T. (2000), “Vision 2000: positioning the new ISO 9000 with respect to total quality
management models”, Total Quality Management, Vol. 10 Nos. 4 and 5, pp. S454-64.
Withers, B. and Ebrahimpour, M. (1996), “An examination of ISO 9000 registration practices of 491
American, German and Japanese firms operating in the USA”, International Journal of
Quality & Reliability Management, Vol. 13 No. 7, pp. 8-22.
About the author

Walid Zaramdini is working in the College of Business and Economics in the United Arab
Emirates University, Al-Ain, UAE, as an Assistant-Professor of Operations Management. Prior
to that, he completed his PhD in Management Science and Engineering from the University of
Tsukuba, in Japan. He published papers in the International Journal of Production Research and
in the Journal of the Japan Industrial Management Association. Dr Zaramdini worked as quality
manager in an international automotive manufacturing company where he had the following
responsibilities: implementing the quality assurance system, dealing with the customers’
complaints and auditing the suppliers’ quality systems. He can be contacted at:
walidz@uaeu.ac.ae

IJQRM
24,5 Assessing international tourists’
perceptions of service quality at
Air Mauritius
492
Girish Prayag
Department of Management, Faculty of Law and Management,
University of Mauritius, Reduit, Mauritius
Abstract
Purpose – The aim of this paper is to assess the service quality of a small airline, Air Mauritius,
using the SERVQUAL model and to determine the relative importance attributed to service quality
dimensions by international tourists.
Design/methodology/approach – A sample of 140 international airline passengers travelling
essentially for a leisure trip was administered the SERVQUAL instrument to determine their
expectations and perceptions of service. A combination of in-flight and mail surveys was used for data
collection leading to an overall response rate of 20.2 per cent.
Findings – The paper finds that service quality structural dimensions are context- and
culture-specific. Four factors are perceived as influencing perceptions of service, service efficiency
and affect being the most important, service personalisation, reliability and tangibles the least
important. Empathy as a service dimension is valued more than assurance. Customer satisfaction and
willingness to recommend the airline is primarily dependent on service efficiency and affect.
Research limitations/implications – The sample size is relatively small and biased towards
South African passengers, but the study is exploratory in nature. The starting-point for service
improvements at Air Mauritius is better management of service reliability. The company is able to
create high perceptions of service using tangible cues.
Practical implications – The generation of a vision of excellence supported by contact employees’
empowerment should improve perceptions of service. There is potentially a link between expectations
of service levels at destination and tourists’ expectations from the flag carrier.
Originality/value – Often smaller airline companies have to resort to strategic alliances and service
differentiation to survive in this globally integrated industry. The paper fulfils the gap in the existing
literature on service quality management in small airlines.
Keywords Customer services quality, Perception, Airlines, SERVQUAL, Customer satisfaction,
Mauritius
Introduction
Mauritius has emerged as one of the best performing tourist destination in Africa over
the last ten years. The country has registered a growth of 82 per cent in tourist arrivals
over the period 1993 to 2002. Gross receipts from tourism amounted to approximately
US$655 million and 681, 648 tourists arrivals were recorded during the year 2002
(Ministry of Tourism and Leisure, Mauritius, 2002). The profile of international visitors
International Journal of Quality & has also changed with the targeting of new markets such as India, China, Russia and
Reliability Management Eastern European Countries. Traditionally, Mauritius has relied on France, Reunion
Vol. 24 No. 5, 2007
pp. 492-514 Island and UK as principal tourist generating markets. The success of the tourism
q Emerald Group Publishing Limited industry can be attributed to the strong and ongoing partnership of the local airline,
0265-671X
DOI 10.1108/02656710710748367 Air Mauritius (MK), with its partners Air France, British Airways and more recently
Emirates (MK (Air Mauritius), 2004). Being a small island economy, the government Perceptions of
has strongly supported the growth of Air Mauritius to enable the tourism industry to service quality
flourish. The flag carrier has reinforced the upmarket image of the destination over the
years.
MK has its early beginning in the 1930s and, like any other national carrier, the
growth of the airline was dependent on government support and funding. Today,
the airline flies 25 destinations around the world with a work force of around 2,600 493
staff worldwide serving the various on and off-line offices. The number of
passengers flying Air Mauritius worldwide was approximately 850, 000 for the
financial year 2002/2003 and an increase of 1.7 per cent was noted for the financial
year 2003/2004 (MK (Air Mauritius), 2004). The “limited open sky policy” and “no
charter flights” of the government has enabled Air Mauritius to survive and grow
successfully over the years. More recently, the pressure from the private sector to
increase tourist arrivals in order to sustain growth of the tourism industry, has led
the government to consider gradually implementing an “open sky policy”. In this
respect, the local airline has to improve its service levels to remain competitive
vis-à-vis the looming competition.
The company is in the process of reviewing its service strategy to increase customer
satisfaction. The airline has a strong commitment to service excellence as it acts as an
ambassador for the country and reveals the image of its people for tourism purposes.
However, intensifying competition between large and small airlines on the global scene
implies not just the opportunity for airline companies to grow and succeed but also to
fail and go out of business. The principal trends driving this sector are privatisation,
globalisation, liberalisation and cooperation (Wirtz and Johnston, 2003; Chan, 2000a;
Driver, 1999; Chidambaram, 1999) that have in essence redefined the role of smaller
airlines. Thus the potential threat for smaller airline companies to be taken over by
mega-carriers is omnipresent. The pace and extent to which small airlines have to be
innovative in their service approach are unquestionable. As such, the long-term
survival of companies in this globally integrated industry is dependent on the quality
of service they provide.
While mega carriers are using strategic alliances as an effective way to compete
internationally and provide higher levels of service to customers, smaller airlines
cannot always resort to such measures for lack of funding. The consensus in the
industry seems to be that both competition and collaboration will soon be inevitable in
all parts of the world to enhance service (Chan, 2000a; Hanlon, 2003). Smaller airlines
have to resort to strategic alliances with mega carriers to survive. Their role has been
the provision of efficient feeder or secondary services for powerful consortia or
alliances that in a way may dictate their future. Air Mauritius has been no different as
it entered into a code share agreement with Air France in 1998 and Austrian Airlines in
2004. The company has also increased its number of flights to destinations such as
Australia, South Africa, UK, India and Madagascar (MK (Air Mauritius), 2004) in order
to sustain growth despite difficult economic conditions in this global industry.
Success for such smaller airlines has been through service differentiation and airline
positioning. Service quality has become a centrepiece of their corporate and marketing
strategy. Already many mega carriers have pushed the boundaries of service quality to
its limit through service personalisation, quality meals, greater variety of in-flight
entertainment, modern aircraft, and frequent flyer programmes amongst others (Zaid,
IJQRM 1994; Sultan and Simpson, 2000; Wirtz and Johnston, 2003). Singapore Airlines, for
example, has created a strong image and reputation for its in-flight service excellence
24,5 through substantial investment in staff training and development (Wirtz and Johnston,
2003). Yet, other carriers such as Southwest Airlines and Easy Jet offer limited in-flight
service but have been equally successful in the market place. In light of such different
service management strategies employed in the airline industry, small and
494 medium-sized carriers have two options for their survival and future growth, either
to pursue a strategy of service excellence or lower cost better known as a “no-frills”
strategy (Johnson and Scholes, 1999). Air Mauritius has decided to restructure around a
service excellence strategy.
The airline has consistently received numerous awards over the years for its service
excellence, for example, it was awarded “African Airline of the Year” in 2002 at the
African Aviation Award, “Gold Award for Excellence and Business Prestige” at the
Quality Summit in New York in July 2003, “National Excellence Award” for tourism in
March 2004 and the “Travel Oscar 2003” by Verkehrsbüro, the largest Austrian travel
agency chain. The latter rates carriers according to customer satisfaction,
in-flight-service, on-time performance, value for money, complaint management and
agency support. Thus, such awards provide an indication that the airline offers
comparatively good service on its European routes and its commitment to enhancing
service quality is evident. However, perceptions of high service levels vary from
customer to customer and from country to country (Sultan and Simpson, 2000). The
influence of culture on expectations and perceptions of service quality cannot be
underestimated. A few authors (Calvert, 2000; Herbig and Genestre, 1996; Ling et al.,
2005; Winsted, 1997; Sultan and Simpson, 2000) have found that national culture
exercises a major influence on service quality, satisfaction levels and repurchase
intentions.
The purpose of this exploratory study is to determine whether customers’
expectations are being met or exceeded. It provides an insight of international tourists’
perceptions of the service level provided by Air Mauritius. It serves as a diagnostic tool
for the airline company. The SERVQUAL model developed by Parasuraman et al.
(1985) is used to measure service quality and assess the level of importance that
passengers attach to the various SERVQUAL dimensions. Therefore, the objectives of
this study are threefold:
(1) To measure customers’ expectations and perceptions of service quality based
on the SERVQUAL model.
(2) To determine customers’ assessment of the relative importance of the five
service quality dimensions of SERVQUAL.
(3) To determine any specific dimensions of service quality which has the greatest
influence on tourists’ overall satisfaction levels and willingness to recommend
the airline to others.
In addressing these objectives, a concise synthesis of the extant literature on the

subject is discussed below.
Service quality
As a concept, service quality has received much attention in the literature because of its
sustainability as a source of competitive advantage. Service quality has been defined in
different ways by researchers. Kasper et al. (1999) define service quality as “the extent Perceptions of
to which the service, the service process and the service organisation can satisfy the service quality
expectations of the user”. Parasuraman et al. (1988) define service quality as “a function
of the difference between service expected and customers” perceptions of the actual
service delivered’. Grönroos (1978) suggests that service quality is made of two
components – technical quality and functional quality. Technical quality refers to
what the service provider delivers during the service provision while functional quality 495
is how the service employee provides the service. In the services marketing literature,
the quality construct can be summarised as providing customer value (Feigenbaum,
1951), conformance to requirements (Crosby, 1979), fitness for use (Juran et al., 1974)
and meeting customers’ expectations (Parasuraman et al., 1985). Service quality is
therefore an enduring construct that encompasses quality performance in all activities
undertaken by management and employees. Customers are the sole judges of service
quality. If they perceive it to be bad service, then it is. They assess service quality by
comparing what they want or expect with what they perceive they are getting.
Few airlines have been able over the years to establish a reputation of high service
quality. This is because of rapid changes in the industry both in terms of changing
needs of customers and definitions of what constitutes the industry itself (Rhoades
et al., 1998). Singapore Airlines (SIA), British Airways (BA) and American Airlines
(AA) are among the few airlines that have successfully positioned themselves globally
as offering excellent service quality (Chan, 2000b). Delivering consistent service
quality is difficult for both large and small airline companies. Mega carriers and small
airlines are working together rather than competing with one another to maintain and
enhance quality standards. Forms of cooperation include sub contracting, code
sharing, franchising and the formation of global marketing networks. Such alliances
allow firms to focus on their respective core competencies, while drawing the benefits
of scale economies (Dana and Vignali, 1999). Firms enter alliances for competitive
reasons, for example Air Mauritius and Air Seychelles have a code share agreement on
the UK route. Such an agreement between the two partnering airlines helps to increase
flight availability and to increase yield from passengers. However, such an alliance is
dependent on both airlines offering similar service levels and having similar market
positioning for its success. Image of the two cooperating airlines has to be consistent to
avoid negative perceptions of service levels. As rightly pointed by Wirtz and Johnston
(2003), customers adjust their expectations according to brand image of the airline
company.
Service quality contributes significantly towards service differentiation, positioning
and branding. SIA and BA have long been widely acknowledged within the airline
industry as the industry’s strategic benchmark airlines, as well as the industry leaders
and innovators of service branding as a source of strategic competitive advantage
(Chan, 2000a). Companies that search for the most effective ways to incorporate the
best service methods and processes tend to be winners in the long term in terms of
favourable customer perceptions. Such companies excel in relation to their competitors
and are able to build a solid foundation for customer loyalty based on segmented
service. Service, both poor and outstanding, has a strong emotional impact on the
customer, creating intense feelings about the organisation, its staff and its service, and
influencing the loyalty to it (Wirtz and Johnston, 2003). Several authors have shown
empirically that there is a positive link between customer service improvements and
IJQRM customer satisfaction, customer loyalty and profitability (Buzzell and Gale, 1987;
24,5 Boulding et al., 1993; Rust and Oliver, 1994). At Air Mauritius, top management
recognises that it has to review its service branding strategy and in-flight service to
better meet customer needs.
Services are more subject to social, cultural and national boundaries influence,
which predetermine customers’ evaluation of service quality (Philip and Hazlett, 1997).
496 Few studies have focused on the relationship between a passenger’s cultural
background and perceptions of service quality (Ling et al., 2005). Sultan and Simpson
(2000) indicated that customer expectations and perceptions varied by nationality in an
international environment. Service quality ratings of European passengers were
significantly lower than those of US passengers. Cunningham et al. (2002), Furrer et al.
(2000), Herbig and Genestre (1996) found that there were some significant relationships
between culture and perception of service quality. Cross cultural comparison between
US and Mexican consumers revealed that Mexicans had poorer perceptions of service
quality compared to their US counterparts on the evaluation of products and services
in general. Service quality has been shown to lead to different behavioural intentions
with respect to customers from different cultures (Liu et al., 2001). Therefore, the
cultural background of passengers cannot be ignored in assessing service quality as it
contributes to building long-term brand recognition (Ling et al., 2005).
Methods of measuring service quality

While there may be general agreement that the evaluation of services is more
subjective than that of tangible products, there has been less agreement about how to
operationalise service quality as a construct (Gabbott and Hogg, 1997; Cronin and
Taylor, 1992, 1994; Dabholkar et al., 2000). Firms throughout the world use a
combination of methods to assess trends about changing customer needs and
perceptions of service quality. Zaid (1995) suggests that regular service ratings by
passengers through in-flight surveys can be used to monitor service quality in the
airline industry. Otherwise survey audits, market studies, complaint and compliment
monitoring are other available methods to measure service levels. Airlines tend to use a
combination of methods simultaneously to get a rounded picture of their service
management strategies.
The Gaps model proposed by Parasuraman et al. (1985) has been the most
comprehensive and widely used model to understand service quality. Its
operationalisation through SERVQUAL using a battery of 22 statements have been
proven to be reliable and valid across many service industries. The SERVQUAL scale
has been applied to airlines (Nel et al., 1997; Sultan and Simpson, 2000), hotels (Ingram
and Daskalakis, 1999; Juwaheer, 2004), travel agencies (Luk, 1997; Johns et al., 2004),
financial services (Kagis and Passa, 1997; Lassar et al., 2000; Newman, 2001), health
care (Desombre and Eccles, 1998; Kilbourne et al., 2004) and the public sector (Donnelly
et al., 1995; Wisniewski, 2001; Brysland and Curry, 2001). At the heart of the
SERVQUAL model is an understanding of the nature and determinants of customer
expectations and perceptions of service quality. Consumers’ expectations and
perceptions are measured to identify any shortfall in service levels, better known as the
disconfirmation paradigm in the services marketing literature. A customer will
perceive quality in a positive way only when the service provider meets or exceeds
his/her expectations (Parasuraman et al., 1985, 1988; Bitner, 1990; Robledo, 2001). In the
airline industry, customers’ expectations are shaped at the “moment-of-truth” by Perceptions of
reservations department of the airline, telephone sales, ticketing, cabin crew, cabin service quality
services, baggage handling, flight delays and others (Albrecht, 1992).
SERVQUAL uses a concise 22-item scale to measure expectations and perceptions.
The model suggests the existence of five dimensions namely: tangibles, reliability,
responsiveness, assurance and empathy that can discriminate well across customers
with differing quality perceptions. The last two dimensions contain items representing 497
seven of the original dimensions namely: communication, credibility, security,
competence, courtesy, understanding/knowing customers, and access (Parasuraman
et al., 1985). Various researchers such as Carman (1990), Cronin and Taylor (1992),
Babakus and Boller (1992), Boulding et al. (1993), Teas (1993, 1994), Buttle (1996),
Asubonteng et al. (1996), Llosa et al. (1998), Sureshchandar et al. (2001) and Coulthard
(2004) have criticised the model. Carman (1990, p. 44) suggests that “it is better to
collect data in terms of the perception/expectation difference directly rather than to
asks about each separately. It is also important to take into account the level of
experience of the customer with the service.” Cronin and Taylor (1992) through the
SERVPERF model argue that service quality should be measured as an attitude and
support the use of perception statements only in the measurement of service quality.
Numerous studies have been undertaken to assess the superiority of the two scales but
consensus continues to elude as to which one is better (Jain and Gupta, 2004). One of
the main criticisms of SERVPERF has been the way it measures customer satisfaction.
Parasuraman et al. (1988) argues that quality is an enduring global attitude towards a
service while SERVPERF measures satisfaction related to a specific transaction.
The various criticisms of SERVQUAL centre on its theoretical paradigm and the
operationalisation of the survey instrument. Buttle (1996), for example, argue that there
is little evidence that customers assess service quality in terms of the gap between
perceptions and expectations of a service and the model focuses on the service delivery
process as opposed to the service outcome. On the operational side, criticisms are, for
example, the administration of the instrument twice, which can lead to respondent
boredom and confusion. Also, Babakus and Mangold (1992) suggest the use of a five
point likert scale as opposed to the seven-point Likert scale proposed by Parasuraman
et al. (1985). In response to the growing literature on the weaknesses of SERVQUAL,
Parasuraman et al. (1991, 1994a, b) amended the original SERVQUAL model to include,
for example, levels of importance for each dimension by allocating a number of points
to each dimension, and embodiment of not only the discordance between perceived
service and desired service (labelled as measurer of service superiority) and the
discrepancy between perceived service and adequate service (labelled as measure of
service adequacy).
Furthermore, criticisms were also levelled at the applicability of the SERVQUAL
scale to service industries in developing countries (Jain and Gupta, 2004) and
cross-cultural influences on expectations and perceptions scores (Llosa et al., 1998;
Zhao et al., 2002; Nel et al., 1997). It has been demonstrated empirically that based on
cultural norms, UK and US customers have the same reaction to good service but they
tend to respond differently to poor service encounters (Voss et al., 2004). The
applicability of SERVQUAL as an instrument for measuring service quality across
cultures can therefore be questioned. Mattila (1999) found that customers from Western
cultural backgrounds are more likely to rely on tangible cues from the physical
IJQRM environment to evaluate service quality compared to customers from Asia. Furrer et al.
24,5 (2000) conclude that customers from different cultures assigned different importance
weights to the five SERVQUAL dimensions, which in turn is reflected in their
perceptions of service quality. Yet surveys conducted in New Zealand and China
revealed that customers in these apparently diverse countries display many
similarities in their expectations of service quality (Calvert, 2000).
498 Taking all these concerns into account, Philip and Hazlett (1997) proposed the P-C-P
model that takes the form of a hierarchical structure based on three main classes of
attributes namely: pivotal, core and peripheral. The pivotal attributes are considered
collectively to be the single most determining influence on why consumers approach a
particular organisation in the first instance and exert the greatest influence on
satisfaction levels. Core attributes are an amalgamation of people, processes and the
organisational structure through which the consumers must interact and/or negotiate
to achieve/receive the pivotal attributes. Finally, peripheral attributes are those
attributes that can be classified as service extras to delight the customer. This model
provides only a skeletal framework and still needs to be tested empirically for its
reliability and validity as a superior measure of service quality across industries and
cultures. Thus considering the various models proposed in the service quality
literature, SERVQUAL remains the most widely accepted and used method of
measuring service quality despite its weaknesses (Coulthard, 2004; Zhao et al., 2002;
Wisniewski, 2001; Cook and Thompson, 2000; Lam and Woo, 1997; Jain and Gupta,
2004).
Research methodology
Survey instrument
This study applies the disconfirmation theory methodology of the SERVQUAL model.
The survey instrument was designed and customised for Air Mauritius using the
battery of expectations and perceptions statements proposed by Parasuraman et al.
(1985, 1988). The survey instrument consisted of three sections. The first section
comprised of airline service quality expectations and perceptions statements
measuring the five dimensions of service quality. Only 22 items of the expectation
and perception paradigm were included in the survey instrument as they have proven
to be reasonably good predictors of service quality in its entirety (Sureshchandar et al.,
2001). The SERVQUAL model measures five dimensions: tangibility, reliability,
responsiveness, assurance and empathy. Research has shown that consumers evaluate
providers of service along dimensions, which are groupings of criteria (Robledo, 2001).
The second section of the questionnaire dealt with specific service attributes that
customers’ associated with Air Mauritius. Included were also questions pertaining to
overall level of customer satisfaction and willingness to recommend the airline to
others. The final section dealt with flying habits and preferences as well as
demographic data. The SERVQUAL statements were all worded positively as
suggested by Babakus and Boller (1992) since item wording significantly influences
the result of factor analysis on the data set. A five-point Likert scale was used to reduce
the frustration level of respondents and increase response rate and quality as
suggested by the following researchers (Babakus and Mangold, 1992; Krosnick and
Fabrigar, 1997; Preston and Colman, 2000; Johns et al., 2004). The scales were
hypothesised to have the five dimensional structure used in the SERVQUAL model
(Robledo, 2001). A pilot study was carried out to validate the survey instrument, which Perceptions of
involved mailing 50 customers that had travelled Air Mauritius over the last 12 months, service quality
only ten questionnaires were returned. Mail survey is problematic in South Africa as
observed by Nel et al. (1997) in their assessment of service quality in the airline
industry, which led to a response rate of only 33.75 per cent.
Sample design and method of data collection 499

