See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/222693856

Utilizing data envelopment analysis to

benchmark safety performance of construction
contractors

ARTICLE in INTERNATIONAL JOURNAL OF PROJECT MANAGEMENT · JANUARY 2010

Impact Factor: 1.53 · DOI: 10.1016/j.ijproman.2009.04.002

CITATIONS DOWNLOADS VIEWS

31 1,043 180

3 AUTHORS, INCLUDING:

Shaher Rababeh Khaled Hesham Hyari

Hashemite University Hashemite University
21 PUBLICATIONS 41 CITATIONS 23 PUBLICATIONS 109 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Shaher Rababeh

Retrieved on: 09 September 2015
 Available online at www.sciencedirect.com

International Journal of Project Management 28 (2010) 61–67

www.elsevier.com/locate/ijproman

Utilizing data envelopment analysis to benchmark safety

performance of construction contractors
Mohammad S. El-Mashaleh a,*, Shaher M. Rababeh b, Khalied H. Hyari a
a
Department of Civil Engineering, The Hashemite University, P.O. Box 150459, Zarqa 13115, Jordan
b
Department of Architectural Engineering, The Hashemite University, P.O. Box 150459, Zarqa 13115, Jordan

Received 20 August 2008; received in revised form 9 April 2009; accepted 14 April 2009

Abstract

The purpose of this paper is to utilize data envelopment analysis (DEA) to benchmark safety performance of construction contrac-
tors. DEA has been recognized as a robust tool that is used for evaluating the performance of business organizations. The proposed
approach is deployed based on empirical data collected from 45 construction contractors. On a scale of 0–1.0, DEA analysis assesses
the relative efficiency of every contractor relative to the rest of the contractors in terms of safety performance. For inefficient contractors,
DEA analysis provides quantitative guidance on how to become efficient.
Ó 2009 Elsevier Ltd and IPMA. All rights reserved.

Keywords: Organization resources; Safety and health; Benchmarking; Performance measurement; Data envelopment analysis

1. Introduction (EMR) [27,28]. OSHA recordable incidence rates are based

The purpose of this paper is to utilize data envelopment
analysis (DEA) to benchmark safety performance of con-
struction contractors. DEA has been recognized as a
robust tool that is used for evaluating the performance of
organizations such as business firms, hospitals, government
agencies, educational institutions, etc. DEA is well-
deployed in other industries. DEA is a nonparametric lin-
ear programming approach that produces a single measure
of efficiency for each unit relative to its peers. It enables
firms to assess their relative efficiency compared to other
firms in the industry.
Construction literature includes several methods for
assessing safety performance of construction contractors.
Two of the most commonly used ones are OSHA record-
able incidence rates and experience modification rating
* introduced in the development of the weights and the
Corresponding author. Tel.: +962 5 390 3333x4829; fax: +962 5 382
6348.
E-mail address: mashaleh@hu.edu.jo (M.S. El-Mashaleh).
on the US Occupational Safety and Health Act (1970),
which requires employers to record and report accident
information. Incidents are recorded and a formula is used
to compute the incidence rates.
EMR, on the other hand, is established by independent
rating bureaus. It dictates the contractor’s premium of the
workers’ compensation insurance. EMR formula is criti-
cized for its complexity and because of the existence of dif-
ferent versions in practice [19]. It is also argued that EMR
is sensitive to company size [12,23,33].
Ng et al. [31] develop a safety performance evaluation
(SPE) framework for evaluating contractor’s safety perfor-
mance. The model includes a range of organization-related
and project-related SPE factors. Based on a survey, the
authors assign weights to the different SPE factors to calcu-
late a weighted average safety performance score for each
contractor. Generally, it is well-accepted that weighted
average scores have an inherent weakness due to the biases
additive assumptions utilized in the computations of the
weighted score average.

0263-7863/$36.00 Ó 2009 Elsevier Ltd and IPMA. All rights reserved.

doi:10.1016/j.ijproman.2009.04.002
62 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67
Teo and Ling [37] develop a model to measure the effec-
tiveness of safety management systems (SMS) of construc-
tion sites. The authors utilize surveys and experts
interviews and workshops to collect the important factors
affecting safety. The analytic hierarchy process and factor
analysis are used to identify the most crucial factors and
attributes affecting safety. Using the model, a construction
safety index can be calculated. The authors indicate that
the limitations of their model include the small number
of experts and respondents involved in the study. The
importance weights and attributes are developed within
the context of Singapore. Another limitation is that their
model includes 590 attributes that must be evaluated on
the site.
Despite the limitations associated with some of the
existing methods, they are useful measures of construc-
tion safety performance. However, new methods are still
needed as they offer new insights to both researchers
and practitioners. A point of departure for the DEA
approach compared to existing methods is that DEA
relates resources expended on a certain performance to
the level of success for that particular performance.
Under existing methods, two contractors that suffer the
same numbers and types of accidents are considered of
identical performance. This is clearly not the case if one
contractor is expending more resources (i.e., money,
etc.) on safety than the other contractor. It makes more
sense to consider the contractor that commits fewer
resources to arrive at a certain safety performance as a
better performer.
The rest of the paper unfolds as follows: data envelop-
ment analysis, data collection, results and analysis, future
extension of the research, and conclusions.
2. Data envelopment analysis (DEA)
Data envelopment analysis (DEA) was initiated by
Charnes et al. [3–5]. Since that time, many studies across
different disciplines have utilized DEA [11]. In the con-
struction domain, only few studies made use of DEA. El-
Mashaleh et al. [17] propose the DEA methodology to
measure and compare subcontractors’ productivity at the
firm-level. El-Mashaleh [14] and El-Mashaleh et al.
[15,16] deploy DEA to evaluate the firm-level performance
of construction contractors. Vinter et al. [39] compare pro-
ject efficiency in a multi-project environment based on
DEA. McCabe et al. [29] utilize DEA to prequalify con-
tractors. Pilateris and McCabe [32] evaluate contractors’
financial performance based on DEA.
El-Mashaleh [14] and El-Mashaleh et al. [18] makes use
of DEA to quantify the impact of information technology
on contractors’ performance. Chiang et al. [8] combine
DEA to the I–O tables to examine repercussions of con-
sumptions placed on the construction sector. Cheng et al.
[6] propose a DEA approach for credit scoring to evaluate
borrowers with respect to certain types of projects.
DEA is concerned with evaluations of performance of
organizations (i.e., business firms, hospitals, government
agencies, universities, etc.), where the presence of multi-
ple input, multiple output makes comparison difficult.
In DEA, the organization under study is called a deci-
sion making unit (DMU). A DMU is regarded as the
entity responsible for converting inputs (i.e., resources,
money, etc.) into outputs (i.e., sales, profits, certain per-
formance measures, etc.) and whose performance is to be
evaluated. In this study, a DMU refers to a construction
contractor.
DEA utilizes mathematical linear programming to
determine which of the set of DMUs under study form
an envelopment surface. This envelopment surface is
referred to as the efficient frontier. DEA provides a com-
prehensive analysis of relative efficiency for multiple
input-multiple output situations by evaluating each DMU
and measuring its performance relative to this envelopment
surface. Units that lie on (determine) the surface is deemed
efficient in DEA terminology. Units that do not lie on the
surface are termed inefficient and the analysis provides a
measure of their relative efficiency.
Cooper et al. [11] and Coelli et al. [9] argue that DEA
has gained its popularity from three inherent powerful fea-
tures. First, its capability to incorporate multiple inputs
and multiple outputs as a result of the use of linear pro-
gramming. Linear programming can handle large numbers
of variables and relations (constraints). Second, DEA has
no priori assumptions. There is no need to assign weights
to the different inputs and outputs. The weights are derived
directly from the data relaxing the user from arbitrary sub-
jective weighting. DEA provides a set of weights, which
optimize a unit’s performance subject to the weights not
leading to any other unit violating the bounds of the fron-
tier. Third, the measurement units of the different inputs
and outputs need not be congruent. Some may involve
number of persons, or areas of floor space, money
expended, etc.
2.1. Charnes–Cooper–Rhodes (CCR) DEA model
This study makes use of the Charnes–Cooper–Rhodes