Ideally to obtain an exact measure of the service quality offered by Air Mauritius, one
would have to administer the survey instrument on every flight to every destination
that the national airline flies to. Such a large-scale distribution of questionnaires was
unrealistic given budget constraints for this study. Consequently, it was decided to
administer questionnaires only to passengers travelling on the route South Africa to
Mauritius, given the exploratory nature of this research. Air Mauritius flies to three
cities in South Africa namely Cape Town, Durban and Johannesburg. The target
population for this study was defined as all passengers having flown Air Mauritius in
the last 12 months. A combination of self-administered in-flight survey and mail
survey was used to collect the data.
Inability to have access to a database of customers from Air Mauritius led to a
non-probability sampling procedure being adopted. A convenience sample was drawn
for the in-flight survey. The survey instrument was administered on two flights from
Cape Town to Mauritius. All seats on both flights were sold leading to 362
questionnaires (181 seats on each flight) being distributed by cabin crew. Only 95
questionnaires were returned, a response rate of 26.2 per cent. The mail survey was
administered to customers that flew Air Mauritius from Durban and Johannesburg in
the last 12 months. A list of 450 sampling elements was obtained from travel agents
and tour operators that dealt with Air Mauritius customers in South Africa. All
customers were sent a questionnaire but only 69 (response rate of 15.3 per cent) of them
returned the questionnaire to their corresponding travel agent or tour operator.
Therefore the study had an overall response rate of 20.2 per cent, which is similar to
previous studies on service quality measurement in the airline industry, which
achieved response rates of between 20 per cent and 40 per cent using similar data
collection methods (Sultan and Simpson, 2000; Frost and Kumar, 2001; Robledo, 2001;
Nel et al., 1997). All together data collection spanned over six weeks and of all
questionnaires received, 24 were unusable due to errors and incomplete responses. A
total of 140 questionnaires were used for data analysis.
Both the self-completion survey and mail survey led to a low response rate. It is well
documented that self-completion airline surveys do not generate high response rates
(Sultan and Simpson, 2000; Nel et al., 1997). This is because the researcher has no
control on the administration of the survey instrument in-flight. For this study, cabin
crew was asked to distribute the questionnaire to all passengers, except to those of
Mauritian nationality, and to collect them at the end of the flight. There is always a
probability of self-selection bias, that is, only those passengers that felt the service was
either excellent or poor to fill in and return the questionnaire to cabin crew. As noted by
Goh and Uncles (2003), both self-completion surveys and mail surveys suffer from
limitations such as prior conditioning, question order bias and problems in recalling
previous behaviour. The mail survey as a method of data collection has well known
limitations such as absence of mailing lists, poor mail services, and high levels of
IJQRM illiteracy (Aaker et al., 1998). The mailing list for this study was limited in terms of
number of respondents, accuracy of recorded addresses, and time frame to return the
24,5 questionnaire to travel agents and tour operators was short. South Africa is also
notorious for its poor mail services (Nel et al., 1997). To increase response rate, a
personalised cover letter was sent along with the survey instrument to respondents for
the mail survey.
500
Data analysis
The discussion of the research findings begins with a brief demographic profile of
respondents in terms of age, monthly household income level and nationality. On
average respondents were 37 years old, with a monthly household income in excess of
US$1,267. Most of the respondents (73 per cent) were of South African nationality while
British respondents accounted for 15 per cent of the sample and the remainder were
respondents from Greece, Germany, Spain, India and Portugal. Regarding the purpose
of visit to Mauritius, about 82 per cent were essentially on an all leisure trip. It was also
noted that 46.4 per cent of tourists had flown Air Mauritius for the first time and 55 per
cent had bought an all-inclusive package covering airfares and accommodation. Of all
respondents, 62.8 per cent flew economy class, 22.2 per cent business class and 15 per
cent first class.
Owing to the multidimensionality of the service quality construct, it was deemed
necessary to assess the reliability of the SERVQUAL instrument administered to Air
Mauritius customers. Reliability coefficients (Cronbach alpha), which is based on the
internal consistency of the items pertaining to the different service dimensions, were
computed separately for the five dimensions on the raw SERVQUAL scores. As shown
in Table I, all the dimensions were reasonably reliable, with greater than 0.7 alpha
values. The reliability coefficients are more or less similar to previous studies in the
airline industry (Nel et al., 1997; Frost and Kumar, 2001).
In addressing the first objective of the study, which is to assess customers’
expectations and perceptions of service, the former was on average highest for
Assurance, with a mean value of 4.629 followed by Reliability with a mean value of
4.587 as shown in Table II. The average score for the service dimension Assurance
imply that customers expect employees of airlines in general to be trustworthy and
polite, they should get the necessary support to perform their job well and that
customers should feel safe when transacting with the company. On average
expectations of customers were lowest for tangible features (the appearance of physical
facilities, equipment, personnel and communication materials).
International tourists had the highest perceptions of service quality for the
dimension Tangibles and lowest for Reliability as shown in Table II. The results
Dimensions Item coefficient (alphas) Reliability item alpha Standardised
Tangibles 4 0.7323 0.7471

Reliability 5 0.8841 0.8833
Responsiveness 4 0.8436 0.8453
Table I. Assurance 4 0.7169 0.7286
Service quality scores: Empathy 5 0.7131 0.7723
Cronbach’s alpha Total scale reliability 22 0.9258 0.9327
suggest the existence of service gaps at Air Mauritius. While expectations of Air Perceptions of
Mauritius customers are lowest for “Tangibles” in comparison to other service service quality
dimensions, customers tend to agree (mean ¼ 4:125) that the company has up to date
equipment, visually appealing physical facilities, well-dressed employees and visually
appealing communication materials. It is also worth noting that customers expect the
company to be reliable (mean ¼ 4:587) but the same dimension has obtained the lowest
perception scores. This is a real cause of concern and provides a clear starting point for 501
service improvements at Air Mauritius. The results of the SERVQUAL gap scores
confirm the existence of Gap 5 as suggested by Parasuraman et al. (1985, 1988).
The overall perceived service quality of Air Mauritius will be higher, the more
positive is the perception minus expectation score. The more negative the SERVQUAL
score, the more serious is the service quality shortfall in the eyes of the customer. The
mean SERVQUAL score on each service quality dimension was computed. The results
are shown in Table II. Air Mauritius does not exceed customer expectations on any of
the five service dimensions. The most problematic service dimension is reliability with
an overall mean score of 2 0.729 and the least is tangibles (mean ¼ 20:315). A
negative score for reliability shows that customers on average perceive that Air
Mauritius does not: keep its promises, show sincere interest in solving customer
problems, perform the service right the first time, provide the service at the time it
promises to do so and insist on error free records. Earlier SERVQUAL research (Nel
et al., 1997; Asubonteng et al., 1996; Frost and Kumar, 2001; Zhao et al., 2002; Ingram
and Daskalakis, 1999) indicates that businesses find it easier to meet customer
expectations in the dimension found least important by customers generally
(tangibles), while finding it most difficult to meet expectations in the dimension most
valued by customers (reliability). These assessments would suggest that airlines intent
on improving customer service quality should spend more management attention and
resources on improving the reliability dimension and less on tangibles (Sultan and
Simpson, 2000).
The overall SERVQUAL mean score for Air Mauritius is 2 0.531 and it shows the
extent of the discrepancy between customer expectations and perceived service (Gap
5). The P-E framework suggests that the highest service quality score for an attribute
occurs when the expectation score is þ 1 (strongly disagree) and the perception score is
þ 5 (strongly agree), giving a service quality score of þ 4. Also when service quality
perceptions are less than expectations, customers are dissatisfied, this suggests that on
average customers are dissatisfied by the service quality of the airline. The company
Unweighted
gap scores Weighted
Dimension Expectations Perceptions (P-E) Weightings average
Tangibles 4.440 4.125 2 0.315 14.7 20.046

Reliability 4.587 3.858 2 0.729 28.1 20.205
Responsiveness 4.543 4.072 2 0.471 18.9 20.089
Assurance 4.629 4.098 2 0.531 17.5 20.093
Empathy 4.470 3.859 2 0.611 20.8 20.127 Table II.
Overall average SERVQUAL scores for
SERVQUAL score 2 0.531 20.560 Air Mauritius
IJQRM must strive harder to align customers’ expectations with perceptions on all service
24,5 dimensions. However, assessing the importance of the service dimensions to customers
is crucial, as any investment improving service quality will payoff in the long term
only if customers perceive such investment to be of value to them. Consequently, the
relative importance of each dimension were computed using Parasuraman et al. (1991)
suggestion of allocating 100 points among the five service dimensions. Customers
502 allocated the highest weighting to Reliability and the lowest to Assurance as shown in
Table II. The overall service gap using the weighted SERVQUAL model is 2 0.56 and
reveals a larger discrepancy between tourists’ expectations and perceptions. These
findings confirm previous studies (Llosa et al., 1998; Sureshchandar et al., 2001;
Coulthard, 2004) that show weighted dimensions contribute marginally in enhancing
understanding of service quality. The negative SERVQUAL scores both weighted and
unweighted indicates that there is room for improvement in the service quality at Air
Mauritius.
The second objective of the study was further investigated using the suggestions of
Sultan and Simpson (2000) where passengers were asked to rank the importance of the
service dimensions (rank 1 indicating the most important attribute and rank 5 the least
important one). The results showed that 42.7 per cent of passengers indicated
reliability to be the most important service quality feature, 30.7 per cent ranked
empathy as the second most important, 29.9 per cent indicated responsiveness as the
third most important, 30 per cent Assurance as the fourth most important and 30.7 per
cent indicated tangibles as the least important. it is evident that the order of importance
(reliability, empathy, responsiveness, assurance and tangibility) is not the same as
noted by Parasuraman et al. (1988). Sultan and Simpson (2000) found reliability to be
the most important dimension for US and European passengers, followed by
responsiveness, assurance, empathy and tangibles.
International tourists flying Air Mauritius value reliability and empathy more than
any other service dimensions. This can be attributed to different expectations on behalf
of passengers when travelling specifically for holiday purposes. The higher
importance attached to empathy is perhaps the result of tourists also taking into
consideration expected service levels at their holiday destination in their evaluation of
a flag carrier. This link is emphasised by Juwaheer (2004), in her study of international
tourists’ perceptions of service quality provided by hotels in Mauritius. Personalisation
of service (empathy) emerged as a significant influencing factor on tourists’
perceptions of service quality. It is interesting to note that tourists perceive assurance
to be of lesser importance than empathy. This is most probably due to repeat
customers (53.6 per cent) feeling that employee trust and confidence is a given in a
transactional relationship between a company and its customers, otherwise the
relationship does not exist in the first place. Hence, empathy as a construct can only be
displayed after an employee has inspired trust in the customer at the moment of truth.
In the short term, Air Mauritius can still be profitable but if service reliability is not
improved, it will become increasingly more difficult for the company to retain existing
customers and attract new ones. Investments in service quality improvement may not
show immediate financial returns, but in the long term it is the only true source of
competitive advantage that will ensure continuous cash flows (Buzzell and Gale, 1987).
To further verify the dimensionality of the 22 items, the raw SERVQUAL scores
were factor analysed. A principal component procedure with Varimax normalised
rotation was used to ensure reliability of results. The aim of a principal component Perceptions of
procedure is to explain the greatest amount of variance in the data set, thus providing service quality
data reduction (Aaker et al., 1998). There are various rotational strategies such as
Varimax, Biquartimax, Quartimax, and Equamax (Jennrich and Sampson, 1966) that
enable a researcher to obtain a clear pattern of loadings, that is, factors that are
somehow clearly marked by high loadings for some variables and low loadings for
others. The option Varimax rotation of the normalised factor loadings was chosen in 503
order to maximise the variances of the dquared normalised factor loadings across
variables for each factor. It is the method that is most commonly used (Clarkson and
Jennrich, 1988).
Kaiser’s criterion ensured that only factors with eigenvalues greater or equal to one
were reported and chosen for interpretation. In essence such a criterion proposed by
Kaiser (1960) suggests that unless a factor extracts at least as much as the equivalent
of one original variable, it should be dropped from factor analysis. The analysis was
constrained a priori to five factors as the SERVQUAL model was hypothesised to have
a five dimensional structure (Frost and Kumar, 2001). Various researchers have used
different criteria to interpret factor loadings, Nunnally (1967) suggests loadings of 0.7
or greater to be reliable while Frost and Kumar (2001) as well as Juwaheer (2004) used
0.4 as cut off points. Consequently, variables with factor loadings greater or equal to
0.45 were deemed to load heavily on a particular factor. The resulting procedure led to
a four-factor structure that explained 63.1 per cent of the variation in the SERVQUAL
model. The corresponding factor loading matrix, including commonalities is shown in
Table III. Commonalities show how much of the variance in the variable has been
accounted for by the four factors (Kinnear and Gray, 1994).
Factor one is made up of seven statements that explain most of the variation (43.6
per cent). It contains statements pertaining to service dimensions such as
responsiveness, assurance and empathy. The result suggests that tourists perceive
enhancing factors of service delivery such as, helpfulness of employees, feelings of
trust and safety dealing with employees, promptness of service, and having the best
interest of passengers at heart, to be all related and part of the same construct. This
factor can be named “service efficiency and affect”. This finding is in line with the
research carried out by Khan and Su (2003), which showed that indeed a hierarchy of
service dimensions exist and that tourists tend to expect more of those services that are
courteous and informative, and convey a feeling of trust and confidence. Five
statements loaded heavily on factor two, explaining 8.4 per cent of total variance. This
factor can be named “service personalisation”, which is already a significant service
dimension in the hospitality industry (Kandampully et al., 2001; Stevens et al., 1995;
Mittal and Lassar, 1996). Research has shown that international tourists visiting
Mauritius attach a lot of importance to “warmth” and reliability of service in hotels
(Juwaheer, 2004). Also, the destination has over the years built a strong reputation for
service excellence, which creates high expectations among passengers as Air
Mauritius is part of the destination experience. Wirtz and Johnston (2003) confirm that
competitive advantage for airlines lies in service delivery, where softer skills such as
warmth, care and anticipation of needs rather than the technicalities of flight
schedules, seat comfort, and punctuality are more important to passengers. The latter
most airlines can master. Factor three pertains to the service dimension of reliability
and explains 6.1 per cent of variance and factor four measured tangibility with 5 per
IJQRM
Factor 1 Factor 2 Factor 3 Factor 4
24,5
Q1 0.5739
Q2 0.6751
Q3 0.5152
Q4 0.7421
504 Q5 0.5958 0.4889
Q6 0.7440
Q7 0.7723
Q8 0.7541
Q9 0.7040
Q10 0.6939
Q11 0.4954
Q12 0.6744
Q13 0.6397
Q14 0.7181
Q15 0.7808
Q16 0.5123
Q17 0.7407
Q18 0.6676
Q19 0.6144
Q20 0.4800 0.5723
Q21 0.4822 0.4552
Q22 0.6303
Table III. Explained variance 4.2323 2.6220 4.0554 2.9577
Factor loadings with Proportion of total variance 0.1924 0.1192 0.1843 0.1344
Varimax rotation of
SERVQUAL scores Note: Factor loadings below 0.45 are not shown in the table
cent of variance explained. Thus, exploratory factor analysis revealed that “service
efficiency and affect” was the most important dimension for international tourists
followed by service personalisation, reliability and tangibility. Hence, as noted by
authors of several studies (Nel et al., 1997; Babakus and Boller, 1992; Carman, 1990), the
SERVQUAL instrument in this study does not consistently measure the same factors
as suggested by Parasuraman et al. (1988) and that indeed dimension structure may be
context specific.
The third objective of the study was to determine the extent to which the identified
service dimensions influenced overall satisfaction and willingness to recommend the
airline to others. The four service quality factors were entered into regression analysis
for that purpose. Table IV reports the results of the stepwise regression model, using
tourists’ overall judgement of service quality as the dependent variable and the four
factors presented above as independent variables. It should be noted that the
dependent variable was measured on a five-point Likert scale from not at all satisfied
to very satisfied. As it can be seen, there are two factors (service efficiency and affect
and tangibility), which explain 21.1 per cent of variance in passengers rating of
satisfaction levels. Correlation analysis between the dependent variable of the model
and service dimensions revealed that correlation exists between the independent
variables. This phenomenon can be explained by the multidimensionality of service
quality constructs in the data set. Using linear combinations such as these reduces, but
cannot eliminate, a potential multi-collinearity problem.
As can be seen from Table IV, Service efficiency and affect appear to be a core Perceptions of
dimension (b ¼ 0:415) in tourists’ evaluation of satisfaction with the airline service. service quality
The F statistic for the regression model was 9.041 with a p value less than 0.000. The
result differs from the study of Cunningham et al. (2002), who found that reliability and
empathy were significant predictors of customer satisfaction among US customers.
The cultural and social background of respondents for this study might account for the
difference. The model also indicates that 78.9 per cent of the variance cannot be 505
explained and therefore further research is needed to identify other factors that
influence service quality at Air Mauritius. Earlier studies have shown that service
quality perceptions can also be influenced by factors such as value (Rust and Oliver,
1994; Zeithaml, 1988), attributions and equity (Bolton and Drew, 1991), information
availability, perceived risk of flying (Cunningham et al., 2002) and atmosphere (Peyrot
et al., 1993). These factors could perhaps explain at least part of the unexplained
variance.
Similarly, a regression model was developed to determine which service quality
factors influence tourists’ likeliness to recommend the airline to others. As shown in
Table V, three factors namely “Service efficiency and affect”, “Tangibility” and
“Reliability” explained 21.7 per cent of variance. The F statistic for the regression
b t p-level
Dependent variables
Overall customer satisfaction
Independent variables
Factor 1 (service efficiency and affect) 0.415 5.424 0.000
Factor 4 (tangibility) 0.155 2.023 0.044
Multiple R 0.459
R2 0.211 Table IV.
Adjusted R 2 0.188 Stepwise regression
Standard error 0.755 analysis results for
F statistic 9.041 customer satisfaction
p-level 0.000 levels
b t p-level
Dependent variables
Overall willingness to recommend airline to others
Independent variables
Factor 1 (service efficiency and affect) 0.334 4.404 0.000021
Factor 4 (tangibility) 0.279 3.679 0.000033
Factor 3 (reliability) 0.168 2.219 0.028112
Multiple R 0.467 Table V.
R2 0.217 Stepwise regression
Adjusted R 2 0.205 analysis results for
Standard error 0.702 willingness to
F statistic 12.618 recommend airline to
p-level 0.000 others
IJQRM model was 12.618 with a p value less than 0.000. The model also indicates that 78.3 per
24,5 cent of variance can be explained by other factors and further research is needed to
identify additional factors influencing tourists’ willingness to recommend the airline.
The regression results for the dependent variables satisfaction levels and
willingness to recommend the airline reflect Bitner’s (1992) proposition that the
service-scape (tangibility) is used by consumers as surrogate indicators of service
506 quality and hence influence satisfaction levels. It is surprising to note that reliability of
service does not contribute to evaluation of satisfaction levels and marginally
contributes to willingness to recommend, while other studies (Cunningham et al., 2002;
Johns et al., 2004; Juwaheer, 2004) have found this service dimension to be a significant
predictor. Cross-cultural differences have been found to be a plausible explanation for
the above (Nel et al., 1997; Winsted, 1997; Ling et al., 2005).
Discussion and conclusions

The overall results suggest that the underlying factors defining service quality seem to
be inconsistent across service providers or contexts. The results of the empirical study
show that tangibility and reliability are clearly perceived by international tourists
while original service dimensions, responsiveness, assurance and empathy are
uni-dimensional to some extent. Further research is required to explicitly conclude if
empathy plays a bigger role than responsiveness and assurance in tourists’ perceptions
of service quality. In fact, Carman (1990) highlighted the multi-faceted nature of some
services that resulted in a number of inter-relationships between dimensions. It would
appear that the number of service dimensions is very much influenced by the context
under evaluation and methods of evaluation (Coulthard, 2004). The SERVQUAL model
showed good internal reliability with high Cronbach alpha scores on all dimensions in
line with other studies. Like some researchers who have suggested that overall
reliability can be improved by changing negatively worded items into positive ones
(Babakus and Boller, 1992; Babakus and Mangold, 1992; Carman, 1990), this study
confirms the latter. The discriminant validity of SERVQUAL was poor for this study
as well, in line with other studies that found great overlap among service dimensions
(Peter et al., 1993). The number of distinct dimensions based solely on factor analysis
results is not the same as with other studies in the airline industry (Frost and Kumar,
2001; Sultan and Simpson, 2000). The variation may be due to differences in data
collection, analysis procedures and nationality of respondents.
The majority of interviewees being of South African nationality could account for
the difference in the hierarchy of dimensions found in this study. In fact, Nel et al.
(1997) made various recommendations on the applicability of SERVQUAL among
South Africans, one of which was that respondents for surveys tend to be Caucasians
with a very good command of the English language. The respondents for this study
are no different as Mauritius as a tourist destination attracts largely a Caucasian
crowd. In the same line of thought, Sultan and Simpson (2000) demonstrated
empirically that there are significant differences in service quality expectations and
perceptions based on nationality as well as overall perceptions of service quality. The
SERVQUAL model differs in its features, that is, the various dimensions and
importance of the dimensions in an international context. Thus, cultural and social
background of respondents may have led to different service quality expectations and
perceptions. Clearly what matters to South Africans when evaluating a service
encounter differs from other countries. Personalisation of service is much more Perceptions of
important to them. As pointed out by Khan and Su (2003), culture prescribes how the service quality
service is delivered, perceived and consumed.
For Air Mauritius, a strategy of service differentiation should start on improving
reliability of service and not only focus on reducing costs. This type of strategy is
common in the airline industry, but few can master the complexities. The company is
able to create high perceptions using tangible cues but under performs on process 507
dimensions, which seem to be of more importance to customers. Customers expect
personalised service, employee’s sincerity and personal warmth in service delivery, to
make the service experience memorable. The regression analyses have shown the
extent to which service efficiency and affect contributes to overall satisfaction and
willingness to recommend the airline to others. Excellent customer service is the result
of all the components being in place, from the right strategic focus and service culture,
to a clear understanding of the service, to good training and people, to good systems
and processes. A service culture should be developed in the organisation that promotes
values and beliefs such as excellence and zero tolerance for mediocrity. Berry et al.
(1994) have outlined the following ten critical lessons for improving service quality that
should become embedded in the organisational culture: listening, reliability, basic
service, service design, recovery, surprising customers, fair play, teamwork, employee
research and servant leadership. Marketing activity of Air Mauritius should centre on
the development of branding strategies to integrate their service, establish identities
and images both for their employees and customers.
The Gaps model, in particular Gaps 1 to 4, should be investigated to identify
potential factors contributing to shortfalls in service delivery at the moment of truth.
Management actions could include empowerment and training of employees which
might increase feelings of ownership of the service encounter, resulting in more
personalised and accountable service for customers (Bowen and Lawler, 1992).
Employees are the most important element in the service delivery process thus
ensuring customer satisfaction (Frost and Kumar, 2001). Successful management of
service encounters can only happen if management has an excellent understanding of
customer expectations and adequate internal communication of these expectations
within the organisation. Benchmarking has become vital as a source of identifying
gaps that exist between customer expectations and company performance as perceived
by customers. The extent to which internal marketing is successful within the
organisation should be investigated. Employee role conflict and ambiguity can act as a
barrier to effective service delivery (Hartline and Ferrell, 1996). Self-efficacy, job
satisfaction and adaptability influence employee’s ability to serve customers in a
manner that enhances the service experience. Service quality issues should not be seen
as a means of solving actual and potential service delivery problems. It must be viewed
as a strategic orientation of a company with the belief that service quality provides real
business advantages over the long term.
A strategy of superior customer service can be generated through a vision of
excellence throughout the organisation. A service organisation that does not have a
shared vision and culture of service excellence will have a tough task acquiring it, as it
cannot be bought. It must be built (Chan, 2000a). Service quality helps in supporting an
organisation’s competitive position in much the same way as cost synergies,
organisational expertise and brand equity do. The consequences of such a strategy is
IJQRM the ability to command premium prices, with high margins, in a global industry where
24,5 small airlines are dependent on repeat business for survival, and with word-of-mouth
praise by customers as an important channel of marketing. Customers want to be
respected and rewarded for their business and not just with frequent-flyer miles, which
have become a commodity, a price of entry into the market (Albrecht, 1992). The thrust
for Air Mauritius should be on providing an experience as opposed to a function. Such
508 a strategy is that anyone can fly airplanes but few organisations can excel in serving
people. And because it is a competence that is hard to build, it is also hard for
competitors to copy or match (Chan, 2000a). Successful service companies have shown
to be consistently excellent listeners to their customers and service quality strategies
are often characterised by customer segmentation, customised service, guarantees,
continuous customer feedback, and comprehensive measurement of company
performance (Sultan and Simpson, 2000).
In addition, perceptions of high service are the result of very carefully planned
service processes. Management and employees should be encouraged to own up their
mistakes and much effort should be spent in sharing experiences of service failures to
use them as learning opportunities. At Southwest Airlines for example, improvement
plans are generated based on past problems to enable the organisation to learn from
mistakes and move on (Laszlo, 1999). Such a climate encourages innovation in service
delivery. Operational systems, procedures, processes should be all oriented towards
serving the needs of customers. Once the right culture is in place, internal
communication processes should be geared towards motivating employees to deliver
high service while external communications should be managed to set the right level of
expectations among consumers. Air Mauritius should also consider the trend of
outsourcing of non-core services (catering, administration, ground handling,
maintenance etc.) and the rationalisation of distribution channels (Driver, 1999) to
focus resources on service delivery.
Survey limitations and directions for future research

This research was exploratory in nature and its limitations should be noted. The
sample size should be considered adequate only for exploratory analysis and a larger
sample is needed to validate the study. Also, the in-flight self completion survey and
mail survey led to a very low response rate due to a number of factors that were
discussed earlier on. The non-response problem for both the in-flight survey and the
mail survey could potentially have led to a bias in the sample.
Passengers other than international tourists should be surveyed to provide a more
holistic picture of service quality at Air Mauritius. A more in-depth study is required to
assess customer expectations and perceptions of service quality from passengers of
different nationalities. Sultan and Simpson (2000) showed that significant differences
in service quality perceptions exist among European and US airline passengers.
Cross-cultural comparisons of a similar nature based on the SERVQUAL model will
provide for rewarding research in Southern Africa. This study focused on literate
passengers with at least a command of the English language, it may be important to
develop ways of measuring service quality in different languages in an international
context.
The link between international tourists’ perceptions of service quality for airlines
and both expectations and perceptions of service at the tourist destination should be
researched. Repetitive studies within the same industry should be conducted to Perceptions of
examine the stability of findings among international tourists visiting Mauritius. Other service quality
factors influencing customer satisfaction and willingness to recommend the airline
should be identified to improve the predictive validity of the regression models used.
Moreover, comparative research on service quality perceptions is needed within other
airlines from the Indian Ocean such as Air Seychelles, Air Austral and Air
Madagascar. Air Mauritius should also undertake benchmarking studies with airlines 509
such as South African Airways, Air France, Emirates and British Airways in order to
improve its service levels. Quality does not improve unless it is measured
(Sureshchandar et al., 2001) and hence longitudinal studies would enhance the
usefulness of the data collected during this research.
References
Aaker, D.A., Kumar, V. and Day, G.S. (1998), Marketing Research, John Wiley & Sons, New York,
NY.
Albrecht, K. (1992), The Only Thing that Matters, HarperCollins, New York, NY.
Asubonteng, P., McCleary, K.J. and Swan, J.E. (1996), “SERVQUAL revisited: a critical review of
service quality”, The Journal of Services Marketing, Vol. 10 No. 6, pp. 62-81.
Babakus, E. and Boller, G.W. (1992), “An empirical assessment of the SERVQUAL scale”, Journal
of Business Research, Vol. 24 No. 3, pp. 253-68.
Babakus, E. and Mangold, G. (1992), “Adapting the SERVQUAL scale to hospital services:
an empirical investigation”, Health Services Research, Vol. 26 No. 6, pp. 767-86.
Berry, L.L., Parasuraman, A. and Zeithaml, V.A. (1994), “Improving service quality in America:
lesson learned”, The Academy of Management Executive, Vol. 7 No. 2, pp. 32-44.
Bitner, M.J. (1990), “Evaluating service encounters: the effects of physical surroundings and
employee responses”, Journal of Marketing, Vol. 54 No. 2, pp. 69-82.
Bitner, M.J. (1992), “Servicescapes: the impact of physical surroundings on customers and
employees”, Journal of Marketing, Vol. 56 No. 2, pp. 57-71.
Bolton, R.N. and Drew, J.H. (1991), “A multi-stage model of customers’ assessment of service
quality and value”, Journal of Consumer Research, Vol. 17 No. 4, pp. 375-84.
Boulding, W., Kalra, A., Staelin, R. and Zeithaml, V.A. (1993), “A dynamic process model of
service quality: from expectations to behavioural intentions”, Journal of Marketing
Research, Vol. 30 No. 1, February, pp. 7-27.
Bowen, D.E. and Lawler, E.E. (1992), “The empowerment of service workers: what, why, how and
when”, Sloan Management Review, Vol. 33 No. 3, pp. 31-9.
Brysland, A. and Curry, A. (2001), “Service improvements in public services using SERVQUAL”,
Managing Service Quality, Vol. 11 No. 6, pp. 389-401.
Buttle, F. (1996), “SERVQUAL: review, critique, research agenda”, European Journal of
Marketing, Vol. 30, January, pp. 8-32.
Buzzell, R.D. and Gale, B.T. (1987), The PIMS Principles, Free Press, New York, NY.
Calvert, P. (2000), “International variations in measuring customer expectations”, paper
presented at the ARL Measuring Service Quality Symposium, Washington, DC,
20-21 October.
Carman, J.M. (1990), “Consumer perceptions of service quality: an assessment of the SERVQUAL
dimensions”, Journal of Retailing, Vol. 66 No. 2, pp. 27-45.
IJQRM Chan, D. (2000a), “The development of the airline industry from 1978 to 1998 – a strategic global
overview”, The Journal of Management Development, Vol. 19 No. 6, pp. 489-514.
24,5
Chan, D. (2000b), “Beyond Singapore girl – grand and product/service differentiation strategies
in the new millennium”, The Journal of Management Development, Vol. 19 No. 6,
pp. 515-42.
Chidambaram, V. (Deputy Chief Executive of Air Mauritius) (1999), “Managing an airline in the
510 next millennium”, African Airlines Conference, 22 October, pp. 1-22.
Clarkson, D.B. and Jennrich, R.I. (1988), “Quartic rotation criteria and algorithms”,
Psychometrika, Vol. 53, pp. 251-9.
Cook, C. and Thompson, B. (2000), “Reliability and validity of SERVQUAL scores used to
evaluate perceptions of library service quality”, The Journal of Academic Librarianship,
Vol. 26 No. 4, pp. 248-58.
Coulthard, L.J.M. (2004), “Measuring service quality – a review and critique of research using
SERVQUAL”, International Journal of Market Research, Vol. 46 No. 4, pp. 479-97.
Cronin, J. and Taylor, S.A. (1992), “Measuring service quality: a re-examination and extension”,
Journal of Marketing, Vol. 56, July, pp. 55-67.
Cronin, J. and Taylor, S.A. (1994), “SERVPERF versus SERVQUAL: reconciling performance
based and perceptions minus expectations measurement of service quality”, Journal of
Crosby, P.B. (1979), Quality Is Free: The Art of Making Quality Certain, New American Library,
New York, NY.
Cunningham, L.F., Young, C.E. and Lee, M. (2002), “Cross-cultural perspectives of service quality
and risk in air transportation”, Journal of Air Transportation, Vol. 7 No. 1, pp. 3-26.
Dabholkar, P.A., Shepherd, D.C. and Thorpe, D.I. (2000), “A comprehensive framework for
service quality: an investigation of critical, conceptual and measurement issues through a
longitudinal study”, Journal of Retailing, Vol. 76 No. 2, pp. 139-73.
Dana, L.P. and Vignali, D. (1999), “British Airways plc”, International Marketing Review, Vol. 16
Nos 4/5, pp. 278-91.
Desombre, T. and Eccles, G. (1998), “Improving service quality in Trust hospitals: lessons from
the hotel sector”, International Journal of Health Care Quality Assurance, Vol. 11 No. 1,
pp. 21-7.
Donnelly, M., Wisniewski, M., Dalrymple, J.F. and Curry, A.C. (1995), “Measuring service quality
in local government: the SERVQUAL approach”, International Journal of Public Sector
Driver, J.C. (1999), “Developments in airline marketing practice”, Journal of Marketing Practice:
Applied Marketing Science, Vol. 5 No. 5, pp. 134-50.
Feigenbaum, A.V. (1951), Quality Control: Principles, Practice and Administration, McGraw-Hill,
New York, NY.
Frost, F.A. and Kumar, M. (2001), “Service quality between internal customers and internal
suppliers in an international airline”, International Journal of Quality & Reliability
Furrer, O., Liu, B.S. and Sudharshan, D. (2000), “The relationships between culture and service
quality perceptions: basis for cross-cultural market segmentation and resource allocation”,
Journal of Service Research, Vol. 2 No. 4, pp. 355-71.
Gabbott, M. and Hogg, G. (1997), Contemporary Services Marketing Management, The Dryden
Press, London.
Goh, K. and Uncles, M. (2003), “The benefits of airline global alliances: an empirical assessment Perceptions of
of the perceptions of business travellers”, Transportation Research Part A: Policy and
Practice, Vol. 37 No. 6, pp. 479-97.
service quality
Grönroos, C. (1978), “A service oriented approach to marketing of services”, European Journal of
Marketing, Vol. 12 No. 8, pp. 588-601.
Hanlon, P. (2003), Global Airlines, Butterworth-Heinemann, Oxford.
Hartline, D.M. and Ferrell, O.C. (1996), “The management of customer contact service employees:
511
an empirical investigation”, Journal of Marketing, Vol. 60, October, pp. 52-70.
Herbig, P. and Genestre, A. (1996), “An examination of the cross cultural differences in service
quality: the example of Mexico and the USA”, Journal of Consumer Marketing, Vol. 13
No. 3, pp. 45-53.
Ingram, H. and Daskalakis, G. (1999), “Measuring quality gaps in hotels: the case of Crete”,
International Journal of Contemporary Hospitality Management, Vol. 11 No. 1, pp. 24-30.
Jain, S.K. and Gupta, G. (2004), “Measuring service quality: SERVQUAL vs. SERVPERF scales”,
Vikalpa, Vol. 29 No. 2, pp. 25-37.
Jennrich, R.I. and Sampson, P.F. (1966), “Rotation for simple loadings”, Psychometrika, Vol. 31,
pp. 313-23.
Johns, N., Avci, T. and Karatepe, O.M. (2004), “Measuring service quality of travel agents:
evidence from Northern Cyprus”, The Services Industries Journal, Vol. 24 No. 3, pp. 82-100.
Johnson, G. and Scholes, K. (1999), Exploring Corporate Strategy, Prentice-Hall Europe, London.
Juran, J.M., Gryna, F. and Bingham, R.S. (1974), Quality Control Handbook, McGraw-Hill, New
York, NY.
Juwaheer, T.D. (2004), “Exploring international tourists; perceptions of hotel operations by using
a modified SERVQUAL approach – a case study of Mauritius”, Managing Service Quality,
Vol. 14 No. 5, pp. 350-64.
Kagis, P. and Passa, V. (1997), “Awareness of service charges and its influence on customer
expectations and perceptions of quality in banking”, Journal of Services Marketing, Vol. 11
No. 2, pp. 105-16.
Kaiser, H.F. (1960), “The application of electronic computers to factor analysis”, Educational and
Psychological Measurement, Vol. 20, pp. 141-51.
Kandampully, J., Mok, C. and Sparks, B.A. (2001), Service Quality Management in Hospitality,
Tourism and Leisure, Haworth Hospitality Press, London and New York, NY.
Kasper, H., Van Helsdingen, P. and De Vries, V. (1999), Services Marketing Management, John
Wiley & Sons, New York, NY.
Khan, M.M. and Su, K.D. (2003), “Service quality expectations of travellers visiting Cheju Island
in Korea”, Journal of Ecotourism, Vol. 2 No. 2, pp. 114-25.
Kilbourne, W., Duffy, J.A., Duffy, M. and Giarchi, G.G. (2004), “The applicability of SERVQUAL
in crossnational measurements of health-care quality”, Journal of Services Marketing,
Vol. 18 No. 7, pp. 524-33.
Kinnear, P.R. and Gray, C.D. (1994), SPSS for Windows Made Simple, Erlbaum (UK), Taylor and
Francis, London.
Krosnick, J.A. and Fabrigar, L.R. (1997), “Designing rating scales for effective management in
surveys”, in Lyberg, L., Biemer, P., Collins, M. and deLeeum, E. (Eds), Survey Management
and Process Quality, Wiley, New York, NY, pp. 141-64.
IJQRM Lam, T. and Woo, K.S. (1997), “Measuring service quality: a test retest reliability investigation of
SERVQUAL”, Journal of Market Research Society, Vol. 39 No. 2, pp. 381-96.
24,5
Lassar, W.M., Manolis, C. and Winsor, R. (2000), “Service quality perspectives and satisfaction in
private banking”, Journal of Services Marketing, Vol. 14 No. 3, pp. 244-71.
Laszlo, G.P. (1999), “Southwest Airlines – living total quality in a service organisation”,
Managing Service Quality, Vol. 9 No. 2, pp. 90-5.
512 Ling, F.I., Lin, K. and Lu, J.L. (2005), “Difference in service quality of cross-strait airlines and its
effect on passengers’ preferences”, Journal of Eastern Asia Society for Transportation
Studies, Vol. 6, pp. 798-813.
Liu, B.S.C., Furrer, O. and Sudharshan, D. (2001), “The effect of culture on behavioural intentions
through service quality perceptions”, paper presented at Rethinking Services, 30th EMAC
Conference, Bergen, 8-11 May.
Llosa, S., Chandon, J.L. and Orsingher, C. (1998), “An empirical study of SERVQUAL’s
dimensionality”, The Service Industries Journal, Vol. 18 No. 2, pp. 16-44.
Luk, S.T.K. (1997), “An examination of the role of marketing culture in service quality”,
International Journal of Contemporary Hospitality Management, Vol. 9 No. 1, pp. 3-21.
Mattila, A.S. (1999), “The role of culture in the service evaluation processes”, Journal of Service
Research, Vol. 1 No. 3, pp. 250-61.
Ministry of Tourism and Leisure, Mauritius (2002), Handbook of Statistical Data on Tourism,
Republic of Mauritius,Vol. 22, Ministry of Tourism and Leisure, Port Louis, Mauritius,
pp. 1-39.
Mittal, B. and Lassar, W.M. (1996), “The role of personalisation in service encounters”, Journal of
Retailing, Vol. 72 No. 1, pp. 95-109.
MK (Air Mauritius) (2004), Annual Report, 2003/2004, MK (Air Mauritius), Port Louis,
Mauritius.
Nel, D., Pitt, L.F. and Berthon, P.R. (1997), “The SERVQUAL instrument: reliability and validity
in South Africa”, South African Journal of Business Management, Vol. 28 No. 3, pp. 113-23.
Newman, K. (2001), “Interrogating SERVQUAL: a critical assessment of service quality
measurement in a high street retail bank”, The International Journal of Bank Marketing,
Vol. 19 No. 3, pp. 126-39.
Nunnally, J.C. (1967), Psychometric Theory, McGraw-Hill, New York, NY.
Parasuraman, A., Berry, L.L. and Zeithaml, V.A. (1985), “A conceptual model of service quality
and its implications for future research”, Journal of Marketing, Vol. 49, Autumn, pp. 41-50.
Parasuraman, A., Berry, L.L. and Zeithaml, V.A. (1988), “SERVQUAL: a multiple-item scale for
measuring consumer perceptions of service quality”, Journal of Retailing, Vol. 4 No. 1,
pp. 12-37.
Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1991), “Refinement and reassessment of the
SERVQUAL scale”, Journal of Retailing, Vol. 67 No. 4, pp. 420-50.
Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1994a), “Alternative scales for measuring
service quality: a comparative assessment based on psychometric and diagnostic criteria”,
Journal of Retailing, Vol. 70 No. 3, pp. 201-29.
Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1994b), “Reassessment of expectations as a
comparison in measuring service quality: implications for further research”, Journal of
Peter, J.P., Gilbert, A.C. and Brown, T.J. (1993), “Caution in the use of difference scores in Perceptions of
consumer research”, Journal of Consumer Research, Vol. 19, March, pp. 655-62.
service quality
Peyrot, M., Cooper, P.D. and Schnapf, D. (1993), “Consumer satisfaction and perceived quality of
outpatient health services”, Journal of Health Care Marketing, Vol. 13 No. 1, pp. 24-32.
Philip, G. and Hazlett, S.A. (1997), “The measurement of service quality: a new P-C-P attributes
model”, International Journal of Quality & Reliability Management, Vol. 14 No. 3,
pp. 260-86. 513
Preston, C.C. and Colman, A.M. (2000), “Optimal number of response categories in rating scales:
reliability, validity, discriminating power and respondent preferences”, Acta Psychologica,
Vol. 104 No. 1, pp. 47-55.
Rhoades, D.L., Waguespack, B. and Treudt, E. (1998), “Service quality in the US airline industry:
progress and problems”, Managing Service Quality, Vol. 8 No. 5, pp. 306-11.
Robledo, M.A. (2001), “Measuring and managing service quality: integrating customer
expectations”, Managing Service Quality, Vol. 11 No. 1, pp. 22-31.
Rust, R.T. and Oliver, R.L. (1994), Service Quality – New Directions in Theory and Practice, Sage
Publications, New York, NY.
Stevens, P., Knutson, B. and Patton, M. (1995), “DINESERV: a tool for measuring service quality
in restaurants”, Cornell Hotel and Restaurant Administration Quarterly, April, pp. 56-60.
Sultan, F. and Simpson, M.C. (2000), “International service variants: airline passenger
expectations and perceptions of service quality”, Journal of Services Marketing, Vol. 14
No. 3, pp. 188-216.
Sureshchandar, G.S., Rajendran, C. and Kamalanabhan, T.J. (2001), “Customer perceptions of
service quality: a critique”, Total Quality Management, Vol. 12 No. 1, pp. 111-24.
Teas, R.K. (1993), “Expectations, performance evaluation and consumers’ perception of quality”,
Journal of Marketing, Vol. 57, October, pp. 18-34.
Teas, R.K. (1994), “Expectations as a comparison standard in measuring service quality:
an assessment of a reassessment”, Journal of Marketing, Vol. 58, January, pp. 132-9.
Voss, C.A., Roth, A.V., Rosenzweig, E.D., Blackmon, K. and Chase, R.B. (2004), “A tale of two
countries’ conservatism, service quality, and feedback on customer satisfaction”, Journal
of Service Research, Vol. 6 No. 3, pp. 212-30.
Winsted, K.F. (1997), “Service encounter expectations: a cross cultural analysis”, Journal of
Transnational Management Development, Vol. 2 Nos. 4, Spring, pp. 5-32.
Wirtz, J. and Johnston, R. (2003), “Singapore Airlines: what it takes to sustain service excellence –
a senior management perspective”, Managing Service Quality, Vol. 13 No. 1, pp. 10-19.
Wisniewski, M. (2001), “Assessing customer satisfaction with local authority services using
SERVQUAL”, Total Quality Management, Vol. 12 No. 7 and 8, pp. 995-1002.
Zaid, A.M. (1994), “Malaysia airlines’ corporate vision and service quality strategy”, Managing
Service Quality, Vol. 4 No. 6, pp. 11-15.
Zaid, A.M. (1995), “Measuring and monitoring service quality at Malaysia Airlines”, Managing
Service Quality, Vol. 5 No. 2, pp. 25-7.
Zeithaml, V.A. (1988), “Consumer perceptions of price, quality and value: a means-end model and
synthesis of evidence”, Journal of Marketing, Vol. 52, July, pp. 2-22.
Zhao, X., Bai, C. and Hui, Y.V. (2002), “An empirical assessment and application of SERVQUAL
in a Mainland Chinese department store”, Total Quality Management, Vol. 13 No. 2,
pp. 241-54.
IJQRM Further reading
24,5 Carr, C. (1990), Front-line Customer Service: 15 Keys to Customer Satisfaction, John Wiley, New
York, NY.
Genestre, A. and Herbig, P. (1996), “Service expectations and perceptions revisited: adding
product quality to SERVQUAL”, Journal of Marketing Theory and Practice, Vol. 4 No. 4,
pp. 72-82.
514 Ghobadian, A., Speller, S. and Jones, M. (1994), “Service quality – concepts and models”,
About the author

Girish Prayag is a Lecturer in Marketing and Tourism in the Department of Management at the
University of Mauritius. His areas of interest include, consumer behaviour in tourism, airline
service quality, destination marketing and branding, and SMEs. Girish Prayag can be contacted
at: girish@webmail.co.za