(CCR) model of DEA to benchmark safety performance
of construction contractors. The mathematical form is
shown below (Eqs. (1),(2),(3),(4),(5). Interested readers
may refer to Cooper et al. [11] for details.
X s
max h0 ¼ ur y r0 ð1Þ
r¼1
Xm
subject to vi xi0 ¼ 1 ð2Þ
i¼1
X s X
m
ur y rj  vi xij 6 1 ð3Þ
r¼1 i¼1
i ¼ 1; . . . ; m; j ¼ 1; . . . ; n; r ¼ 1; . . . ; s ð4Þ
ur ; vi P 0 ð5Þ
 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67
Where:
h0 the measure of efficiency for DMU0 (the DMU under
evaluation), which is a member of the set j = 1, . . ., n
DMUs
ur the output weight. It is determined by the solution of
the model and is assigned to the observed rth output
vi the input weight. It is determined by the solution of
the model and is assigned to the observed ith input
yr0 the known amount of the rth output produced by
DMU0
xi0 the known amount of the ith input used by DMU0
yrj the known amount of the rth output produced by
DMUj
xij the known amount of the ith input used by DMUj
The objective function is to maximize the efficiency of
DMU0 (the DMU under evaluation). This is done by max-
imizing the sum of its outputs (Eq. (1)), while forcing the
sum of its inputs to be equal to 1.0 (Eq. (2)). Eq. (3) means
that the efficiency of all DMUs is 61.0. This implies that all
DMUs are either on the efficient frontier or below it, and
that the efficiency scores range between 0 and 1.0.
Briefly stated, the CCR model of DEA will be used to
benchmark safety performance of construction contrac-
tors. The model yields efficiency scores that range between
0 and 1.0. A contractor is considered efficient if it has an
efficiency score of 1.0. This means that this contractor is
on the efficient frontier. Compared to the rest of contrac-
tors, this particular contractor effectively converts its
inputs into outputs.
2.2. Inputs and outputs selection
As mentioned earlier, DEA considers a DMU as the
entity responsible for converting inputs (i.e., resources,
money, etc.) into outputs (i.e., sales, profits, certain perfor-
mance measures, etc.). In the context of construction safety
performance, Fig. 1 shows the related input and outputs. In
terms of input, safety performance of a construction con-
tractor is impacted by the contractor’s expenses on safety
as a percentage of total revenues. These expenses include
contractor’s annual cost of safety programs and salaries
of safety personnel.
As for the outputs, a contractor’s safety performance is
measured by the numbers of the different types of accidents
that are suffered by that construction contractor. Since this
study is implemented in the Jordanian construction indus-
Inputs
Expenses on safety as a % of
total revenues
Fig. 1. Input and outputs of a construction contractor safety performance.
63
try, the Jordanian Social Security Organization (SSO) clas-
sification of accidents is adopted. According to SSO, work
accidents are classified into 5 types as follows:
 Type 1: accidents that do not cause any disability and do
not involve any lost work days.
 Type 2: accidents that do not cause any disability but
involve lost work days.
 Type 3: accidents that cause temporary disability.
 Type 4: accidents that cause permanent partial disability.
 Type 5: accidents that cause permanent full disability or
fatality.
Since the numbers of accidents are unfavorable, the
reciprocals of the numbers of the different types of acci-
dents that are defined by SSO are used as outputs.
3. Data collection
Potential participants in the study were selected at ran-
dom from member lists of the Jordanian Contractors Asso-
ciation. These potential participants were assured that their
identities and their companies’ identities will be kept confi-
dential. Additionally, the researchers pointed out that the
responses will be coded and kept separate from the names
of the participants and their companies’ names. It was
explained that when the study is completed and the data
is analyzed, the codes will be destroyed. Potential partici-
pants were informed that the results of the research study
will be published, but the identities of the participants
and their companies’ names will not be used. It was
emphasized that the anticipated benefit for respondents’
participation is benchmarking their firms’ safety perfor-
mance compared to other firms in the construction indus-
try. Potential participants were offered to receive a free
hard copy of the research report after it has been published
if they choose so.
Seventy contractors, out of 164 contractors that were
contacted for potential participation, agreed to participate
in the research project. However, only 45 contractors sup-
plied the required data to be included in the DEA model.
For a particular fiscal year, contractors were asked to
report the following:
 Expenses on safety as a% of total revenues.
 Number of the 5 types of accidents (as classified by SSO)
that are suffered by the contractor.
DMU Outputs
1/Number of Type 1 accidents
1/Number of Type 2 accidents
Construction
1/Number of Type 3 accidents
contractor
1/Number of Type 4 accidents
1/Number of Type 5 accidents
64 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67

Table 1
Descriptive statistics for the collected data.
Mean Standard deviation Minimum Maximum
Expenses on safety as a% of total revenues 0.0768 0.0989 0.01 0.5
Number of Type 1 accidents 1.76 1.33 1 6
Number of Type 2 accidents 1.84 1.35 1 6
Number of Type 3 accidents 1.18 0.53 1 3
Number of Type 4 accidents 1.22 0.6 1 4
Number of Type 5 accidents 1.22 0.56 1 3

Table 1 shows descriptive statistics for the collected Table 3

data. Efficiency scores and ranks of construction contractors.
No. DMU Score Rank
4. Results and analysis 1 A 0.20 16
2 E 1.00 1
The DEA-Solver software of Cooper et al. [11] is used to 3 G 0.20 16
run the CCR model. Table 2 summarizes the descriptive 4 H 0.20 16
5 I 0.11 30
statistics of the results. Out of 45 contractors, only 8 are 6 J 0.50 10
efficient in their safety performance. The maximum effi- 7 K 0.05 42
ciency score is 1.0, while the minimum efficiency score is 8 M 0.20 16
0.02. The efficiency scores average is 0.32. This means that 9 N 0.08 38
the input for an average unit could be reduced by 68%. 10 O 0.10 33
11 P 0.07 41
Table 3 shows the efficiency scores and the ranks of the 12 S 0.13 28
45 contractors. The results show how each contractor per- 13 U 0.29 14
formed in comparison with the rest of the contractors. The 14 V 0.20 16
following 8 contractors are deemed efficient and are consid- 15 W 0.10 33
ered to have superior safety performance: E, CC, RR, SS, 16 X 0.08 38
17 CC 1.00 1
TT, ZZ, AAA, and CCC. These efficient contractors have 18 DD 0.33 12
an efficiency score = 1.0. They are on the efficient frontier. 19 EE 0.33 12
Compared to the rest of the contractors, these 8 contrac- 20 FF 0.17 23
tors are more efficient in converting the money they spent 21 GG 0.20 16
on safety into less number of accidents. 22 HH 0.17 23
23 JJ 0.13 28
Table 3 shows that the efficiency scores went down from 24 LL 0.14 27
1.0 for efficient contractors to 0.59 (Contractor PP) then to 25 NN 0.1 33
0.5 (Contractor J and Contractor OO). We need to note 26 OO 0.5 10
here that the efficiency values are valid within this particu- 27 PP 0.59 9
lar group of contractors. The efficiency scores will vary 28 QQ 0.25 15
29 RR 1 1
depending on the contractors that are included in the anal- 30 SS 1 1
ysis. As mentioned before, the best performers create an 31 TT 1 1
envelopment surface and every DMU’s performance is 32 UU 0.09 37
measured against this envelopment surface. As such, the 33 VV 0.08 38
gaps in scores depend on the specific performance numbers 34 YY 0.17 23
35 ZZ 1 1
for the DMU in question. To explain the gaps in efficiency 36 AAA 1 1
scores, assume that we added a new contractor that has 37 CCC 1 1
better performance than Contractor G (efficiency 38 FFF 0.02 44
score = 0.59), but yet is not as efficient as the industry lead- 39 HHH 0.04 43
ers. So, in this case, the efficiency score of this added new 40 III 0.02 44
41 JJJ 0.1 33
42 LLL 0.11 30
Table 2 43 OOO 0.11 30
Descriptive statistics for DEA results. 44 PPP 0.17 23
45 RRR 0.2 16
Total number of DMUs 45
Number of efficient DMUs 8
Number of inefficient DMUs 37
Scores average 0.32 contractor will fall within the following range: 0.59 < effi-
Scores standard deviation 0.34 ciency score < 1.0.
Maximum score 1.0 Table 4 compares averages of the input and outputs
Minimum score 0.02 between efficient and inefficient contractors. Note that
 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67 65