Measuring
Measuring information quality in information
the order fulfilment process quality
Helena Forslund
Logistics Department, School of Management and Economics, Växjö University, 515
Växjö, Sweden
Abstract
Purpose – The purpose of this paper is to develop a scale for measuring information quality in the
order fulfilment process.
Design/methodology/approach – Measurement scales are developed out of the practical
information needs and a theory review. A state-of-the-art description and further scale development
are based on a survey of the most important suppliers of 136 Swedish companies.
Findings – The paper finds that information quality can be measured with variables as in time,
accurate, convenient to access and reliable. The scales were found to possess unidimensionality,
validity and reliability.
Research limitations/implications – The implications of this paper are mainly theoretical,
providing a foundation for further empirical research.
Practical implications – Even though established customer-supplier relations were studied,
information quality deficiencies were found on all variables. This is an indication of improvement
possibilities.
Originality/value – There is a lack of research on measuring the information quality construct,
which might be the reason for little empirical research on the impact of information quality on logistics
performance.
Keywords Information, Quality systems, Order processing, Performance measurement (quality),
Sweden
Introduction and objective

One of the key issues in supply chain management is the exchange of logistics related
information (e.g. Cooper et al., 1997; Manrodt et al., 1997; Stank et al., 1996a). The fact
that information exchange is related to logistics performance is an axiom; however, few
studies showing this can be found. Stank et al. (1996b) found information exchange
being positively related to supplier performance. Gustin et al. (1995) found that
information availability was related to logistics performance. The implication would
be that the larger quantity of information that is exchanged, the better the supplier’s
performance should be. One problem in past research is that information issues are
discussed in an unspecified and aggregated way (Moberg et al., 2002). Little research
interest is directed towards the quality of information exchanged. Petersen (1999)
concluded that while much has been written about supply chain integration, little
empirical research has been done to determine whether information quality helps to
create better performing supply chains. Quality aspects of information could be related International Journal of Quality &
to information possessing features as timeliness or accuracy. Reliability Management
Vol. 24 No. 5, 2007
This article focuses on information quality the order fulfilment process, which is a pp. 515-524
central process for operational logistics flows (Croxton, 2003). At the operational level, q Emerald Group Publishing Limited
0265-671X
this process focuses on transactions, and includes activities such as order entry, order DOI 10.1108/02656710710748376
IJQRM processing, order fulfilment and order delivery. The value of exchanging operational
24,5 information between companies may sometimes be overlooked due to a focus on
strategic goals (Whipple et al., 2002). The functioning of the order fulfilment process is
affected both by order and forecast information. The customers’ orders put the supply
chain in motion (Croxton, 2003) and trigger the order fulfilment process. If the supplier
pursues a make-to-order strategy, timely order information would seem to be critical
516 for logistics performance, e.g. measured as on-time delivery. The exchange of forecasts
also influences the supplier’s ability to perform. If the supplier pursues a make-to-stock
strategy, reliable forecast information would make it possible to react proactively to
secure logistics performance, e.g. measured as inventory availability.
One reason for the lack of empirical studies highlighting the relationship between
information quality and logistics performance could be the lack of basic research on
defining and measuring constructs, as indicated by Flynn et al. (1990). Mentzer and
Flint (1997) emphasise that without clear definitions and operationalisation, research
results rest on weak ground. Finding a valid and reliable scale for measuring
information quality would be necessary for further empirical research on the impact of
information quality on logistics performance. Theory development and measurement
issues related to reliability and validity are particularly weak in research on quality
issues (Flynn et al., 1994).
The overall objective of this article is to develop a scale for measuring information
quality in the order fulfilment process. A first step is to specify the information need in
the order fulfilment process. Quality constructs are measured multi-dimensionally
(Flynn et al., 1994). A second step is therefore to find measurement scales for order and
forecast information quality. As the state of information quality in the order fulfilment
process is rarely described, a state-of-the-art description is provided. By further scale
development, reliability and validity of the scales are assessed. Results are discussed
and suggestions for further research are provided together with the implications of the
results.
Information need in the order fulfilment process

In the order fulfilment process the supplier is dependent on both the customer’s
information and information internal to the supplier. The information focused on in
this study, is the order-trigging information supplied by the customer. Internal
information from a supplier perspective such as information of inventory levels is not
addressed. Information on price, terms of payment and terms of delivery is exchanged
in the order fulfilment process but is not addressed in this study, as it has a weaker
relation to logistics performance. Both order and forecast information do affect order
fulfilment process performance. Thus the information types exchanged in the order
fulfilment process are restricted to current and future demand expressed as product
content, order quantity and delivery time.
In the activities of the order fulfilment process, different information is needed and
different information features are appreciated. The activity order entry deals with
receiving, entering and editing orders. As short lead times are demanded, the order
arriving in time is critical for on-time delivery. Changes in product content, order
quantity or delivery time are often challenging during the suppressed lead-time. Order
information remaining unchanged is therefore interesting. Unreliable order
information means uncertainty to the supplier, which has to be compensated for by
expensive buffers. The condition of order information is also critical. If the order is Measuring
inconveniently received and needs editing, such as translating product content into the information
company’s item code or manually entering information in the order system, precious
lead time is consumed. Errors in entering an order can be costly (Croxton, 2003). quality
The activity order processing deals with checking inventory, planning order flow
and transportation. Mainly internal information is used. The activity order fulfilment
deals with picking and packing the order. To avoid picking errors it is necessary that 517
the accurate product content be specified. Finally, the activity order delivery deals with
document preparation and transportation. It is obvious that more features than the
quantity of information are valuable and needed in the order fulfilment process. In the
following section, this discussion is enriched by a literature review.
Measuring information quality

A researcher cannot identify variation in a variable unless the variable can be
measured. Can information quality be measured? Quality is a matter of perception, and
is often difficult to measure objectively. Like all other quality measures, it should be
judged by the receiver. The supplier is the information customer in the order fulfilment
process. Measurement is important in accurately representing the construct in
question. Basically there are two types of measurement data, non-metric/qualitative
and metric/quantitative (Hair et al., 1998). As no metric ways of measuring information
quality seem to be available, non-metric measurements are hereafter concentrated on.
Non-metric measurements can be made with either a nominal or an ordinal scale.
Measuring information quality on a nominal scale would imply indicating the presence
or absence of the variable. Using an ordinal scale would mean a higher level of
measurement precision. Variables can be ordered or ranked in relation to the amount of
the attribute possessed. Many scales in the behavioural sciences fall into this category
(Hair et al., 1998).
To generate a scale for measuring information quality, a literature review was
conducted. As few information quality studies were found, closely related constructs
were also examined, which can be referred to as theory building (Flynn et al., 1990). An
overview of studies reviewed is shown in Table I, together with a specification of
construct and context used.
It can be seen that only one study is relevant regarding both construct (information
quality) and context (the order fulfilment process), namely Petersen (1999). This fact
strengthens the insight that this issue is rarely studied. The studies reviewed use
disparate constructs and contexts, and there is no established taxonomy. Variables
have to be carefully analyzed. The following discussion on variables relates to those
authors who defined their variables.
From the studies reviewed, timeliness has been described as being in the (agreed)
time (table), the time when the information customer wants it (English, 1999; Lindau,
1995). Timeliness has also been described as information being concurrent relative to
the situation that the state of information used for decision making corresponds to the
state in the real system, so the data represented are not time-phased with regard to
when it was registered and presented (Closs et al., 1997; English, 1999; Petersen, 1999).
This was referred to as “information float” by English (1999). To make it possible to
capture this variable, order information is operationalised as to whether it arrives
before lead-time is frozen. This variable is referred to as in time. In time was judged to
IJQRM
Type of study Authors Construct Context
24,5
Construct measured Petersen (1999) Information quality Customer order
in empirical study fulfilment cycle
Whipple et al. (2002) Information exchange Alliances between
customer and supplier
518 Gustin et al. (1995) Information availability Order processing
Stank et al. (1996a) Information exchange Between customer and
supplier
Moberg et al. (2002) Information exchange Supply chains
Closs et al. (1997) Information Logistics information
systems
Lindau (1995) Information quality Internal production
planning
Table I. Construct not Coyle et al. (2003) Information quality Not specified
Overview of studies measured English (1999) Information quality General business
reviewed Byrne and Markham (1991) Information Not specified
be possible to measure on an ordinal scale, as orders can occur more or less frequently
in time.
Accuracy, or the state of information received, is also interesting. Order information
coming from a customer might be impaired by obvious mistakes, which must be
corrected by the supplier (information customer) before the order can be confirmed
(Croxton, 2003). Several authors mentioned information accuracy. The variable
accurate seems also relevant to measure on an ordinal scale.
The convenience of accessing information without processing was mentioned by
Closs et al. (1997), English (1999), Keebler et al. (1999) and Petersen (1999). Orders or
forecasts might arrive on fax or in an Excel spread sheet, which implies that the
supplier (information customer) needs to do a considerable amount of work to be able
to use it. The aggregation level might also be un-adapted to the supplier’s (information
customer’s) needs, for example, be on product group level where a product level would
have been preferred. Problems of convenience could cause mistakes and delays.
Convenient to access is another variable judged to be relevant to measure on an ordinal
scale.
Reliability refers to the probability that an order will remain unchanged. Planning
information often has reliability problems (Mattsson, 2002). Moberg et al. (2002)
discuss the importance of information reliability without measuring it. This variable is
referred to as reliable. Reliability could also vary – orders are unchanged more or less
often – and was therefore measured on an ordinal scale. Hence the information quality
construct will be measured using four variables for order information quality. Studies
reviewed have not treated forecast information quality; an adaptation of theory
concerning order information quality was made. Also forecast information quality was
measured with four variables.
Methodology
The measuring of information quality was done as in Table II.
Definition – order Definition – forecast Source (order
Measuring
Variable information quality information quality information quality) Type of scale information
In time Arrives in the agreed Arrives in the agreed English (1999), Ordinal/Likert
quality
time – before lead time – within the Lindau (1995)
time is frozen supplier’s planning
horizon
Accurate Free from obvious Free from obvious Byrne and Markham Ordinal/Likert
519
mistakes mistakes (1991), Croxton (2003),
English (1999),
Petersen (1999),
Whipple et al. (2002)
Convenient to Easy access without Easy access without Closs et al. (1997), Ordinal/Likert
access further processing further processing English (1999),
Keebler et al. (1999),
Moberg et al. (2002),
Petersen (1999),
Reliable The probability that The probability that a Mattsson (2002), Ordinal/Likert Table II.
an order remains forecast remains Moberg et al. (2002) Variables to measure
unchanged unchanged information quality
Eight items were hence used, four for order information quality and four for forecast
information quality. They were formulated as “How do you value the customer’s
order/forecast information: in time, accurate, etc.” on a Likert scale from 1 (strongly
disagree) to 7 (strongly agree). The four information quality variables were pre-tested
on some purchasing managers and planning managers (together with other items as a
larger study was conducted). These pre-tests confirmed the four variables of each scale.
Survey was selected as the empirical data-generating method. A large number of
respondents can be addressed with surveys, which makes it possible to use statistical
analysis techniques such as factor analysis (Hair et al., 1998; Mentzer and Flint, 1997).
To avoid treating information quality issues in an unspecified way, as concluded by
Moberg et al. (2002), it was necessary to let suppliers (information customers) judge
information quality related to a specific customer.
The order fulfilment process in fabricated metal industry, machinery, electronic
equipment and automotive industry is characterised by converging material flows.
Purchases are expected to be frequent and recurrent, and lead-time is expected to be
short. All 432 Swedish purchasing managers of companies in these industries (with 100
or more employees) were addressed in a first survey. A total of 171 replied to the
survey, concerning the relation with their most important Swedish supplier
(information customer). The respondent decided in what way it was considered the
most important supplier. The response rate was hence 39.6 per cent. Contact
information to that supplier was provided for 149 suppliers. New survey
questionnaires were mailed during spring 2003, covering questions on perceived
information quality from that specific customer. The researcher did not control the
population in the second survey; responding customers in the first survey did that. The
distribution by industry and company size varied. The respondents were mainly
representing sales and marketing. A total of 136 responses were returned,
corresponding to a response rate of 91.3 per cent (136/149). The high response rate
IJQRM could be explained by a lot of reminding, not the least by the customer respondents
24,5 from the first survey. Owing to the fact that the population was hard to define, and to
the prevailing high response rate, no analysis of non-response bias was undertaken.
A state-of-the-art description of information quality

Suppliers’ (information customers’) responses to how they perceive order information
520 quality from the designated customer, together with the results of an analysis of
variance (ANOVA) test (with Bonferroni’s pair-wise comparisons) for differences
between the variables, are shown in Table III.
The variable in time – the order arrives in the agreed time, before lead time is
frozen – was rated significantly lower than accurate and convenient to access.
Information being accurate – order information is free from obvious mistakes – is
rated significantly higher than in time and convenient to access. Accuracy could be
expected to be high since important and established customer/supplier relationships
are studied, and accuracy should have been discussed and dealt with. Convenient to
access – easy access without further processing – differs significantly from all other
variables. Reliable information – the probability that an order remains unchanged –
has a mean significantly lower than convenient to access. Reliability was not measured
by the studies reviewed, even if it was mentioned as a variable of information quality
by some authors. Quality deficiencies in order information quality were found. The
standard deviations indicate a spread in the answers.
Suppliers (information customers) were also asked how they perceived information
quality of forecasts from the designated customer, which is shown in Table IV,
together with the results of an ANOVA test (with Bonferroni’s pair-wise comparisons)
for differences between the variables. It was an option not to answer this question if
forecasts were not exchanged. The lower n compared to order information quality
reflects this. Of the suppliers, 18 per cent were not provided with forecast information.
The mean score for convenient to access indicates that suppliers (information
customers) need to transform forecast information to be able to use it. The low mean
score for reliable could also reflect problems in the customer’s forecasting process.
Standard deviations are overall high. Also forecast information quality shows quality
Significant difference with

Order information quality variable n Mean St. dev. variable *
In time – arrives in the agreed time – 135 5.53 1.55 Accurate

before lead time is frozen Convenient to access
Accurate – free from obvious 135 6.18 1.23 In time
mistakes Convenient to access
Convenient to access – easy access 136 5.93 1.51 In time
without further processing Accurate
Reliable
Reliable – the probability that an 136 5.71 1.49 Convenient to access
order remains unchanged
Table III.
Perceived order Notes: * Variables with significant (p , 0:05) different mean using Bonferroni’s tests; ANOVA with
information quality F-statistics: 21.52
Measuring
Significant difference with
Forecast information quality variable n Mean St. dev. variable * information
In time – arrives in the agreed time – 113 5.43 1.67
quality
within the planning horizon
Accurate – free from obvious 112 5.84 1.52 Reliable
mistakes 521
Convenient to access – easy access 111 5.41 1.73
without further processing
Reliable – the probability that a 111 4.82 1.81 Accurate
forecast remains unchanged
Table IV.
Notes: * Variables with significant (p , 0:05) different mean using Bonferroni’s tests; ANOVA with Perceived forecast
F-statistics: 2.72 information quality
deficiencies. Using paired samples t tests the forecast information quality variables
accurate, convenient to access and reliable were found to be significantly (p , 0:05)
lower than the corresponding order information quality variables.
Scale validity and reliability assessment

The primary approach when a theoretical foundation drives scale development is to
rely on factor analysis to ensure scale unidimensionality, followed by scale reliability
and construct validity assessments (Gerbing and Anderson, 1988). To confirm the
importance of certain variables and to reduce the number of variables, factor analyses
of information quality were conducted. The number of factors was determined by
including only those that accounted for eigenvalues higher than one. Following the
principles of Hair et al. (1998), the goal was to obtain measures with factor loadings
greater than 0.4 for only one factor.
After having conducted a factor analysis, it is possible to make a summated scale. A
summated scale is formed by combining several individual variables into a single
composite measure (Hair et al., 1998). Two specific benefits are then provided. First, the
measurement error that might occur in a single variable is reduced. Second, a
summated scale has an ability to represent multiple aspects of a construct in a single
variable. Cronbach’s coefficient alpha is the most commonly used measure for testing
inter-item reliability. Acceptable values are 0.70 or higher, with 0.60 considered
acceptable for new scales (Hair et al., 1998). Both scales were considered new.
Validity is a concept with terminological disparity (Mentzer and Flint, 1997).
Content validity is the assessment of the correspondence of the variables to be included
in a scale and its theoretical definition (Hair et al., 1998). This is the most common way
of measuring validity. Content validity is a subjective measure conducted through
pre-tests. Construct validity indicates whether a scale provides an appropriate
operational definition of a construct. It is an objective measure that can be assessed in
factor analysis (Mentzer and Flint, 1997). In Table V, the result of a factor analysis by
principal components of order information quality is shown.
All variables loaded on one factor, which implied that the solution could not be
rotated (Varimax rotation could not be used). Unidimensionality was thus established
– order information quality could be measured with one summated scale with a mean
of 5.85 and a standard deviation of 1.06. Inter-item reliability for the summated scale
IJQRM measured as Cronbach’s alpha was 0.71. This new scale was therefore considered
24,5 reliable.
Content validity was aimed for as the scale was based on literature reviews and
exposed to pre-tests, as recommended by Flynn et al. (1990). Construct validity was
ensured as all eigenvalues were larger than one and as individual loadings exceed 0.40.
Most of them are above 0.70. Altogether the scale for measuring order information
522 quality was found to possess both content and construct validity.
The variables of forecast information quality also loaded on one single factor. A
Varimax rotation was therefore not possible. Hence, unidimensionality was
established. Forecast information quality could be measured with one summated
scale. The result of the factor analysis is shown in Table VI.
Cronbach’s coefficient alpha revealed an inter-item reliability of 0.87. The scale is
therefore considered reliable. A summated scale could be constructed with a mean of
5.36 and a standard deviation of 1.45.
As no studies of forecast information quality were found, an adaptation from
studies of order information quality was conducted. To ensure content validity,
pre-tests were conducted. Construct validity was further ensured as all eigenvalues
were larger than one, and individual loadings exceed 0.40. In fact, they exceed 0.80. On
the whole, the scale for measuring forecast information quality was considered valid.
Discussion
Even though discussion has occurred through the article, some issues can be further
discussed. State-of-the-art descriptions of order and forecast information quality were
provided, showing quality deficiencies in all variables. The description of order
information quality could be compared to some previous studies. The variable in time
was found to be the lowest ranked variable in studies by Whipple et al. (2002) and Closs
et al. (1997), just as in the current study. Order information accuracy was rated similar
to that in the Whipple et al. (2002) study. Accuracy being ranked highest accords with
Loadings – Cronbach’s
Order information quality variable component 1 alpha Mean St. dev.
In time 0.78
Accurate 0.77
Table V. Convenient to access 0.69
Factor analysis – order Reliable 0.70
information quality Summated scale – order information quality 0.71 5.85 1.06
Loadings – Cronbach’s
Forecast information quality variable component 1 alpha Mean St. dev.
In time 0.83
Table VI. Accurate 0.87
Factor analysis – Convenient to access 0.88
forecast information Reliable 0.80
quality Summated scale – forecast information quality 0.87 5.36 1.45
Closs et al. (1997) who also found accuracy to be the highest ranked variable. The Measuring
variable convenient to access was rated lower than accuracy in the Closs et al. (1997) information
study, which also accords with the results of the current study. The studies reviewed
where information quality or similar constructs were measured all show large standard quality
deviations, which also conforms to the current study. Very little research on forecast
information quality was found. Just as in the study by Gustin et al. (1995), forecast
information was found to be of significantly lower quality than order information. 523
Altogether similarities with previous research were found in this Swedish study.
Addressing pairs of customers and suppliers was necessary in this study, in order
to increase the precision and measure information quality related to a specific
customer. Involving the customer in the order fulfilment process was encouraged by
Croxton (2003). Also Stank et al. (1996a) claimed that customers could increase the
likelihood that their needs will be met by providing high quality information. The
relation between information quality and logistics performance needs further empirical
research.
Conclusion and implications

The overall purpose of this article was to develop a scale for measuring information
quality in the order fulfilment process. The study specified the need for information
concerning product content, quantity and delivery date from the customer. Order
information and its qualities were related to the activities in the order fulfilment
process. The variables in time, accurate, convenient to access and reliable were used to
measure order and forecast information quality. An empirical study resulted in a
state-of-the-art description of the variables. By further scale development, the scales
were found to possess unidimensionality, reliability and validity.
The implications of this paper are mainly theoretical, providing a foundation for
further research. The theoretical contribution is two tested scales for measuring
information quality in the order fulfilment process. As further research, it would be
possible to empirically study the impact of order and forecast information quality on
logistics performance. The descriptive part has managerial implications, as suppliers
perceive deficiencies in order and forecast information quality even in established
customer-supplier relations. This shows promise for improvements.
References
Byrne, P.M. and Markham, W.J. (1991), Improving Quality and Productivity in the Logistics
Process, Council of Logistics Management, Oak Brook, IL.
Closs, D.J., Goldsby, T.J. and Clinton, S.R. (1997), “Information technology influences on world
class logistics capability”, International Journal of Physical Distribution & Logistics
Cooper, M.C., Lambert, D.M. and Pagh, J.D. (1997), “Supply chain management: more than a new
name for logistics”, International Journal of Logistics Management, Vol. 8 No. 1, pp. 1-13.
Coyle, J.J., Bardi, E.J. and Langley, J. Jr (2003), Management of Business Logistics: A Supply Chain
Perspective, Southwestern/Thompson Learning, Cincinnati, OH.
Croxton, K.L. (2003), “The order fulfilment process”, International Journal of Logistics
English, L.P. (1999), Improving Data Warehouse and Business Information Quality, John Wiley &
Sons, New York, NY.
IJQRM Flynn, B.B., Schroeder, R.G. and Sakakibara, S. (1994), “A framework for quality management
research and an associated measurement instrument”, Journal of Operations Management,
24,5 Vol. 11 No. 4, pp. 339-66.
Flynn, B.B., Sakakibara, S., Schroeder, R.G., Bates, K.A. and Flynn, E.J. (1990), “Empirical
research methods in operations management”, Journal of Operations Management, Vol. 9
No. 2, pp. 250-84.
524 Gerbing, D.B. and Anderson, J.C. (1988), “An updated paradigm for scale development
incorporating unidimensionality and its assessments”, Journal of Marketing Research,
Vol. 25, May, pp. 186-92.
Gustin, C.M., Daugherty, P.J. and Stank, T.P. (1995), “The effects of information availability on
logistics integration”, Journal of Business Logistics, Vol. 16 No. 1, pp. 1-21.
Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C. (1998), Multivariate Data Analysis,
Prentice-Hall, Upper Saddle River, NJ.
Keebler, J.S., Manrodt, K.B., Durtsche, D.A. and Ledyard, D.M. (1999), Keeping Score –
Measuring the Business Value of Logistics in the Supply Chain, Council of Logistics
Management, Oak Brook, IL.
Lindau, R.A. (1995), “The impact of high-quality information on performance in manufacturing”,
doctoral thesis, Chalmers University of Technology, Gothenburg.
Manrodt, K.B., Holcomb, M.C. and Thompson, R.H. (1997), “What’s missing in supply chain
management?”, Supply Chain Management Review, Vol. 1 No. 3, pp. 80-6.
Mattsson, S.A. (2002), Logistik i försörjningskedjor, Studentlitteratur, Lund (in Swedish).
Mentzer, J.T. and Flint, D.J. (1997), “Validity in logistics research”, Journal of Business Logistics,
Vol. 18 No. 1, pp. 199-216.
Moberg, C.R., Cutler, B.D., Gross, A. and Speh, T.W. (2002), “Identifying antecedents of
information exchange within supply chains”, International Journal of Physical Distribution
& Logistics Management, Vol. 32 No. 9, pp. 755-70.
Petersen, K.J. (1999), “The effect of information quality on supply chain performance:
an inter-organizational information system perspective”, doctoral thesis, Michigan State
University, East Lansing, MI.
Stank, T.P., Daugherty, P.J. and Ellinger, A.E. (1996a), “Information exchange, responsiveness
and logistics provider performance”, International Journal of Logistics Management, Vol. 7
No. 2, pp. 43-58.
Stank, T.P., Emmelhainz, M.A. and Daugherty, P.J. (1996b), “The impact of information on
supplier performance”, Journal of Marketing Theory and Practice, Fall, pp. 94-105.
Whipple, J.M., Frankel, R. and Daugherty, P.J. (2002), “Information support for alliances:
performance implications”, Journal of Business Logistics, Vol. 23 No. 2, pp. 67-82.
About the author

Helena Forslund is an Assistant Professor of Logistics at Växjö University, Sweden. She received
her PhD from Institute of Technology at Linköping University, Sweden. Helena Forslund can be
contacted at: helena.forslund@vxu.se

Analysing
Modeling and analysing system system failure
failure behaviour using RCA, behaviour
FMEA and NHPPP models
525
Rajiv Kumar Sharma
Department of Mechanical Engineering, National Institute of Technology
(formerly Regional Engineering College), Himachal Pradesh, India, and
Dinesh Kumar and Pradeep Kumar
Department of Mechanical and Industrial Engineering,
Indian Institute of Technology, IIT Roorkee, Roorkee, India
Abstract
Purpose – The aim of this paper is to permit system reliability analysts/managers/practitioners/
engineers to analyse system failure behaviour more consistently and plan suitable maintenance
actions accordingly.
Design/methodology/approach – The paper adopted three important tools, namely, root cause
analysis (RCA), failure mode effect analysis (FMEA), and non-homogeneous Poisson point process
(NHPPP), to build an integrated and helpful framework, able to facilitate the maintenance managers in
decision making. The factors contributing to system unreliability were analysed using RCA and
FMEA. The failure data related to the components are modelled using NHPPP models and are used to
optimise maintenance decisions (repair or replacements) based on cost dimensions.
Findings – The paper finds that the in-depth analysis of a system using RCA and FMEA helps to
create a knowledge base to deal with problems related to process/product unreliability. From the
results it is observed that NHPPP models adequately analyse time-dependent rate of occurrence of
failures. Thus, assisting the maintenance analyst in development of suitable maintenance strategy by
properly understanding the mechanism of failure (through modeling of failure data); adopting
adequate aging management actions (such as predictive or periodic testing) to predict or detect the
degradation of components; and performing cost analysis.
Originality/value – The contemporaneous adoption of the three proposed techniques for failure
analysis will help system reliability engineers/managers/practitioners not only to understand the
failure behaviour of component(s) in the system, but also to plan/adapt suitable maintenance practices
to improve system reliability and availability.
Keywords Failure modes and effects analysis, Fuzzy logic, Modelling
1. Introduction
In today’s turbulent business environment global competition characterised by
both a technology push and a market pull had forced organisations to compete on
various platforms such as faster delivery, price tags, state-of-art-technology and
higher quality dimensions. Various innovative techniques such as total productive
maintenance (TPM), total quality management (TQM), business process International Journal of Quality &
engineering (BPR), manufacturing resources planning (MRP), just in time (JIT), Reliability Management
Vol. 24 No. 5, 2007
etc. are becoming popular among the business houses (Carpinetti et al., 2003; pp. 525-546
Najmi, 2005; Sharma et al., 2005a, 2006, in press). However, benefits obtained from q Emerald Group Publishing Limited
0265-671X
these innovative management practices have often been limited because of the DOI 10.1108/02656710710748385
IJQRM unreliable or inflexible nature of systems/components/parts. The expectation of
24,5 troubled free operation of such systems/components/parts has not been completely
achieved. As a consequence, the processes are becoming vulnerable to various
kinds of disturbances resulting in unexpected sudden and sporadic failures. It is
nearly an unavoidable phenomenon in mechanical systems/components. One can
observe various kinds of failures in past under various circumstances such as
526 nuclear explosions (e.g. the Chernobyl nuclear disaster in 1986), Industrial plant
leakages (the oil pipeline at Jesse, Nigeria in 1998), aero plane crashes, and
electrical network shut downs, etc. which may be due to human error, poor
maintenance, inadequate testing/inspection, unstable operating conditions and so
on. The advances in technology has made the job of reliability/system analyst(s)
more challenging as they have to study, characterise, measure and analyse the
behaviour of system using various techniques such as Markov modeling (MM),
fault tree (FT), failure mode and effect analysis (FMEA), root cause analysis (RCA)
and non-homogeneous Poisson point process (NHPPP) (list of notations used in
this article can be found in Appendix 1). The use of each single technique for
failure analysis can be found in literature studies. For instance, Aneziris and
Papazoglou (2004) used MM to accurately model the dynamics of process. They
presented a method for evaluating the probability of catastrophic failure in process
plants. Hauptmanns (2002, 2004) used the FT approach to analyse propagation of
uncertainties through fault trees for process plant safety using frequency and
probability ranges. Madu (2005) in his paper “Strategic value of reliability and
maintainability” emphasised on the need of maintaining the equipment in good
condition in order to eliminate the sudden and sporadic failures resulting in
production loss. Various models such as RCA, FMEA and Pareto charts were
discussed to uncover the problems related to system unreliability. Sharma (2005)
used the RCA to identify the causes resulting in poor/less/rough plating in
cylinder liners. Numerous researchers such as Teng and Ho (1996), Sankar and
Prabhu (2001), Xu et al. (2002), Guimarães and Lapa (2004),Sharma et al. (2005b)
carried out FMEA research focused on improving traditional FMEA limitations by
using different schemes to identify and prioritise failure causes in engineering
systems. Coetzee (1997) presented application of NHPPP models to analyse failure
data. Saldanha et al. (2001) discussed the application of NHPPP to analyse the
reliability of service water pumps in a typical pressurised water reactor.
As such, the reliability of a system is determined by the constituent sub-systems
and reliability of each subsystem is, in turn, determined by the associated components
and their possible failure modes. In a hierarchical structure (as depicted in Figure 1), it
is usually important that the reliability or system analyst should make use of the
information produced at lower level. There is, therefore, a need to develop a structured
framework to model, analyse and predict the system failure behaviour in a more
realistic manner.
To this effect, in the present paper authors purpose a contemporaneous adoption of
three different methodologies namely RCA, FMEA, and NHPPP to build an integrated
and helpful framework (as depicted in Figure 2), which could prove beneficial to
maintenance/reliability engineers to not only understand the failure behaviour of
component(s) in the system, but also to plan/adapt suitable maintenance practices:
Analysing
system failure
behaviour
527
Figure 1.
Hierarchical structure of
system
Figure 2.
Framework to analyse and
predict system failure
behaviour
.
To ascertain the mechanism of failure (through modelling of failure data) which
will help to adopt adequate aging management actions.
.
To identify the most critical components and select the best maintenance mix
based on cost dimensions.
An industrial case from paper industry is undertaken to discuss the proposed

framework. In brief the framework is summarised as below.
In the qualitative framework, RCA is used to provide comprehensive classification
of causes related to failure of paper machine. To quantify the sources of unreliability
related to process problems a detailed FMEA analysis of forming unit is carried out by
listing all potential failure modes, their causes and effect on system performance. The
numerical values of parameters i.e. Of, S and Od obtained from expert elicitation are
used to compute the RPN score for each failure cause. The discrepancies associated
with the traditional procedure of risk ranking were modelled using a decision support
system based on fuzzy methodology. The system makes use of well-defined
membership functions for Of, S and Od to capture /handle the element uncertainty and
subjectivity. Finally, in quantitative analysis NHPPP models has been used to
IJQRM reproduce the time dependent failure rates and to determine the best maintenance
24,5 strategy for two most important components (wire mat and vacuum pumps) of the
system. By assuming the relevant cost information the economic analysis has been
done to find the trade off between maintenance cost and capital expenditure.
2. Failure analysis methods

528 Among various failure methods as listed in the introduction, the present section
presents a brief account of only those that are used in the study for analysing the
system.
2.1. RCA
RCA is common terminology found in the reliability literature to avoid future occurrence
of failures by pinpointing the causes of problems (Madu, 2000, 2005; Sharma, 2005). It
provides comprehensive classification of causes related to 4 Ms i.e. man, machine,
materials and methods, and thus helps in establishing a knowledge base to deal with
problems related to process/product reliability, availability and maintainability. For
instance, Figure 3 shows how RCA is used to diagnose an unreliable mechanical system.
With respect to man inadequate training, operator’s errors and attitude can contribute to
unreliability, and with respect to machine, problems such as, poor calibrations or
misalignments may result in loss in operational efficiency.
2.2. FMEA
FMEA is yet another powerful tool used by system safety and reliability engineers to
identify critical components/parts/functions whose failure will lead to undesirable
outcomes such as production loss, injury or even an accident. The tool was first
proposed by NASA in 1963 for their obvious reliability requirements. Since then, it has
been extensively used as a powerful technique for system safety and reliability
analysis of products and processes in a wide range of industries – particularly
aerospace, nuclear, automotive and medical (Ebeling, 2001; Connor, 2002). The main
objective of FMEA is to discover and prioritise the potential failure modes by
computing respective RPN, which is a product of Of, S and Od. The critically debated
disadvantage of FMEA based on RPN analysis is that various sets of failure
occurrence probability (Of), severity (S) and detectability (Od) may produce an identical
Figure 3.
Root cause analysis
value; however, the risk implication may be totally different which may result in Analysing
high-risk events going unnoticed. The other disadvantage of the RPN ranking method system failure
is that it neglects the relative importance among Of, S and Od. The three factors are
assumed to have the same importance, but in real practical applications a relative behaviour
importance among the factors exists.
To address these limitations related to traditional FMEA, a decision support system
based on fuzzy set principles is developed to prioritise the failure causes. 529
2.3. NHPPP
Many repairable mechanical systems show a tendency towards long-term reliability
degradation (with repeated overhauls and replacements) of system component(s).
These typically include equipment (systems) and sub-units (sub-systems) where repair
of the system (or sub-system) consists of the replacement or repair of only a small part
of the system (or sub-system). The system is thus not in the “good-as-new” condition
after repair, but in the “bad-as-old” (BAO) condition (the same condition the system
was in prior to failure) known as “minimal repair”. This leads to the system being
subjected to reliability degradation, with an accompanying increase in the failure rate
(ROCOF) (the so-called “sad” trend of Ascher), such systems are not modelled by the
conventional fitting of a statistical distribution function, as successive failures are not
identically and independently distributed. In this case, the NHPPP are used to model
failure/repair process. Log-linear and power law are the two mathematical models,
which are generally used for analysis of NHPPP (Ascher and Feingold, 1984; Calabria
and Pulcini, 2000):
(1) The first NHPPP model with a log-linear rate of occurrence of failures discussed
by Cox and Lewis (1966) behaves well with a1 . 0 and is given as:
r1 ðT Þ ¼ e a0 þa1 T ; 21 , a0 ; a1 , 1; T $ 0: ð1Þ
Using maximum likelihood estimates, the parameters for the model can be
obtained from equation (2):
X
n
2a1 T n 21
T i þ na21
1 2 nT n 1 2 e ¼0 ð2Þ
i¼1

na^1
a^0 ¼ ln : ð3Þ
e a^1 T n 2 1
The process trend is determined by conducting a natural test of hypothesis, i.e.
centeroid or Laplace test. If x0 is the period of observation, and x1, x2, x3 . . . xn
are the arrival values of the independent variables (e.g. time) from x ¼ 0 at
which event occurs, then the test static is given by equation (4):
P
xi =n 2 x0 =2
U¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : ð4Þ
x0 1=ð12nÞ
The statistic compares the centeroid of the observed arrival values with the
mid-point of the period of observation. Under the null hypothesis, U approaches
IJQRM a standard normal distribution.
If H1: a1 – 0, one rejects Ho if U is large. On the other hand, if H1: a1 . 0,
24,5 one rejects Ho if U is large, and if H1: a1 , 0, one rejects Ho if 2 U is large:
.
If U ¼ 0 there is no trend, i.e. the process is stationary.
.
If U , 0 the trend is decreasing, i.e. inter arrival values are tending to
become larger.
530 .
If U . 0 the trend is increasing, i.e. inter arrival values are tending to
become progressively smaller.
(2) The second model (Ascher and Feingold, 1984) based on Weibull distribution is
known as the “Power law process” and is given by equation (5):
r2 ðT Þ ¼ lbT b21 ; l; b . 0; T $ 0 ð5Þ
.
If b . 1, the rate of occurrence of failure increases.
.
When 0 , b , 1, the rate of occurrence of failures decreases.
The maximum likelihood estimates for the parameters of r2 ðTÞ”Power law
process are given as:
n n
b^ ¼ and l^ ¼ : ð6Þ
X
n
Tn Tn
ln
i¼1
Ti
In order to test whether the rate of occurrence of failures is constant, that is

b ¼ 1, the following statistic is employed (Crowder et al., 1996):
Xn t0
V ¼ 2 i¼1 ln ð7Þ
ti
which under the null hypothesis follows a x22n (chi-squared) distribution with 2n
degrees of freedom. Large values of V indicate reliability growth (0 , b , 1),
whereas small ones indicate deterioration (b . 1).
Algorithm for model selection. The steps for selection of appropriate model for rate of
occurrence of failure rðtÞ are:
(1) Obtain the plot of operational time ti against the failure number ni.
(2) Obtain expressions forr1 ðtÞ and r2 ðtÞ using log-likelihood methods.
(3) Perform linear regression (graphical method, as discussed in Appendix 2) and
select rðtÞ.
3. An illustration
As an example a case from process industry (paper mill) situated in northern part of
India (producing 180 tons of paper per day) is taken to discuss the techniques described
in section 2. There are many functional units in a paper mill such as feeding, pulp
preparation, pulp washing, screening, bleaching and preparation of paper (as shown in
Figure 4). The current analysis is based on the study of forming unit of paper machine.
It consists of head box, wire mat and suction box as three main components. Cooked
pulp after processing through number of stages is fed to the head box of a paper Analysing
machine from where (in controlled proportion) it is made to run over the wire mat system failure
running over the rollers. The synthetic belt provides support to run the fibre mat
through the press section and drying section. The dewatering of pulp is accomplished behaviour
with the help of vacuum pumps arranged in parallel. Finally, the moist paper is passed
through the heated rollers along with a synthetic belt (press section) and is dried in a
drier section. The dried paper is finally rolled on the pope reel. Figure 5 shows the 531
components involved in paper manufacturing process.
3.1. Qualitative analysis

RCA. First to diagnose the unreliable aspects of the machine, RCA is carried out by
listing all the possible failure causes related to the machine units i.e. forming, press and
dryer as shown in Figure 6. The comprehensive classification of causes related to the
units help to establish a knowledge base that provides in-depth knowledge regarding
the failure symptoms of the units.
FMEA. Further, to quantify the sources of unreliability related to process problems
and to identify potential system failure modes, their causes and effect on performance
of the system, it is decided to conduct FMEA of one of the unit i.e. forming, by breaking
the unit into its sub-units. In brief the methodology used to compute the scores related
to (Of), (Od), and (S) of failure of various components is discussed as below (Sharma
et al., 2005b):
.
Probability of occurrence of failure (Of). Probability of occurrence of failures is
evaluated as a function of mean time between failures. The data related to mean
time between failures of components are obtained from previous historical records,
Figure 4.
Functional units in paper
mill
Figure 5.
Paper manufacturing
process
IJQRM
24,5
532
Figure 6.
RCA for paper machine
maintenance log-books and is then integrated with the experience of maintenance

personnel. For instance, if MTBF of component lies between two to four months
then probability of occurrence of failure is high (with an occurrence rate 0.5-1 per
cent. and score ranging between 7-8). Table I presents the linguistic assessment of
probability of failure occurrence with corresponding MTBF and scores assigned.
.
Probability of non-detection of failures (Od). The chance of detecting a failure
cause or mechanism depends on various factors such as ability of operator or
maintenance personnel to detect failure through naked eye or by periodical
inspection or with the help of machine diagnostic aids such as automatic
controls, alarms and sensors. For instance, probability of non-detection of failure
of a component through the naked eye is say, 0-5 per cent, and is ranked 1 with
non-detectability remote. The values of Sd for various failure causes reported in
the study are evaluated according to the score reported in Table I.
.
Severity of failure (S). Severity of failure is assessed by the possible outcome of
failure effect on the system performance. If the effect is not noticed and has no
effect on performance than the severity is termed as remote. The slight
deterioration in system performance regards severity as moderate and
significant deterioration regards severity as high. The effect is said to be very
high when non-conforming products are produced or there is appreciable
Likelihood of
Occurrence rate non-detection
Linguistic terms Score/rank no. MTBF (%) Severity effect (%)
Remote 1 .3 years , 0.01 Not noticed 0-5

Low 2 1-3 years 0.01-0.1 Slight annoyance to operator 6-15
16-25
Moderate 4 0.4-1 year 0.1-0.5 Slight deterioration in system performance 26-35
5 36-45
6 46-55
High 7 2-4 months 0.5-1 Significant deterioration in system performance 56-65
8 66-75
Very high 9 , 2 months .1 Production loss and non-conforming products 76-85
10 86-100
behaviour
Analysing
system failure
Linguistic terms and
S and Od)
scale used to measure (Of,
533
Table I.
IJQRM production loss. In the study the data related to mean time to repair (MTTR),
24,5 effect on the quality of the product, are used to obtain score for severity. For
instance, if MTTR of facility/component is less, say lies between 1/4-1/5 hours,
than the effect may be regarded as remote. If external intervention is required for
repairs or MTTR exceeds 1/2 days and there is and if system degrades resulting
in line shut down/production loss than the severity may be regarded as very
534 high.
Table II presents the details of FMEA analysis for the forming unit. The numerical
values of parameters i.e. Of, S and Od are obtained by using the discussed
methodology. The RPN for each failure cause is evaluated by multiplying the factor
scores i.e. (Of £ S £ Od ).For instance, for failure cause FC11 the RPN so obtained is
3 £ 8 £ 4 ¼ 84 respectively. From Table II it is observed that a failure cause FC11 with
high severity, low rate of occurrence, and moderate detectability (7, 3, and 4
respectively) has lower RPN (84) than FC12 where all the parameters are moderate (4, 5,
and 5 yielding an RPN of 100), even though FC11 should have a higher priority for
corrective action. Also, the failure causes FC35 and FC36 though represented by
different sets of Of, S and Od but produce identical RPN number. Such type of
limitations of traditional FMEA are addressed by developing a fuzzy decision support
system (FDSS) based on fuzzy set theory. The basic system architecture of the
proposed system consists of three main modules i.e. knowledge base module and user
input/output interface module as shown in Figure 7.
System input parameters i.e. Of, S and Od, are fuzzified using appropriate
membership functions to determine degree of membership in each input class. For the
output variable, riskiness/priority level both triangular and trapezoidal membership
functions are used (Figure 8(a) and (b)). Multiple experts with different degree of
competencies “C” are used to construct the membership function (Sharma et al., 2005b).
The resulting fuzzy inputs are evaluated in fuzzy inference engine, which makes use
of well-defined rule base. In the study, based on the membership functions of three
input variables Of, S, Od with, with five fuzzy sets in each, a total of 125 rules can be
generated. However, these rules are combined (wherever possible) and the total number
of rules in rule base is reduced to 30. Figure 9 shows the set of rules framed in the
study. Finally, to express the riskiness/criticality level of the failure so that corrective
or remedial actions can be prioritised accordingly, defuzzification is done using
centroid method to obtain crisp ranking from the fuzzy conclusion set. The defuzzified
results are presented in Table III.
Discussion. It is observed from Table III that in traditional FMEA, causes with same
linguistic terms produce different RPN number and hence different ranking but fuzzy
decision support system produces identical output and hence identical ranking. For
instance, FC24, FC29 and FC210, where Of, S and Od are described by Moderate, High
and High, respectively, the defuzzified output is 0.664, for all three events. This entails
that these three should be given the same priority for attention. The RPN method,
however, produces an output of 280, 384 and 320 for these causes and ranks them at
6th, 3rd, and 4th place respectively. This means that FC29 has the highest priority
followed by F210 and F24, which could be misleading. Also, FC24 and FC27 though
represented by different set of linguistic terms i.e. Moderate, High and High and High,
High and Moderate produce identical RPN score by traditional FMEA and both are
Component Function Potential failure mode Potential effect of failure Potential cause of failure Of S Od RPN
Head box
Baffles To discharge pulp on Breaking Non-uniform interrupted flow Broken internals [FC11] 3 7 4 84
Perforated to wire Jamming Corrosion [FC12] 4 5 5 100
plates/bars Blockage Scale build-up [FC13] 6 5 8 240
Slice jet/nozzle Particulate contamination [FC14] 7 7 8 392
Level control To regulate the level Fails to open Unable to provide required conc. Mechanical binding [FC15] 4 7 6 168
valve (pulp þ water)
Fails to open fully Failure to provide full-metered flow Scale building [FC16] 5 8 7 280
Fails to close Loss of flow Broken internals. [FC17] 3 6 7 126
Fourdinier wire
table
Wire mat Carry the pulp Abrasion Holes/marks on the sheet Foreign materials. [FC21] (sand, grit, nails 8 10 9 720
etc.)
Building of fibre mat Rapid wear and shorten the life Lumps/pimples etc. [FC22] 7 9 8 504
Table rolls Support the wire Misalignment Variation in wire tension Roll wear [FC23] 4 9 8 288
Vibrations. [FC24] 5 7 8 280
Suction rolls Drainage of water Buckling/deformation Loss in operation Out of balance [FC25] 6 6 5 180
Improper maintenance [FC26] 3 7 5 105
Dandy rolls Run freely on the Looseness Stock jumps and creates disturbance on Mechanical stresses [FC27] 7 8 5 280
surface Sagging wire
Cough rolls Dewatering the pulp Breaks Sheet formation interrupted (crush and Bearing seizure [FC28] 6 7 6 252
Bearing seizure curl) Jammed shafts [FC29] 6 8 8 384
High temperature [FC210] 5 8 8 320
Pick-up rolls Transfer sheet to Breaks Fails to transfer sheet to pick-up felt Nip pressure [FC211] 6 8 4 192
pickup
felt Bearing seizure Vibrations [FC212] 4 9 8 288
High temperature [FC213] 4 8 9 288
Suction box
Vacuum pumps Complete Fails to operate Piston fails to execute the movement Breaking of piston rod [FC31] 4 9 9 324
dewatering/drainage continuously Seal failure [FC32] 7 6 8 336
Leakage from casing Air may enter the system Excessive radial thrust [FC33] 6 7 4 168
Rotor jamming Pump motor overloaded Lack of lubrication in moving parts. 4 8 6 192
[FC34]
Bearing failure [FC35] 6 7 6 252
Pressure loss Incursion of solid particles into clearances 6 6 7 252
[FC36]
behaviour
Analysing
system failure
FMEA for forming unit