Table 4 money) into lower number of accidents. Contractor G

Comparison of input and outputs averages between efficient and inefficient
contractors.
Expenses on safety as a percentage of
total revenues
Number of Type 1 accidents
Number of Type 2 accidents
Number of Type 3 accidents
Number of Type 4 accidents
Number of Type 5 accidents
should spend less on safety and yet enjoy fewer numbers
of accidents. Clearly, a decision to commit fewer resources
Efficient Inefficient to safety requires substantial consideration by the organi-
contractors contractors
zation. Having an informed result regarding its expenses
0.01 0.09 on safety, the organization needs to examine the specifics
1 1.85
of its safety expenses (i.e., training, salaries, etc.) and iden-
1 1.95 tify cost minimization opportunities without negatively
1 1.2 impacting safety performance of the organization.
1.2 1.23
1 1.25 5. Future extension of the research

The next step for the research team is to attempt to asso-

efficient contractors spent less than their inefficient counter-
parts on safety as a percentage of total revenues. The aver-
age for the first group is 0.01, while that average for the
second group is 0.09. Similarly, efficient contractors suf-
fered less number of accidents compared to the inefficient
contractors.
A major motivation behind measuring performance is to
identify opportunities for possible efficiency improvements
by looking at the differences between efficient firms and
inefficient ones. To realize their potentials, the inefficient
contractors need to compare themselves with the best prac-
tice contractors that ‘‘make-up” the efficient frontier. DEA
analysis provides quantitative guidance for inefficient con-
tractors to be recognized as efficient frontier contractors.
As an example, lets take Contractor G with efficiency
score = 0.2. Table 5 shows both current and projected val-
ues of the input and outputs of Contractor G. To be recog-
nized on the efficient frontier, Contractor G needs to
reduce its input and three of its outputs. In particular,
expenses on safety have to be reduced from 0.05 to 0.01.
Additionally, number of Type 1, Type 2, and Type 5 acci-
dents need to be reduced to 1 down from their current
values.
Note that the above recommendation to decrease
expenses on safety for contractor G should be considered
from a performance point of view. As demonstrated in
the analysis, superior safety performance contractors spent,
on average, 0.01% of their total revenues on safety. Con-
tractor G spent 0.05% of its total revenues on safety, while
suffering higher number of accidents compared to superior
safety performance contractors. The recommendation to
reduce expenses on safety is an indication that Contractor
G is not effectively converting its committed resources (i.e.,
Table 5
Contractor G with efficiency score = 0.2.
Input and outputs Current Projected
values values
Expenses on safety as a percentage of total 0.05 0.01
revenues
Number of Type 1 accidents 6 1
Number of Type 2 accidents 6 1
Number of Type 3 accidents 1 1
Number of Type 4 accidents 1 1
Number of Type 5 accidents 2 1
ciate variability in safety performance with certain explan-
atory variables. Safety research within the construction
domain identifies several factors that drive safety perfor-
mance. Table 6 lists these factors and summarizes key
statements made by previous research. The research team
plans to broaden the size and scope of data collection to
statistically test 8 research hypotheses that are shown in
Table 7. Analysis of Variance (ANOVA), for example, will
examine the existence (or non-existence) of statistical sig-
nificance differences between efficient and inefficient con-
tractors regarding the 8 factors that drive safety
performance.
6. Conclusions
This paper contributes a DEA approach for benchmark-
ing safety performance of construction contractors. A
point of departure for the DEA approach compared to
existing methods is the input–output framework. Com-
pared to each other, DEA measures the efficiency of con-
struction contractors in utilizing their expenses on safety
to minimize the number of suffered accidents. Therefore,
the DEA approach relates resources expended on safety
to safety performance.
DEA analysis scores safety performance of construction
contractors on a scale of 0–1.0. The analysis identifies con-
tractors E, CC, RR, SS, TT, ZZ, AAA, and CCC as effi-
cient frontier contractors. Compared to the rest of the
contractors, these eight contractors are the industry leaders
in safety performance. They serve as the ‘‘benchmark” for
the industry and can be utilized as role models to which
inefficient contractors may adjust their practices in order
to become efficient.
An excellent utilization for the results of this study is
programs like the US Malcolm Baldrige National Quality
Award [35] and the UK Department of Trade and Industry
Business-to-Business Exchange Program [30]. These pro-
grams aim at improving the performance of particular
industries. For example, the UK Department of Trade
and Industry Business-to-Business Exchange Program
offers visits to UK best practicing organizations in manu-
facturing and service industries. The goal of these visits is
to transfer best practices across interested organizations
for the purpose of improving their performance.
66 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67

Table 6
Factors that drive safety performance.
Number Factor
1 Organizational
safety policy
2 Safety training
3 Safety meetings
4 Safety equipment
5 Safety
inspections
6 Safety incentives
and penalties
7 Workers’
attitude towards
safety
8 Labor turnover
rates
Research Key statements
Sawacha et al. [34] Organizational safety policy is the most influential factor that drives safety
performance in the construction industry
Wong et al. [40] Written safety policies are essential to construction safety
Hinze and Wilson [25] Study participants recommend a safety system of checks and balances to
improve safety performance
Jaselisks et al. [27] Better safety performance involves the development of more detailed written
safety programs
Construction Industry Institute (CII) [10] Safety training is one of five high-impact zero accident techniques
Hinze and Wilson [25] Worker training is vital to improved safety performance
Huang and Hinze [26] The lack of safety training is often a contributing factor to many falls
Jaselisks et al. [27] To improve safety performance at the project level, the authors recommend
increasing the number of formal safety meetings with supervisors
Toole [38] Some construction accidents result because safety equipment necessary to
perform the job safely is not present at the location of the work
Duncan and Bennett [13] Both active measures (those that prevent workers from falling, for example,
guardrails) and passive measures (those that protect workers after falling,
for example, safety nets) are useful in reducing fall injuries
Chi et al. [7] Falls are associated with lack of complying scaffolds, unguarded openings,
inappropriate protections, removal of protections, and improper use of
personal protective equipment (PPE)
Wong et al. [40] Safety performance is affected by monitoring of safety compliance
Jaselisks et al. [27] It is recommended to increase site safety inspections
Hinze and Wilson [25] More jobsite visits are needed to improve safety performance
CII [10] Safety incentives are among the top five high-impact zero accident
techniques
Jaselisks et al. [27] There is a need to increase fines to workers with poor safety performance
Tam and Fung [36] Safety incentives should be utilized to improve safety performance
Abdelahamid and Everett [1], Aksorn and Workers’ attitude towards safety is one of the root causes of accidents
Hadikusumo [2], Hinze [21], Toole [38]
Harper and Koehn [20], Hinze and Higher turnover rates are associated with higher injury rates
Gambatese [24]
Hinze [22] New hires are more subjected to accidents