535
Table II.
IJQRM ranked at 6th place but FDSS differentiates them and ranks FC24 at 5th and FC27 at 6th
24,5 place.
Similarly, in case of vacuum pump the causes, FC35 and FC36 which are represented
by different sets of linguistic terms i.e. Moderate, High and Moderate and Moderate,
Moderate and High produce identical RPN score by traditional FMEA, i.e. 252 and are
536
Figure 7.
Modules in fuzzy decision
support system
Figure 8.
Membership function
plots (a) Of, S and Od,(b)
risk priority
Figure 9.
Format of rules
Analysing
Potential cause of failure Traditional output Traditional ranking Fuzzy output Fuzzy ranking
system failure
FC11 84 7 0.533 5 behaviour
FC12 100 6 0.511 7
FC13 240 3 0.617 4
FC14 392 1 0.667 1
FC15 168 4 0.627 3 537
FC16 280 2 0.664 2
FC17 126 5 0.521 6
FC21 720 1 0.699 1
FC22 504 2 0.679 2
FC23 288 5 0.677 3
FC24 280 6 0.664 5
FC25 180 9 0.511 9
FC26 105 10 0.533 8
FC27 280 6 0.646 6
FC28 252 7 0.627 7
FC29 384 3 0.664 5
FC210 320 4 0.664 5
FC211 192 8 0.627 6
FC212 288 5 0.677 3
FC213 288 5 0.674 4
FC31 324 2 0.681 1
FC32 336 1 0.636 2
FC33 168 5 0.611 5
FC34 192 4 0.627 3 Table III.
FC35 252 3 0.627 3 Comparison of traditional
FC36 252 3 0.617 4 and fuzzy results
ranked at 3rd place but FDSS differentiates them and ranks FC35 at 3rd place with
0.627 as output and FC36 at 4th place with 0.617 as output.
3.2. Quantitative analysis

After setting out the component criticalities using FDSS, it is necessary to take
corrective actions in order to improve the system performance. For this purpose proper
understanding of failure behaviour of the components is a must. To this effect the
analysis of failure data is an important facet in the development of maintenance
strategy for equipment. A framework (using NHPPP models as discussed above) which
makes use of modelling of data, goodness of fit tests and regression analysis has been
applied to make repair/replacement decisions regarding the two important components
i.e. wire mat (WM) and vacuum pumps (VP). The WM fails mainly because of corrosion
and abrasion (due to presence of foreign materials clay, sand and other contaminants in
pulp; wear of the roller rubber, roller bearing, roller bending and excessive vibrations
resulting sudden breakdown of roller. Apart from these factors any spot plugging the
wire will give hole in the paper and any foreign material (metallic particles) sticking to
the wire will produce dents and will rapidly wear out the wire. The parameters related
to pump failures that were noticed are lub-oil level of pump systems components,
vacuum pressure of pumps, various leakages, excessive bearing temperatures, seized
bearings, damaged impeller, strainer restricted, vibrations and malfunction indications
IJQRM by sensors or alarms. For the analysis, failure data sets (Table IV) related to both
24,5 components are collected for a period of about eight months. To check the constituency
of rate of occurrence of failures, a plot of the accumulated failures times (operation
times) versus the number of failures is obtained for both components as depicted in
Figure 10. As no linearity is seen for the plotted points so it is concluded that the rate of
occurrence of failures is not constant and clearly time dependent.
538 The model parameters (a1, a0, b and l) for the components are computed and are
listed in Table V. Using the trend test in equation (4) the values are tested. For model
r1 ðtÞ we get, U ¼ 1:235. (WM) and U ¼ 1:19 (VP). Both values are considered large
and the estimated parameters are considered adequate for the model (rejection of the
a1 ¼ 0 hypothesis). Similarly, for model r2 ðtÞ, using equation (7), we get V ¼ 6:844
(WM) and V ¼ 7:7724 (VP). The value is considered small and the estimated
parameters are considered adequate for the model (rejection of the b ¼ 1 hypothesis).
The model equations are represented as:
For wire mat r1 ðt Þ ¼ e {26::5785þ0:0004901t} hr21 r2 ðtÞ ¼ 4:4 £ 1026 t 0:7526 hr21 ð8Þ
For vacuum pump r1 ðtÞ ¼ e {26:92þ0:00062958t} hr21

ð9Þ
r2 ðt Þ ¼ 3:44 £ 1025 t 0:54392 hr21
In order to select which model is applicable for WM and VPs linear regression analysis
Ti (hrs) TBF (hrs) lnTn/Ti

Sr no. WM VP WM VP WM VP
1 5,80 322 – – 1.69 2.049

2 1,400 860 820 538 0.814 1.0671
3 1,790 1,626 390 766 0.568 0.430
4 2,510 2,044 720 418 0.230 0.2013
5 2,800 2,178 290 134 0.120 0.1378
6 P 3,160 P 2,500 360 322 P0 P0
¼ 12,240 ¼ 9,530 ¼ 3:423 ¼ 3:886
Table IV.
Failure data of Notes: WM: wire mat; VP: vacuum pump; TBF: Time between failures; Ti: Time of ith failure, Tn:
components Time of nth failure
Figure 10.
Operational time versus
the number of failures
Parameters
Component Method a1 a0 b l
Wire mat Log-likelihood method 0.0004901 2 6.5789 1.7582 0.0000098

Regression method interval splitting – 1 0.0004632 26.9630 2.1877 0.0000102
Interval splitting – 2 0.0004899 2 6.5933 1.9689 0.0000087
Interval splitting – 3 0.0004460 26.6822 1.9909 0.0000069
Vacuum pumps Log-likelihood method 0.00062058 26.9204 1.5439 0.000034987
Regression interval splitting – 1 0.0004899 26.3214 2.1009 0.000068834
Analysing
estimated parameters)
behaviour
system failure
Table V.
Selection of rðtÞ (based on
539
IJQRM is done (discussed in Appendix 2). The results are presented in Table V. It is observed
24,5 from the table that for WM r1 ðtÞ model adequately fits the rate of occurrence of
failures, considering second interval splitting and log likelihood method. Similarly for
VPs r2 ðtÞ model adequately fits the rate of occurrence of failures, considering third
interval splitting and log likelihood method. For the selected model interval splitting as
shown in Table VI is done. The regression plots of ln rðbj Þ £ bj , for r1 ðtÞ and
540 ln rðbj Þ £ ln bj , for r2 ðtÞ are drawn as shown in Figure 11(a) and (b).
Forecasting rate of occurrence of failures. By performing adequate aging
management actions (predictive maintenance strategies namely condition-based
monitoring (CBM), vibration-based monitoring (VBM) and timely replacements) it is
possible to decrease the expected number of failures. If under the same prevailing
conditions, the two units go on operating i.e. wire mat (3,160, 5,000) and vacuum pumps
(3,000-5,000), then five failures are predicted by using the equations (A1) and (A2)
(Appendix 2). By performing periodic testing, vibration monitoring and timely
maintenance these failures can be reduced to two. The respective failure times (in hrs)
are given in Table VII. To quantify the impact of failure reduction a trend analysis is
Component Selected model Time interval nf bj lnbj rðbj Þ lnrðbj Þ
Wire mat r1 ðtÞ 0-1,000 1 500 6.2146 0.000473409 2 7.65550

1,000-2,000 2 1,500 7.3132 0.001082400 2 6.82855
2,000-3,000 2 2,500 7.8240 0.001590090 2 6.44399
3,000-4,000 1 3,500 8.1605 0.002048800 2 6.19066
Vacuum pumps r2 ðtÞ 0-700 1 350 5.8550 0.000825198 2 7.099
Table VI. 700-1,400 1 1,050 6.9543 0.00149842 2 6.5034
Interval splitting for 1,400-2,100 2 1,750 7.4600 0.00197192 2 6.2259
selected model rðtÞ 2,100-2,800 2 2,450 7.8308 0.00237328 2 6.0432
Figure 11.
Regression plots for (a)
r1 ðtÞ and (b) r2 ðtÞ
performed by means of the NHPPP model, considering the time period (0, 5000) for Analysing
both of them. The developed trend test expressions are represented by equations (10) system failure
and (11) respectively:
behaviour
r1 ðt Þ ¼ e {25:80þð0:0003601t} hr21 ½For wire mat ð10Þ
541
r2 ðtÞ ¼ 4:01 £ 1025 t 0:535 hr21 ½For vacuum pump ð11Þ
Figure 12(a) and (b) displays the comparison of ROCOFs with aging management
actions and without aging management actions for both of the components.
Component First time to failure (hrs) Second time to failure (hrs)
Wire mat 4,000 4,800 Table VII.

Vacuum pumps 3,300 4,200 Failure forecast
Figure 12.
Comparision of ROCOF (a)
wire mat (b) vacuum
pumps
IJQRM 4. Cost analysis
24,5 By assuming the relevant cost information i.e. cost of repair and cost of replacement
the cost analysis has been done to find trade off between maintenance cost and capital
expenditure. Using Type-II and Type-III replacement policies (Coetzee, 1997),
maintenance decisions are optimised in terms of:
. optimal operational time (T *);
542 .
optimal cost/unit time C(T *); and
.
number of minimal repairs (n).
The governing equations, numerical values of parameters, repair and replacement

costs for the components are shown in Table VIII.
From Table VIII it is inferred that, for WM, Type-II policy gives the optimal
replacement frequency T* ¼ 3,226 hours (nearly after five months) and Type-III
policy gives the optimal replacement frequency n* ¼ 11 (approximately) i.e. minimum
11 repairs and then be replaced at the next failure.
Similarly, for VPs we get T* ¼ 12,452 hours (nearly 18 months.) n* ¼ 73
(approximately)
5. Managerial implications
The analysis of system reliability often requires the use of subjective judgments,
uncertain data and approximate system models. Although the results still depends on
the analysts’/experts’ judgement and the quality of the information derived from
different sources, as with any modelling framework one has to exercise great care to
ensure that the data and inputs presented to the method are of good quality because
without these the results could be biased. The application of a proposed framework
that makes use of both qualitative and quantitative approaches (summarised as below)
will help reliability analysts/managers/practitioners to analyse, model and predict the
system failure behaviour in more realistic manner:
Components Wire mat Vacuum pumps

* * *
Governing equation Type-II e a1 T ða1 T 2 1Þ ¼ X 2 1 T ¼ ½C 0 =lðb 2 1Þ1=b
* *
Type-III n ¼ e a0 ðm 2 1Þ=a1 n ¼ C 0 =ðb 2 1Þ
Parametric values a1 4:9 £ 10 2 4 –
a0 26.57 –
b – 1.5439
l – 0.000034987
Approximate costs (Rs £ 1,000) Cp 100 40
Cf 12 10
Results
Optimal operation time T *(hrs) 3,226 12,452
Optimal cost/unit time C (T *)/hr 82.65 11.65
No. of minimal repairs n* 11 73
Table VIII.
*
Cost analysis Note: Where X ¼ C 0 a1 =e a0 , m ¼ X 2 1= ln 2 1 and CðT Þ ¼ C f EðN ðtÞÞ þ C p =T
.
RCA, a failure analysis tool provides in-depth analysis of system unreliability Analysing
aspects by comprehensive classification of causes and also helps to build a system failure
knowledge base for conducting FMEA.
behaviour
.
The application of decision support system based on fuzzy principles not only
addresses the seriously debated disadvantages associated with traditional
procedure for conducting FMEA, but also integrates expert judgment and their
experience in more flexible manner by using well-defined membership functions. 543
.
(The application of NHPPP models assists the maintenance managers in
understanding the failure behaviour of aging components by providing
mathematical model. The model not only helps in forecasting future failures, but
also helps in optimising the maintenance decisions based on cost dimensions
(repair or replacements).
6. Conclusion
In capitalising on the studies available in the literature an attempt has been made by
the authors to develop a contemporary framework that could help the maintenance
engineers/managers/practitioners to model, analyse and predict the system behaviour.
From the study it is concluded that the application of proposed framework which
makes use of both qualitative and quantitative approaches will no doubt help the
maintenance managers/analysts/engineers to adapt/practise suitable maintenance
management strategies in order to improve the reliability and maintainability aspects
of the systems and thereby to improve the overall system performance.
References
Aneziris, O.N. and Papazoglou, I.A. (2004), “Fast Markovian method for dynamic safety analysis
of process plants”, Journal of Loss Prevention in the Process Industries, Vol. 17 No. 1,
pp. 1-8.
Ascher, H. and Feingold, H. (1984), Repairable System Reliability, Modeling, Inference,
Misconceptions and Their Causes, Marcel Dekker, Washington, DC.
Calabria, R. and Pulcini, G. (2000), “Inference and test in modeling the failure and repair process
of repairable mechanical components”, Reliability Engineering and System Safety, Vol. 67
No. 3, pp. 431-3.
Carpinetti, L.C.R., Buosi, T. and Gerolamo, M.C. (2003), “Quality management and improvement:
a framework and a business-process reference model”, Business Process Management
Coetzee, J.L. (1997), “Role of NHPP models in practical analysis of maintenance failure data”,
Reliability Engineering and System Safety, Vol. 56 No. 2, pp. 161-8.
Connor, P.D.T. (2002), Practical Reliability Engineering, Heyden, London.
Cox, D.R. and Lewis, P.A.W. (1966), The Statistical Analysis of Series of Events, Methuen and
Chapman & Hall, London.
Crowder, M.J., Kimber, A.C. and Smith, T. (1996), Statistical Analysis of Reliability Data, Chapman
& Hall, London.
Ebeling, C. (2001), An Introduction to Reliability and Maintainability Engineering, Tata
McGraw-Hill, New York, NY.
IJQRM Guimarães, A.C.F. and Lapa, C.M.F. (2004), “Effects analysis fuzzy inference system in nuclear
problems using approximate reasoning”, Annals of Nuclear Energy, Vol. 31 No. 1,
24,5 pp. 107-15.
Hauptmanns, U. (2002), “Analytical propagation of uncertainties through fault trees”, Reliability
Engineering & System Safety, Vol. 76 No. 3, June, pp. 327-9.
Hauptmanns, U. (2004), “Semi-quantitative fault tree analysis for process plant safety using
544 frequency and probability ranges”, Journal of Loss Prevention in the Process Industries,
Vol. 17 No. 3, pp. 339-45.
Madu, C.N. (2000), “Competing through maintenance strategies”, International Journal of Quality
Madu, C.N. (2005), “Strategic value of reliability and maintainability”, International Journal of
Quality & Reliability Management, Vol. 22 No. 3, pp. 317-28.
Najmi, M.A. (2005), “A framework to review performance measurement systems”, Business
Process Management Journal, Vol. 11 No. 2, pp. 109-22.
Saldanha, P.L.C., Simone, E.A. and Melo, P.F.F. (2001), “An application of non-homogeneous
Poisson point processes to the reliability analysis of service water pumps”, Nuclear
Engineering and Design, Vol. 210 Nos 1-3, pp. 125-33.
Sankar, N.R. and Prabhu, B.S. (2001), “Modified approach for prioritization of failures in a system
failure mode effect analysis”, International Journal of Quality & Reliability Management,
Vol. 18 No. 3, pp. 324-35.
Sharma, R. (2005), “Reduction of process defects in chrome-plated cylinder liners in WDM2 diesel
locomotives using TQM tools”, Industrial Engineering Journal, Vol. 34 No. 3, pp. 17-21.
Sharma, R., Kumar, D. and Kumar, P. (2005a), “FLM to select suitable maintenance strategy in
process industries using MISO model”, International Journal of Quality in Maintenance
Engineering, Vol. 11 No. 4, pp. 359-74.
Sharma, R., Kumar, D. and Kumar, P. (2005b), “Systematic failure mode and effect analysis using
fuzzy linguistic modeling”, International Journal of Quality & Reliability Management,
Vol. 22 No. 9, pp. 886-1004.
Sharma, R., Kumar, D. and Kumar, P. (2006), “Manufacturing excellence through TPM
implementation – a practical analysis”, Industrial Management & Data Systems, Vol. 106
No. 2, pp. 256-80.
Sharma, R., Kumar, D. and Kumar, P. (in press), “Quality costing in process industries through
QCAS – a practical case”, International Journal of Production Research.
Teng, S.H. and Ho, S. (1996), “Failure mode and effects analysis: an integrated approach for
product design and process control”, International Journal of Quality & Reliability
Xu, K., Tang, L.C. and Xie, M. (2002), “Fuzzy assessment of FMEA for engine system”, Reliability
Engineering and System Safety, Vol. 75 No. 1, pp. 17-29.
Further reading
Ascher, H. (1992), “Basic probabilistic and statistical concepts for maintenance of parts and
systems”, IMA Journal of Mathematics Applied in Business and Industry, Vol. 3 No. 3,
pp. 153-67.
Modarres, M., Kaminski, M. and Kritsov, V. (1999), Reliability Engineering and Risk Analysis,
Marcel Dekker, New York, NY.
Appendix 1. Notation Analysing
FMEA failure mode and effect analysis. system failure
RCA root cause analysis. behaviour
NHPPP non-homogeneous Poisson point process.
MTBF mean time between failures. 545
ROCOF rate of occurrence of failures
(Of), (Od), (S) probabilities with respect to failure occurrence, likelihood of non-detection
and severity of failure.
rðtÞ rate of occurrence of failure.
E[N(t)] expected number of failure.
Ti time between the (i 2 1)th failure and ith failure.
a0 ; a1 , parameters of log-linear NHPP model,r1 ðTÞ.
l; b, parameters of power law NHPP model,r2 ðTÞ.
U statistic for testing ROCOF for r1 ðTÞ.
V statistic for testing ROCOF for r2 ðTÞ.
Cp cost of system replacement.
Cf cost of a repair.
C0 cost ratio Cp/Cf.
T* Optimal operational time.
C(T *) Optimal cost per unit time.
x2n chi-squared distribution.
Appendix 2
For r1 ðtÞ
e a0 ða1 tÞ
E ½ N ðt Þ ¼ e 21 : ðA1Þ
a1
For r2 ðtÞ
E ½ N ðt Þ ¼ lt b : ðA2Þ
Considering that the observation period (0, t0) is divided into k arbitrary intervals (0, a1), (a1, a2),
. . . , (ak2 1, t0), an estimate of r½1=2ðaj21 þ aj Þ, is given by:

IJQRM N aj 2 N aj21
r 1=2 aj21 þ aj ¼ : ðA3Þ
24,5 aj 2 aj21
For j ¼ 1, 2, . . . , k, where a0 ¼ 0 and ak ¼ t0 .
Assuming bj ¼ 1=2 (aj21 þ aj ), a plot of rðbj Þ £ bj indicates the shape of the rate of
occurrence of failures, rðtÞ. The choice of k and aj depends on the analyst. However, it is
advisable to test different subdivisions of the observation interval in order to verify that the
546 shape of the plot does not depend on the chosen subdivision. If r1 ðtÞ is appropriate for rðtÞ, then
the plot of ln rðbj Þ £ bj will show a straight line with slope a1 and intercept. a0 On the other hand,
if r2 ðtÞ is appropriate forrðtÞ, the plot of ln rðbj Þ £ ln bj will also show a straight line, but with
slope (b 2 1) and intercept ln l þ ln b.
About the authors

Rajiv Kumar Sharma is Faculty in the Department of Mechanical Engineering, National Institute
of Technology, NIT, Hamirpur, HP, has about four years research/teaching experience and about
20 publications in international journals of repute. He has guided a number of students for their
undergraduate projects. His areas of interest are Quality Control, Production and Operations
Management, Analysis of Quality, Reliability and Maintainability aspects of Systems. Rajiv
Kumar Sharma is the corresponding author and can be contacted at: rksfrne@recham.ernet.in
Dinesh Kumar is Professor in the Mechanical and Industrial Engineering Department, IIT
Roorkee, India. He has about 30 years research/teaching and industrial of experience and 160
publications in international/national journals/conferences. He has guided a number of students
for their undergraduate projects, master dissertations and PhD degrees. His fields of interests are
modelling stochastic behaviour of systems in industry, maintenance engineering and reliability
analysis.
Pradeep Kumar, Professor in the Mechanical and Industrial Engineering Department, IIT
Roorkee, India has about 28 years research/teaching experience. He has guided number of
students for their undergraduate projects, master dissertations and PhD degrees. He has
contributed 282 publications in international/national journals/conferences. His fields of interests
are Advanced Manufacturing Processes, Quality Engineering, Taguchi Methods, Metal Casting,
Supply Chain Management and Industrial Engineering.