Table 7 order to publicize their strategies and procedures for the

Future research hypotheses.
Number Hypothesis
1 Safety performance and organizational safety policy are
positively correlated
2 Safety performance and safety training are positively correlated
3 Safety performance and safety meetings are positively
correlated
4 Safety performance and safety equipment are positively
correlated
5 Safety performance and safety inspections are positively
correlated
6 Safety performance and safety incentives and penalties are
positively correlated
7 Safety performance and workers’ attitude towards safety are
positively correlated
8 Safety performance and low labor turn over rates are positively
correlated
Therefore, the deployment of the DEA methodology
makes it possible for programs like the ones described
above to utilize the results and target industry leaders in
benefit of the whole industry.
The Jordanian construction industry currently lacks any
readily available safety performance measure to assess
safety performance of construction contractors. To judge
safety performance, the industry currently relies on the seg-
regated reported numbers of the different types of acci-
dents. As such, the DEA approach is well suited to fill
this gap and assesses contractors’ safety performance.
The DEA approach presented in this paper can be uti-
lized by a particular contractor to gauge its own safety per-
formance over time. With data available for several
numbers of years, every year might be considered as a sin-
gle DMU. By conducting such analysis a contractor would
be able to quantitatively determine whether or not the
safety performance of the firm is getting better over time.
Additionally, the proposed methodology is deployable
at the project level. Every project is regarded as a single
DMU and projects as a result are ‘‘benchmarked” against
each other. Consequently, contractors will be able to iden-
tify their best performing projects and to isolate internal
factors that contributed to better performance.
 M.S. El-Mashaleh et al. / International Journal of Project Management 28 (2010) 61–67
Even though the development in this paper is based on
data collected from the Jordanian construction industry,
the methodology would suggest a much broader geograph-
ical applicability on evaluating safety performance for con-
struction projects internationally as well as other projects
in other disciplines like manufacturing projects.
The next step for the research team is to attempt to asso-
ciate variability in safety performance with certain explan-
atory variables (i.e., organizational safety policy, safety
training, safety equipment, etc.). Clearly, to test such
hypotheses, both a larger sample size and a larger scope
of data collection are required.
Acknowledgments
The authors acknowledge two anonymous referrers for
their constructive comments that significantly improved
the final presentation of the paper.
References
[1] Abdelahamid T, Everett J. Identifying root causes of construction
accidents. J Constr Eng Manag 2000;126(1):52–60.
[2] Aksorn T, Hadikusumo B. Critical success factors influencing safety
program performance in Thai construction projects. Saf Sci
2008;46:709–27.
[3] Charnes A, Cooper W, Rhodes E. Measuring the efficiency of
decision making units. Eur J Oper Res 1978;2(6):429–44.
[4] Charnes A, Cooper W, Rhodes E. Short communication: measuring
the efficiency of decision making units. Eur J Oper Res 1979;3:339.
[5] Charnes A, Cooper W, Rhodes E. Evaluating program and mana-
gerial efficiency: an application of data envelopment analysis to
program follow through. Manag Sci 1981;27:668–97.
[6] Cheng E, Chaing Y, Tang B. Alternative approach to credit scoring
by DEA: evaluating borrowers with respect to PFI projects. Build
Environ 2007;42:1752–60.
[7] Chi C, Chang T, Ting H. Accident patterns and prevention measures
for fatal occupational falls in the construction industry. Appl Ergon
2005;36:391–400.
[8] Chiang Y, Cheng E, Tang B. Examining repercussions of consump-
tions and inputs placed on the construction sector by use of I–O
tables and DEA. Build Environ 2006;41:1–11.
[9] Coelli T, Rao D, Battese G. An introduction to efficiency and
productivity analysis. Boston/Dordrecht/London: Kluwer Academic
Publishers; 1998.
[10] Construction Industry Institute (CII). Zero accident techniques.
Source document 86. Austin, Texas: Construction Industry Institute;
1993.
[11] Cooper W, Seiford L, Tone K. Data envelopment analysis: a
comprehensive text with models, applications, references and DEA-
solver software. Boston/Dordrecht/London: Kluwer Academic Pub-
lishers; 2000.
[12] De la Garza M, Hancher D, Decker L. Analysis of safety indicators in
construction. J Constr Eng Manag 1998;124(4):312–4.
[13] Duncan C, Bennett R. Fall protection and debris containment during
construction. In: Chang LM, editor. Preparing for construction in the
21st century. New York: ASCE; 1991. p. 97–102.
[14] El-Mashaleh M. Firm performance and information technology
utilization in the construction industry: an empirical study. Ph.D.
dissertation. Gainesville, Florida: University of Florida; 2003.
67
[15] El-Mashaleh M, Minchin R, O’Brien W. Benchmarking construction
firm performance. In: Proceedings of the 11th joint CIB international
symposium, 13–16 June, Helsinki, Finland; 2005. p. 376–89.
[16] El-Mashaleh M, Minchin R, O’Brien W. Management of construc-
tion firm performance using benchmarking. J Manag Eng
2007;23(1):10–7.
[17] El-Mashaleh M, O’Brien W, London K. An envelopment method-
ology to measure and compare subcontractor productivity at the firm
level. In: Proceedings of the 9th annual conference of the interna-
tional group for lean construction, 6–8 August, 2001. Singapore:
National University of Singapore; 2001. p. 17.
[18] El-Mashaleh M, O’Brien W, Minchin R. Firm performance and
information technology utilization in the construction industry. J
Constr Eng Manag 2006;132(5):499–507.
[19] Everett J, Thompson W. Experience modification rating for workers’
compensation insurance. J Constr Eng Manag 1995;121(1):66–79.
[20] Harper R, Koehn E. Managing industrial construction safety in
southeast Texas. J Constr Eng Manag 1998;124(6):452–7.
[21] Hinze J. Human aspects of construction safety. J Constr Div
1981;107(1):61–72.
[22] Hinze J. Construction safety. New Jersey: Prentice-Hall; 1997.
[23] Hinze J, Bren D, Piepho N. Experience modification rating as a
measure of safety performance. J Constr Eng Manag
1995;121(4):455–8.
[24] Hinze J, Gambatese J. Factors that influence safety performance of
specialty contractors. J Constr Eng Manag 2003;129(12):159–64.
[25] Hinze J, Wilson G. Moving toward a zero injury objective. J Constr
Eng Manag 1999;126(5):399–403.
[26] Huang X, Hinze J. Analysis of construction worker fall accidents. J
Constr Eng Manag 2003;129(3):262–71.
[27] Jaselisks E, Anderson S, Russell J. Strategies for achieving excellence
in construction safety performance. J Constr Eng Manag
1996;121(1):61–70.
[28] Levitt R, Samelson N. Construction safety performance. New York:
McGraw-Hill; 1987.
[29] McCabe B, Tran V, Ramani J. Construction prequalification using
data envelopment analysis. Can J Civil Eng 2005;32:183–93.
[30] McGeorge D, Palmer A. Construction management: new directions.
Oxford: Blackwell Science Ltd; 1997.
[31] Ng S, Cheng K, Skitmore R. A framework for evaluating the safety
performance of construction contractors. Build Environ
2005;40:1347–55.
[32] Pilateris P, McCabe B. Contractor financial evaluation model
(CFEM). Can J Civil Eng 2003;30(3):487–99.
[33] Samelson N, Levitt R. Owner’s guidelines for selecting safe contrac-
tors. J Constr Div 1982;108(4):617–23.
[34] Sawacha E, Naoum S, Fong D. Factors affecting performance on
construction sites. Int J Proj Manag 1999;17(5):309–15.
[35] Spendolini M. The benchmarking book. AMACOM: American
Management Association; 1992.
[36] Tam C, Fung I. Effectiveness of safety management strategies on
safety performance in Hong Kong. Constr Manag Econ
1998;16(1):49–55.
[37] Teo E, Ling F. Developing a model to measure the effectiveness of
safety management systems of construction site. Build Environ
2006;41:1586–92.
[38] Toole M. Construction site safety roles. J Constr Eng Manag
2002;128(3):203–10.
[39] Vinter G, Rozenes S, Spragget S. Using data envelop analysis to
compare project efficiency in a multi-project environment. Int J Proj
Manag 2006;24:323–9.
[40] Wong K, Chan P, Lo K. Factors affecting the safety performance of
contractors and construction sites. In: Proceedings of the second
international conference of CIB working commission W99, Honolulu,
Hawaii; 1999.