Model to
Using a reliability capability benchmark
maturity model to benchmark
electronics companies
547
Sanjay Tiku
Microsoft Corporation, Redmond, Washington, USA, and
Michael Azarian and Michael Pecht
CALCE Electronic Products and Systems Center, University of Maryland,
College Park, Maryland, USA
Abstract
Purpose – This paper seeks to introduce a set of key practices that can be used to assess whether an
organization has the ability to design, develop and manufacture reliable electronic products.
Design/methodology/approach – The ability to design, develop and manufacture reliable
electronic products is defined in the paper in terms of a reliability capability maturity model, which
is a measure of the practices within an organization that contribute to the reliability of the final
product, and the effectiveness of these practices in meeting the reliability requirements of customers.
Findings – The paper presents a procedure for evaluating and benchmarking reliability capability.
Criteria for assigning different capability maturity levels are presented. The paper also presents a case
study corresponding to reliability capability benchmarking of an electronics company.
Originality/value – The paper provides a set of practices for evaluating and benchmarking
reliability capability.
Keywords Product reliability, Electronics industry, Benchmarking
1. Reliability as a competitive opportunity

For any electronics business, time-to-profit is a key metric for establishing product
design, product operation and high level management goals, including cost, schedule,
and social responsibility. Reliability is the ability of a product or system to perform as
intended (i.e. without failure and within specified performance limits) for a specified
time, in its life cycle application environment (Vichare et al., 2004). Since reliability is
associated with preventing or minimizing the likelihood of failure occurrences,
reliability is a risk factor associated with profit making.
Failures generally lead to costs that extend the time-to-profit for a product.
Financial losses can be in the form of loss of market share due to damaged consumer
confidence[1], increase in insurance rates, costs to replace parts, claims for damages[2],
and maintenance of a service infrastructure to handle failures (Dummer et al., 1997).
Legally, most states in the USA operate on the theory of strict liability. Under this law,
a company can be liable for damages resulting from a defect for no reason other than
that one exists, and a plaintiff does not need to prove any form of negligence to win International Journal of Quality &
their case (Pecht and Biagini, 2002). A history or reputation of poor reliability can also Reliability Management
Vol. 24 No. 5, 2007
prevent potential future customers from buying a product, even if the causes of past pp. 547-563
failures have been corrected. Therefore to be competitive, electronics manufacturers q Emerald Group Publishing Limited
0265-671X
need to know how things fail, in addition to knowing how things work. DOI 10.1108/02656710710748394
IJQRM The last decade of the twentieth century witnessed a rapid globalization of all
24,5 businesses. Competitive and regulatory pressures have driven electronics
manufacturers to low-cost manufacturing and to the evolution of a worldwide
supply chain. Electronics manufacturers have become dependent on worldwide
suppliers who provide materials, parts, sub-assemblies and processes. Failures in
electronic products may originate from parts and materials supplied by second or third
548 tier suppliers, or due to defects introduced during assembly and manufacture. These
failures can have significant consequences.
System integrators, who are near the top of the supply chain, generally set the
requirements for system reliability. Parts and manufacturing processes purchased on
the market, often as commodities, are selected based on information provided by
suppliers. However, system integrators cannot wait until they receive the parts or
sub-assemblies to assess if they are reliable. This can be an expensive iterative process.
An upfront evaluation of suppliers based on their ability to meet reliability
requirements can provide valuable competitive advantage. A manufacturer’s
capability to design for reliability and to implement a reliable design through
manufacturing and testing can yield important information about the likelihood that
the company is “mature” enough to provide a reliable product.
2. Capability maturity models

The maturity approach to determine organizational abilities has roots in quality
management. Crosby’s (1996) Quality Management Maturity Grid describes the typical
behavior of a company, which evolves through five phases (uncertainty, regression,
awakening, enlightenment and certainty) in their ascent to quality management
excellence. Since then, maturity models have been proposed for a wide range of
activities, including software development (Bamberger, 1997; Bollinger and McGowan,
1991; Paulk et al., 1993), supplier relationships (Macbeth and Fergusson, 1994),
research and development effectiveness (Szakonyi, 1994a, b), product development
(McGrath, 1996), innovation (Chiesa et al., 1996), collaboration (Fraser et al., 2002;
Fraser and Gregory, 2002), product design (Fraser et al., 2001; Strutt, 2001; Williams
et al., 2003), and reliability information flows (Boersma et al., 2004; Brombacher, 1999;
Sander and Brombacher, 2000, 1999). This paper proposes a metric, called reliability
capability maturity, which electronics manufacturers can use to evaluate the maturity
of the reliability practices of themselves and their suppliers (Tiku and Pecht, 2003a, b;
Tiku et al., 2003):
Reliability capability maturity metric is a measure of the practices within an organization
that contribute to the reliability of the final product, and the effectiveness of these practices in
meeting the reliability requirements of customers.
3. Key reliability practices

The IEEE Reliability Program Standard 1332 (IEEE Standards Board, 1998; Pecht and
Ramakrishnan, 2000), defines broad guidelines for the development of a reliability
program, based on three reliability objectives:
(1) The supplier, working with the customer, should determine and understand the Model to
customer’s requirements and product needs so that a comprehensive design benchmark
specification can be generated.
(2) The supplier should structure and follow a series of engineering activities so
that the resulting product satisfies the customer’s requirements and product
needs with regard to product reliability.
(3) The supplier should include activities that adequately verify that customer’s 549
reliability requirements and product needs have been satisfied.
For each of the IEEE reliability objectives, key practices for evaluating reliability
capability can be assigned. Figure 1 presents eight key practices identified from a
study of reliability standards from the electronics industry (American Institute of
Aeronautics & Astronautics, 2004; Bell Communications Research, 1993; IEC
Technical Committee – 56, 2001; SAE Standards Board, 1998; US DoD, 1980) and
reliability literature (Condra, 2001; Lewis, 1994; Pecht, 1995). These key practices
encompass all aspects of operation in a company from the product reliability
perspective. The purpose of each of these reliability key practices is briefly described in
Table I (Tiku and Pecht, 2003a, b). The following sections illustrate the use of
reliability tasks associated with these key practices in assigning capability maturity
levels to them.
4. Reliability capability maturity levels

Maturity is “the state of being fully grown or developed” (Fraser et al., 2002). From a
reliability perspective, maturity implies that reliability practices within a company are
well understood, are supported by documentation and training, are being continually
monitored and improved by the users, and are effective and efficient.
In our methodology, the reliability capability of a company is assigned five levels
that represent stages in the evolutionary “maturity” of a company. Some of the
nomenclature is adapted from the Software Engineering Institute’s (SEI’s) Capability
Maturity Model (CMM (Paulk et al., 1993; Strutt, 2001)). Associated with each level,
there are reliability tasks (see Table II) that are typically applied by best-in-class
companies. The assignment of reliability tasks was reviewed by reliability
Figure 1.
Key reliability practices
IJQRM
Key reliability practice Purpose
24,5
Reliability requirements and planning To understand the customer’s reliability requirements
(RRP) To generate reliability requirements for products
To plan reliability activities to meet requirements
Training and development (TAD) To enhance the technical and specialized skills of people
550 To ensure that employees understand reliability plans and
goals for products
To track or develop techniques or methods that can impact
reliability
Reliability analysis (RA) To conduct design analysis to identify potential failure
modes and mechanisms
To determine criticality levels of parts or sub-systems
through system modeling
To make reliability predictions for products
Reliability testing (RTST) To explore design limits for products and identify design
flaws
To demonstrate the reliability of products by running tests
To make or modify reliability predictions for products based
on testing
Supply chain management (SCM) To identify sources of parts or processes to satisfy product
reliability requirements
To manage vendors and sub-contractors
To track change notices for sustaining a product through its
lifecycle
Failure data tracking and analysis To track failures from manufacturing, reliability testing and
(FDTA) from field
To conduct failure analysis and identify the root causes of
failures
To record possible corrective actions to remove the root
causes of failures
Verification and validation (VAV) To verify the implementation of the reliability plan
To conduct internal or external audits of reliability activities
To validate reliability predictions from field performance
and record anomalies
Reliability improvements (RIMP) To track changes in reliability requirements of products
To improve product reliability through implementation of
corrective actions
To improve reliability through the use of new methods or
Table I. techniques
Key reliability practices
and their purpose Sources: Tiku and Pecht (2003a, b)
professionals from the electronics industry. For a company to be assigned a level of

maturity, requirements listed at that level and all lower levels need to be fulfilled[3].
The generic definitions of maturity levels are provided below.
4.1. Solely reactive

The “solely reactive” level is defined by the absence of qualities linked to the higher
levels. Companies at this level are essentially ad hoc in their approach to reliability.
These companies are characterized by a lack of written procedures and an ad hoc, or
sometimes chaotic, nature of design, manufacturing and reliability practices. The
Level
Practice Level 1: solely reactive Level 2: repeatable Level 3: defined Level 4: managed Level 5: proactive
Reliability Reliability plans or requirements A separate reliability department Reliability goals are expressed as Reliability goals are established Reliability plan includes details
requirements and that exist are ad hoc, and changed exists a distribution instead of a point for sub-assemblies and on reliability analysis and testing
planning continuously Reliability requirements are estimate components in a product for specific products
based on customer inputs and Reliability goals are based on Reliability goals and plans are Contingency planning is used and
specifications for competitive specific lifecycle conditions for a documented for all products decision criteria for altering the
products product including the schedule of reliability goals are identified
Reliability goals are expressed as activities Reliability plan includes a process
point estimates A reliability plan exists and for communicating results from
includes a list of required reliability activities
resources like materials,
personnel and equipment
Training and Only some informal on-the-job New technologies, modeling or Reliability engineers are trained Reliability engineers are trained Formally trained reliability
development training is provided to employees analysis techniques that impact in statistical methods for to identify failure modes and engineers are part of the
reliability are constantly tracked, reliability prediction and data mechanisms in a product design reliability department
but are not used to make any analysis Reliability engineers are trained Training is provided to reliability
changes Training is provided to business in root cause analysis and managers on how specific
Some reliability training is managers to appreciate how suggesting corrective actions reliability activities can impact
provided to personnel including reliability impacts business A generic reliability training plan reliability
those who are not directly or program exists Proactive support is provided by
associated with the product top management for reliability
training
Reliability Analysis of product design is Point reliability predictions are Materials used in product design Potential failure modes and single Potential failure mechanisms are
analysis minimal, mainly based on made for products using are characterized points of failure are identified for identified for products
manufacturing issues modeling or reliability prediction Adherence to design rules is products Critical failure modes and
handbooks verified The criticality of components in a mechanisms are identified for all
Lifecycle costs of a product are The warranty cost estimates and product design is quantified products
optimized based on reliability vs spares provisioning is made Reliability predictions are Reliability analysis is used to
cost trade-offs based on reliability predictions provided as distributions, and not design specific reliability tests for
as point estimates a product
Reliability testing Only some functional tests are Reliability testing is based on Tests to identify design limits and Detailed reliability test plans The reliability test data are
conducted to determine product customer specifications destruct limits are conducted for exist including sample sizes and analyzed to determine statistical
operation prior to shipping Products are subjected to burn-in all products confidence intervals failure distributions for products
or screening before shipping Reliability testing based on Accelerated tests are tailored for Models for specific failure
Design verification and generic specifications is expected failure mechanisms in mechanisms are used to make
qualification tests are conducted conducted for all products full lifecycle conditions for reliability predictions for
for all products specific products products
Reliability test results are used to Reliability test requirements for
make design changes in products parts supplied by vendors are
prior to production modified and updated
(continued)
practice
Requirements definition
benchmark
levels for each key

at different maturity
551
Table II.
Model to
24,5
552
Table II.
IJQRM
Level
Practice Level 1: solely reactive Level 2: repeatable Level 3: defined Level 4: managed Level 5: proactive
Supply chain Components are procured from Component engineers manage the Contractual agreements In-coming lots are rejected based A supplier rating system is
management any source depending upon parts selection and management containing quality and reliability on supplier’s manufacturing created and maintained
necessity process requirements are signed with quality data Product change notices are
Components are procured from suppliers Technology maturity is evaluated for their effect on
multiple suppliers (with some Vendor or supplier assessments considered during the selection of manufacturability and product
certification) without any further or audits are conducted components reliability
evaluation In-coming lots are rejected based Approved parts and supplier lists Component traceability markings
Techniques like uprating are used are created and maintained based
on supplier’s reliability tests data are tracked to identify any
for qualifying parts for use on qualification reports and changes
outside specs audits Part obsolescence is tracked to
Parts are procured only from ensure continued supply of parts
authorized distributors and not
from part brokers
Failure data Failures during functional testing Pareto charts based on failure Pareto charts based on failure All manufacturing defects, Reliability testing failures are
tracking and are only recorded as yield data sites and failure modes are mechanisms are created and production testing failures and tracked and recorded in a
analysis created and updated regularly updated regularly without any field failures are tracked and database
without any further action further action recorded in a database Traceability of a failed part is
Failure and root cause analysis is Failure analysis reports detailing ensured from manufacture to
conducted on failed products underlying failure mechanisms failure
from all sources from are generated for all products A database of corrective actions
manufacturing to field Failure mechanisms are based on failure modes and
correlated with specific materials mechanisms is maintained and
or processes updated regularly
Verification and The company is in a process of External certifications like ISO Warranty cost estimates and The statistical failure Reliability test conditions are
validation getting some external are obtained for organizational spares provisioning is modified distributions used for reliability modified for current and future
certification processes including the reliability based on field returns predictions are modified based on products based on observed field
activities field failure data failure mechanisms
Reliability predictions are The failure modes and
updated for the products based on mechanisms database is updated
field failure distributions based on new modes and
Internal audits are conducted for mechanisms observed in field
reliability planning, analysis and
testing activities
Verification and Improvements are made only in Corrective actions based on field Engineering change notices for New modeling and analysis New technologies are evaluated
validation processes and not in product failure modes are implemented reliability improvement are techniques are evaluated and and implemented to improve
designs Product reliability requirements issued and implemented implemented to improve product product reliability
are updated due to business or Bill of materials is modified to reliability Failure information is included
marketing considerations exclude parts that have reliability Changes in lifecycle operating for updating the design rules and
problems in field environment initiate a design process control requirements
Recurrence of identified failures is change for a product Corrective actions based on field
prevented in future products failure mechanisms are
implemented
reliability practices, if there exist any, are often changed or modified as a reaction to Model to
crisis situations. Reliability performance depends primarily on the capabilities and benchmark
motivation of individuals, in the absence of any effort at the organizational level. As a
consequence, these companies generally produce products with unstable reliability.
4.2. Repeatable 553

The “repeatable” level is characterized by consistent and repeatable design,
manufacturing and reliability practices. At this level, reliability practices are
disciplined and successes can be repeated. Planning and managing new products are
based on precedents or prior experience with similar products. The company is
generally able to satisfy written customer requirements. Practices that satisfy
established standards or that have become accepted by industry are repeated. These
companies are able to deliver products that can show conformance to codes, standards
or requirements. However, there are little or no data on actual reliability of products.
The reliability activities, such as testing, are generic for all products, and not tailored
for specific applications. Reliability of the products is not assessed based on an
understanding of the actual lifecycle conditions.
4.3. Defined
The “defined” level companies understand and define the reliability requirements and
goals for their products. There is a standardized and consistent documentation for
reliability activities, and a common understanding among employees about their roles
and responsibilities. At this level, specific reliability training is provided to reliability
engineers and managers to ensure that the employees have the knowledge and
expertise to fulfill their assigned roles. These companies are responsive to test and field
failures and conduct analysis of all failures. Companies at this level have established
practices to satisfy initial product reliability requirements, but their practices are not
mature enough to make design changes in existing products. These companies have
limited ability to use feedback to initiate reliability improvements in products.
4.4. Managed
At the “managed” level, companies can design for reliability. The major improvement
over the level-3 company is that defined reliability practices are used to influence
product designs during development as well as during the rest of the product lifecycle.
A documented reliability plan includes a schedule of product specific reliability
activities. These companies can improve reliability by changing product designs to
achieve desired reliability targets. The impact of changes in reliability requirements or
general operating environment also initiates a product design evaluation. All the
failure mechanisms affecting the reliability of the products are investigated and
documented.
These companies are also successfully able to use their supply chain members in
ensuring reliability of products. They create and update a select list of parts and
suppliers based on defined criteria, and the criticality of components used in design is
quantified. These companies lay down requirements for all reliability activities, and
through audits or reviews ensure that these are met.
IJQRM 4.5. Proactive
24,5 The “proactive” level companies are the best-in-practice companies. They are
characterized as being responsive, adaptive and pro-actively focused on continuous
reliability improvement across product lines. These companies use experience, not
only to correct problems, but also to change the nature of the reliability practices that
they use. The feedback from different stages of a product life cycle, such as predictions,
554 simulations, testing, analysis and field performance, is disseminated widely
throughout the company.
The lessons learned from the feedback are incorporated at the development phase of
new products. The feedback influences the manufactured products and may also
impact the reliability management process. In these companies, improvements can
occur by incremental advances in the existing reliability practices or through
innovations using new technologies and methods. Innovations in design of products as
well the manufacturing processes that exploit the best reliability engineering practices
are identified and transferred throughout the company.
5. The evaluation process

The reliability capability evaluation process is comprised of three phases. In the first
phase, initial information about the process is sent to the company being evaluated. A
reliability capability evaluation questionnaire is included for the company to answer
and collect evidence supporting the answers. In the second phase, evaluators visit the
facility, and verify the responses to the questions with the supporting evidence. The
third phase involves the compilation of an evaluation report.
The first phase is initiated by sending information about the concept of reliability
capability and maturity evaluation to the company being evaluated. This helps the
personnel within the company to appreciate the benefits of such evaluation and
encourage them to answer the questions asked during the evaluation with a positive
frame of mind. A questionnaire for the evaluation is sent at least 20 days before the
evaluators visit the company. The evaluation consists of nine sub-sections – eight
sections pertaining to each of the key practices essential to reliability achievement, and
one section on background information about the company. A schedule for the second
phase involving physical evaluation is also included. The respondents are required to
provide “objective evidence”[4] in support of their responses. The evidence may be in
the form of data, reports, policy drafts or other documents.
In the second phase of the evaluation, evaluators visit the facility. The evaluation
team usually includes one representative from the company. The company presents an
overview of their reliability objectives and practices. The evaluation team then reviews
the responses to the questionnaire and the supporting evidence. Additional follow-up
questions are asked and additional supporting information is identified to clarify some
responses and obtain the correct information. Evidence is sought and judgments are
made based on:
.
commitment to perform (leadership, resources);
.
ability to perform (experience, training, tools);
. methodology used to perform (logic, framework, planning);
.
what has been performed (tasks, activities); and
.
how are the results of product performance used (integration at organizational Model to
level). benchmark
In the third phase, the company is provided a draft evaluation report summarizing the
evaluation team’s observations and recommendations for reliability improvement. The
company is typically given one week to review the draft report and provide comments.
A final report incorporating the feedback comments and clarifications is sent to the 555
company, usually within four weeks after the evaluation. Based on the documented
information and responses received, a reliability capability level is assigned to the
company.
6. Case study: a defined company

To assess the practicality of the reliability capability evaluation process, and as a part
of the reliability capability maturity model development, four case studies were
conducted. The details about one of the case studies is presented here. This section
provides a brief profile of the company in terms of its reliability activities, followed by
the results of the evaluation, and the recommendations made.
This company is a leading manufacturer of electronic control products providing
thousands of products to customers in many countries. The warranty of the products
usually ranges from one year to two years, with a limited warranty of five years
provided for some products.
Most of their products are high-end products with specific reliability requirements,
established based on past experience with similar products and customer feedback
questionnaires. Reliability tasks are part of a quality plan, which is different for each
business unit. A custom quality plan is generated for each product keeping in view the
requirements of the customer. Prior to implementation, the quality plan is reviewed by
a cross-functional team, including people dealing with reliability. The company has
reliability testing and failure analysis facilities, although some testing work is also
outsourced to leading test laboratories.
The company does not offer specific “in-house” training to its employees in broad
areas of reliability. However, some of the employees have had outside training in
specific topics like six-sigma, physics-of-failure (PoF) approach, and highly accelerated
life testing (HALT).
The company conducts very limited failure modes and effects analysis (FMEA) for
their product designs. They believe in designing systems and using parts that are
tested to work beyond the expected usage cycles in the application environment. They
“feel” that by adopting this approach, predicting reliability for their products becomes
unnecessary. However, the company does have regular meetings with their service
departments to inform them about potential component failures. Yearly meetings are
also held to plan for reduction in field returns and component failure rates.
The company designs most of their products for a worst case environment for a
nominal ten-year useful life, and to have cumulative failures of less than a fraction of a
percent over the life of the product. Most of the products are designed to internal
specifications. Internal derating guidelines and thermal imaging are used in design.
Materials used in product manufacture are also characterized for their heat resistance
at elevated temperature usage. Any design changes made during a product
development process are followed by re-qualification of the product.
IJQRM An internal product testing guideline has been developed to test a product design.
24,5 The guideline incorporates tests including HALT, temperature cycling, mechanical
cycling, elevated temperature tests, maximum load testing, minimum load testing, and
electrostatic discharge (ESD) resistance tests. A standard series of tests is conducted
for all products within a business unit. The company also conducts 100 percent
end-of-line functional testing for their products. A documented new product checklist is
556 completed before any product goes into mass production.
The company is proficient at understanding and monitoring life-cycle application
conditions for their products. In some products, built-in software is used to assess the
usage. The company also conducts a simulation of the application and collects
customer surveys to get the information. The purpose of these activities is to match
application requirements with conducted tests. The company is currently also looking
at methods for stress-health monitoring.
An approved vendor list is used for parts selection. This is accompanied by regular
supplier audits conducted by the quality assurance group and statistical multiple-lot
sample analysis of incoming parts and materials. The sample analysis includes
mechanical and electrical testing. The selection of parts is generally made by the
design group. The purchasing group is only used to keep track of the schedule and cost
issues. Suppliers of critical parts are controlled directly by engineering. Otherwise after
initial selection, purchasing maintains the control to ensure scheduled supplies. The
company generally prefers to single source parts, except for some commodity items
that are multiple sourced.
The company very rarely uses parts outside their datasheet or supplier
specifications. They use an internally maintained database to specify design ratings
for supplied parts. All the parts used on existing products are approved for use on
other products. Repeated “failures” of parts from a supplier will initiate action at the
corporate level through the quality assurance department. The action can include
exclusion of a supplier from future consideration.
The company relies on its suppliers for testing of parts and for providing
information about any product changes. The company is currently in a process of
developing a new system for assessing and updating the information about the impact
of product change notices (PCNs) on their products. They believe in re-engineering or
redesigning their products and systems rather than rely on finding obsolete parts for
older systems.
The company uses a failure tracking system during and after manufacture.
Manufacturing defects are tracked by corporate quality assurance, which may initiate
a corrective action in some cases if defects rates are high. The post-warranty service
and parts replacement provided by the company to their customers is noteworthy.
Field failures are tracked even after the warranty period is over. Information of failures
is obtained through a failure hotline, defective returns and warranty returns.
All tracked failures are included in a database providing information on the date of
manufacture and date of return. However, shipping and sale dates are not tracked. All
products that are returned from the field are analyzed. If a new failure mode is found, a
new unit is subject to tests to reproduce the failure. The company uses the data from
field returns database to make improvements in their products by removing the failure
causes or defective components. Field failures are tracked through successive
generations of products to identify discrepancies. An improvement or deterioration
initiates an investigation for the cause of the change. Some reliability tests have been Model to
redesigned based on field failures. benchmark
6.1. Evaluation results and recommendations
It was recommended that the company should increase the education and training of
employees responsible for reliability functions in different reliability topics including
component failure mechanisms. Lessons learned from failure analysis could also be 557
incorporated as short courses.
The company should review and update component derating guidelines for all
parts. The older derating guidelines currently used are not useful for new technologies
and products. The process of how a supplier is obtaining derating curves for their parts
also requires revision.
The company does not incorporate failure mechanism identification in their
reliability tests. The testing conducted is customer driven and focused on testing the
operation of the products using power cycling. Although electrical or mechanical
failures may be precipitated by these tests, the company does not conduct specific tests
for precipitating device level failure mechanisms in semiconductor devices. The
company must design these tests for their product, or have these conducted by their
semiconductor suppliers. Generating a repository of cause and effect diagrams for
different failure mechanisms affecting their products is also useful.
There is a need for a better understanding on life tests conducted by suppliers on
parts to determine the service life of these parts under the life-cycle conditions for
company products. For example, lifetime information about a part at 1508C may not be
enough to obtain information about its expected life at 708C without any information
about the failure mechanism. If the failure mechanism is understood, and the model for
the failure mechanism is known, the qualification data from a supplier may
supplement the company’s test data. A better understanding of exactly how long a
product will work without failure in a particular life-cycle application environment is
also useful for adjusting warranties of products. Mapping from application conditions
to distinct failure mechanisms could be valuable to the company.
The parts database and its use should be evaluated. The database appears to be
updated only if some severe problem is observed for some part. The company should
routinely review the reliability test data from a part manufacturer and also consider
not using parts for which there is no qualification data provided by the manufacturer.
If qualification data for a part cannot be obtained from a supplier, the supplier should
be avoided.
Although some tracking is conducted for PCNs, the company should have a
cross-functional team to evaluate all PCNs in terms of their impact on reliability. The
team can also assess the effect of product changes in terms of availability and expected
obsolescence of parts used in existing designs of products. Any issued PCNs should be
mapped to potential failure mechanisms in terms of risks associated with change of
specifications. There should be a further mapping from the PCNs to the bill of
materials (BOM) for the company products. This mapping will ensure that each
business unit gets a list of the “critical” PCNs potentially affecting their products.
The company must assess the hazard rate (possible non-constant failure rate) of all
the field return data to assess trends. This is especially important if an early wearout
mechanism arises. The company should also conduct more data analysis and
IJQRM experimentation to assess the actual reliability of their products. This may provide the
24,5 company with a product differentiation opportunity, which they are currently not
utilizing.
The company currently specifies failure modes as the failure causes for
semiconductor devices. Understanding the root cause of failures and associated
loads can help to effectively remove problems. A fundamental understanding of failure
558 mechanisms should help to improve the lessons learned program. Design should be
verified to make changes to ensure that the loads that precipitate the failure
mechanism are eliminated or reduced. The company was also advised to assess the
effect of any manufacturing change within the company or any manufacturing change
made by suppliers of parts to assess the potential impact on reliability.
The company has engineers that stay aware of the current reliability issues and
conduct some studies to assess “unresolved” reliability concerns. For example, the
company is addressing lead-free solder reliability challenges. However, a dedicated
reliability resource would supplement the knowledge base. The company should utilize
failure analysis laboratory personnel to keep up on industry failure trends on specific
parts. There is also a need to get up to date with current reliability issues with parts
used in products.
6.2. Benchmarking
The company has a separate quality plan for each business unit. Reliability tasks for
each business unit are part of this quality plan. It uses good quality control processes,
complimented by a 100 percent end-of-line functional testing of products. The company
has also invested in reliability engineering and created an infrastructure for reliability
testing and failure analysis that is used as per the quality plan for each product.
The company does not have defined testing procedures that are conducted to
evaluate or guarantee reliability of products. Accelerated testing to prove life-time
reliability for an intended application is not used. Any additional testing is based only
on specific customer requirements. The company does not evaluate PCNs in terms of
their impact on product reliability. Only if a serious problem occurs, an informal
discussion (usually verbal) is used to determine the cause and the effect.
The company does not conduct benchmarking or an internal review of its reliability
practices. There is no reliability improvement plan for products since all products are
designed for a life of more than ten years without an analysis of the actual reliability of
the products. The company does not use the knowledge from failure analysis of field
returns to improve designs and reliability practices across product lines. Only defective
components are replaced in new designs.
The characteristics of this company are typical for a company at the “Defined” level.
The company is assigned a Level-3 maturity in its reliability capability, according to
the characteristics listed above.
7. Use of radar charts for supplier selection

For the reliability capability maturity model, it is not appropriate to sum up a
supplier’s individual scores on different key practices into a single total, and to use this
figure to compare against scores of other suppliers. First, the customer may not require
the same level of capability maturity for all key practices from its suppliers. Second,
each key practice may carry unequal weighting in terms of its contribution to the Model to
overall reliability capability. benchmark
To compare suppliers, and to indicate the extent of match between the customer’s
requirements and the supplier’s capabilities in various key practices, we use the
Surface Measure of Overall Performance (SMOP) or the radar chart approach (Schütz
et al., 1998). The first step in using radar charts is to create a target reliability
capability octagon for the customer, based on the maturity level required for each key 559
practice (Table II). The required maturity levels are then plotted for each key practice
along the eight different axes, and the plotted points are joined to form the customer’s
requirement octagon for reliability capability. In the second step, reliability capability
octagons are drawn for each supplier following the same procedure. Figure 2 illustrates
the comparison between two suppliers against customer’s requirement octagon.
Supplier A whose octagon has the larger area overlap with the customer’s octagon is
selected.
8. Summary and conclusions

In the last decade of the twentieth century, competitive and regulatory pressures have
driven all types of electronics manufacturers to low-cost manufacturing and to the
evolution of a worldwide supply chain. It is essential that reliability is managed across
all the tiers of the supply chain, because reliability is a risk factor associated with profit
making, Therefore, there is a need for pre facto reliability capability evaluation of
electronics suppliers for cost and schedule considerations.
This paper introduces a set of key practices that can be used to assess whether an
organization has the ability to design, develop and manufacture reliable electronic
products. This ability is defined in terms of a reliability capability maturity metric.
Reliability capability maturity is a measure of the practices within an organization
that contribute to the reliability of the final product, and the effectiveness of these
practices in meeting the reliability requirements of customers. This paper defines eight
key reliability practices that form the basis of a strategy for reliability management,
and for reliability capability evaluation.
Five levels of reliability capability maturity along with their characteristics have
been discussed. The five levels represent stages in the evolutionary transition for a
company. To assign a maturity level to a key practice, requirements in terms of
Figure 2.
Using radar charts for
supplier selection
IJQRM reliability tasks have been enumerated. An assessment based on key practices can
24,5 place companies at one of the five maturity levels. Use of radar charts for supplier
selection based on these maturity levels is also presented.
A reliability capability benchmarking process is outlined. Reliability capability
evaluations were conducted for three companies. Details of one case study are
presented. The suggestions and recommendations made in the evaluation reports to
560 the three companies were well received, and steps have already been initiated for
improvement. In one of the companies, the reliability department has been
re-organized, and more resources and personnel have been allocated to reliability
activities. A revised reliability plan is being developed based on our recommendations,
and training of personnel in specific reliability topics has been initiated. In the second
company, with a better maturity rating, existing data collection and root cause analysis
procedures are being remodeled. The database of lessons learned is being made more
comprehensive and made available across different product divisions so that the
design teams can avoid previous mistakes.
The results of the case studies indicate that reliability capability evaluation of a
company can be conducted, not only to assign a maturity level, but also to add value. It
was found that the evaluation can help a company to understand how they can
improve reliability of products by focusing on set of activities identified during the
benchmarking process.
Notes
1. A month after its release in July 2000, Intel recalled its new 1.13 GHz Pentium III
microprocessors. The chips had a hardware glitch that caused them to crash or hang under
certain conditions. Apparently, pressure from AMD led Intel to push products to market
faster than it had in the past, leaving less time for testing. Although less than 10,000 units
were affected, the recall led to embarrassment and a loss of reputation for Intel at a time
when competition in the microprocessor market was at its fiercest (Jayant, 2000).
2. For example, Toshiba was sued in 1999 for selling defective laptop computers. More than 5
million laptops were allegedly built with a defective floppy disk drive controller chip that
would randomly corrupt data without warning. Lawsuits claimed that Toshiba knew about
the defects since the 1980s, but failed to correct them or notify customers. Toshiba agreed to
a $2.1 billion settlement to prevent the case from going to trial (Pasztor and Landers, 1999).
3. An exception arises when some reliability task conducted at a lower level of maturity is
made redundant by a task at a higher level. For example, under reliability analysis, at
level-2, only point reliability estimates are made for products, while at level-4, by making
reliability predictions in the form of distributions, the need to make point estimates is
precluded.
4. Objective evidence is any piece of information that leads two or three independent evaluators
to the same conclusion.
References
American Institute of Aeronautics & Astronautics (2004), “Objective-oriented reliability and
maintainability program data product requirements”, AIAA S-102 Draft Document,
American Institute of Aeronautics & Astronautics, Reston, VA.
Bamberger, J. (1997), “Essence of the capability maturity model”, Computer, June, pp. 112-14.
Bell Communications Research (1993), Generic Requirements for Assuring the Reliability of Model to
Components Used in Telecommunication Equipment, Technical Reference
TR-NWT-000357, Bell Communications Research, Morristown, NJ. benchmark
Boersma, J., Loke, G., Petkova, V.T. and Sander, P.C. (2004), “Quality of information flow in the
backend of a product development process: a case study”, Quality and Reliability
Engineering International, Vol. 20 No. 4, pp. 255-63.
Bollinger, T.B. and McGowan, C. (1991), “A critical look at software capability evaluations”, 561
IEEE Software, Vol. 8 No. 4, July, pp. 25-41.
Brombacher, A.C. (1999), “Maturity index on reliability: covering non-technical aspects of
IEC61508 reliability certification”, Reliability Engineering & System Safety, Vol. 66 No. 2,
pp. 109-20.
Chiesa, V., Coughlan, P. and Voss, C. (1996), “Development of a technical innovation audit”,
Journal of Product Innovation Management, Vol. 13 No. 2, pp. 105-36.
Condra, L.W. (2001), Reliability Improvements with Design of Experiments, Marcel Dekker, New
York, NY.
Crosby, P.B. (1996), Quality Is Still Free: Making Quality Certain in Uncertain Times,
McGraw-Hill, New York, NY.
Dummer, G.W.A., Tooley, M.H. and Winton, R.C. (1997), An Elementary Guide to Reliability,
5th ed., Butterworth-Heinemann, Oxford.
Fraser, P. and Gregory, M. (2002), “A maturity grid approach for the assessment of product
development collaborations”, Proceedings of 9th International Product Development
Management Conference, Sophia, Antipolis, 27-28 May.
Fraser, P., Moultrie, J. and Gregory, M. (2002), “The use of maturity models/grids as a tool in
assessing product development capability”, IEEE International Engineering Management
Conference, IEMC ’02,Vol. 1, pp. 244-9.
Fraser, P., Moultrie, J. and Holdway, R. (2001), “Exploratory studies of a proposed design
maturity model”, Proceedings of 8th International Product Development Management
Conference, University of Twente, Holland, 11-12 June.
IEC Technical Committee – 56 (2001), “Process for assessing reliability of equipment”,
International Electrotechnical Commission (IEC) New Work Item Proposal no. 56/775/NP,
July.
IEEE Standards Board (1998), IEEE Standard Reliability Program for the Development and
Production of Electronics Systems and Equipment, IEEE STD 1332-1998, IEEE Standards
Board, Piscataway, NJ.
Jayant, M. (2000), “Intel recalls fastest Pentium”, Electronic News, September 4.
Lewis, E.E. (1994), Introduction to Reliability Engineering, John Wiley & Sons, New York, NY.
Macbeth, D. and Fergusson, N. (1994), Partnership Sourcing: An Integrated Supply Chain
Management Approach, Pitman Publishing, London.
McGrath, M.E. (Ed.) (1996), Setting the PACE in Product Development: A Guide to Product and
Cycle Time Excellence, Butterworth-Heinemann, Oxford.
Pasztor, A. and Landers, P. (1999), “Toshiba to pay $2b settlement on laptops”, Wall Street
Journal, November 1, p. 1.
Paulk, M.C., Weber, C.V., Garcia, S.M., Chrisis, M.B. and Bush, M. (1993), Key Practices of the
Capability Maturity ModelSM, Version 1.1, Technical Report CMU/SEI-93-TR-025,
ESC-TR-93-178, Software Engineering Institute, Carnegie Mellon University, Pittsburgh,
PA.
IJQRM Pecht, M. (1995), Product Reliability, Maintainability, and Supportability Handbook, CRC Press,
New York, NY.
24,5
Pecht, M. and Biagini, R. (2002), “The business, product liability and technical issues associated
with using electronic parts outside the manufacturer’s specified temperature range”,
Proceedings of the Pan Pacific Microelectronics Symposium, Maui, Hawaii, February 5-7,
pp. 391-8.
562 Pecht, M. and Ramakrishnan, A. (2000), “Development and activities of the IEEE Reliability
Standards Group”, Journal of the Reliability Engineering Association of Japan, Vol. 22 No. 8,
pp. 699-706.
SAE Standards Board (1998), Board Design/process Checklist for Vehicle Electronic Systems, SAE
Document No. 1938 (revised May), SAE Standards Board, Warrendale, PA.
Sander, P.C. and Brombacher, A.C. (1999), “MIR: the use of reliability information flows as a
maturity index for quality management”, Quality and Reliability Engineering International,
Vol. 15 No. 6, pp. 439-47.
Sander, P.C. and Brombacher, A.C. (2000), “Analysis of quality information flows in the product
creation process of high-volume consumer products”, International Journal of Production
Economics, Vol. 67 No. 1, pp. 37-52.
Schütz, H., Speckesser, S. and Schmid, G. (1998), “Benchmarking labour market performance and
labour market policies: theoretical foundations and applications”, Discussion Paper No. FS
I 98-205, Social Science Research Center, Berlin.
Strutt, J.E. (2001), “Reliability capability maturity briefing document”, Report No. R-03/2/1,
Reliability Engineering & Risk Management Centre, Cranfield University, Cranfield.
Szakonyi, R. (1994a), “Measuring R&D effectiveness – I”, Research Technology Management,
Vol. 37 No. 2, pp. 27-32.
Szakonyi, R. (1994b), “Measuring R&D effectiveness – II”, Research Technology Management,
Vol. 37 No. 3, pp. 44-55.
Tiku, S. and Pecht, M. (2003a), “Auditing the reliability capability of electronics manufacturers”,
Proceedings of IPACK 03: International Electronic Packaging Technical Conference and
Exhibition, Maui, Hawaii, July 6-11.
Tiku, S. and Pecht, M. (2003b), “Reliability capability assessment methodology”, Proceedings of
IMAPS Brazil 2003, the International Technical Symposium on Packaging, Assembling
and Testing & Exhibition, Campinas – SP, Brazil, August 6-8.
Tiku, S., Pecht, M. and Strutt, J. (2003), “Organizational reliability capability”, Proceedings of
Canadian Reliability and Maintainability Symposium, Ottawa, Canada, October 16-17.
United States Department of Defense (US DoD) (1980), United States Department of Defense (US
DoD) Reliability Program for Systems and Equipment Development and Production,
MIL-STD-785B, US DoD, Washington, DC.
Vichare, N., Rodgers, P. and Pecht, M. (2004), “In situ temperature measurement of a notebook
computer – a case study in health and usage monitoring of electronics”, IEEE
Transactions on Device and Materials Reliability, Vol. 4 No. 4, pp. 658-63.
Williams, K., Robertson, N., Haritonov, C.R. and Strutt, J. (2003), “Reliability capability
evaluation and improvement strategies for subsea equipment suppliers”, Journal of the
Society for Underwater Technology, Vol. 25 No. 4, pp. 165-74.
About the authors Model to
Sanjay Tiku has a MS and PhD degree in Mechanical Engineering from the University of
Maryland, College Park. He currently works for Microsoft Corporation. Previously, he worked at benchmark
the Research Center of Tata Motors in India, and he also held a lecturer position in Mechanical
Engineering at Government College of Engineering and Technology, Jammu, India. His research
interests include quality and reliability of electronic products and electronic parts selection and
management. He has written several papers and book chapters in this area. He is a member of the
IEEE, IMAPS, and the academic honor society Phi Kappa Phi. Sanjay Tiku is the corresponding 563
author and can be contacted at: stiku22@gmail.com
Michael Azarian holds a PhD in Materials Science and Engineering from Carnegie Mellon
University, a Masters degree in Metallurgical Engineering and Materials Science from Carnegie
Mellon, and a Bachelors degree in Chemical Engineering from Princeton University. He has
published in the fields of nano-tribology, scanning probe microscopy, structure and properties of
thin films, and colloid science, and holds five US patents for inventions in data storage and
contamination control. His research interests include reliability of photonic and high-speed
electronic devices, failure mechanisms in electronic components, tribology of the magnetic
head-disk interface, and sensor technology. He has also been an invited conference speaker on
nano-tribology, instructor of tribology at IBM, and guest lecturer on optoelectronic reliability at
San Jose State University.
Michael Pecht has a BS in Acoustics, an MS in Electrical Engineering and an MS and PhD in
Engineering Mechanics from the University of Wisconsin at Madison. He is a Professional
Engineer, an IEEE Fellow and an ASME Fellow. He has received the 3M Research Award, and
the IMAPS William D. Ashman Memorial Achievement Award for his contributions. He served
as chief editor of the IEEE Transactions on Reliability for eight years. He has written 18 books on
electronic products development, use and supply chain management. He has also edited a series
of books on the Asian electronics industry including a recent book titled The Chinese Electronics
Industry. He is the founder and the Director of the CALCE Electronic Products and Systems
Center at the University of Maryland and a Chair Professor. He is chief editor for Microelectronics
Reliability and an associate editor for the IEEE Transactions on Components and Packaging
Technology. He has consulted for over 50 major international electronics companies, providing
expertise in strategic planning, design, test, IP and risk assessment of electronic products and
systems.

View publication stats

A Meta-Analysis of The Effect of TQM On Competitiv

Uploaded by

Copyright:

Available Formats

You might also like

A Meta-Analysis of The Effect of TQM On Competitiv

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Meta-Analysis of The Effect of TQM On Competitiv

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

A meta-analysis of the effect of TQM on competitive advantage

Article in International Journal of Quality & Reliability Management · May 2007

The user has requested enhancement of the downloaded file.

Volume 24 , Issue 5 , pp. 442-563 , 2007

A meta-analysis of the effect of TQM on competitive advantage

Assessing international tourists' perceptions of service quality at Air Mauritius

Measuring information quality in the order fulfilment process

Using a reliability capability maturity model to benchmark electronics companies

TQM and competitive advantage

Major components of TQM

Leadership and top management commitment

Training and education

Variables and methodology

Data collection methods

Construct validity and variation of variables

Measurements of independent and dependent variables

context is to approximate the regression database and resulting estimated regression

demonstrated if all independent variables are statistically significant. H3 would be

Analysis, results, and discussion

Culture Leadership Teams Training Process

Table III. Culture 0.61 0.60 0.88 0.67

Bias Coefficient/VIF Culture Leadership Process Teams Training

0 Coef 1.44 0.64 1.61 2 0.37 2 0.45

H3 is indeterminate since all confidence intervals involving leadership cover zero.

Independent variables Observations Prediction interval at means

(1) culture, (2) process 18 (1.39, 5.71)

With these Padjusted

Culture Leadership Process Teams Training

Q 343.26 373.59 353.68 332.94 413.10

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Motives for seeking certification

Reliability and validity tests

Analysis of the means

Improving productivity and/or efficiency 0.782 20.039

Increased productivity and/or efficiency 0.767 0.076

Expected and perceived benefits analysis

Pair 1 (I) (Improving – increased) productivity and/or

484 Factor analysis

Reduced internal costs 0.789

Operational improvement 0.435 * * 0.259 * * 0.347 * * 20.074 0.046

Recommendations for practitioners

About the author

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

In addressing these objectives, a concise synthesis of the extant literature on the

Methods of measuring service quality

Sample design and method of data collection 499

Dimensions Item coefficient (alphas) Reliability item alpha Standardised

Tangibles 4 0.7323 0.7471

Tangibles 4.440 4.125 2 0.315 14.7 20.046

Discussion and conclusions

Survey limitations and directions for future research

About the author

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Introduction and objective

Information need in the order fulfilment process

Measuring information quality