Transportation Letters

The International Journal of Transportation Research

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ytrl20

Identifying strategies for improving airport

services: introduction of the Gap-IPA to an Italian
airport case study

Jaime Allen , Maria Grazia Bellizzi , Laura Eboli , Carmen Forciniti & Gabriella
Mazzulla

To cite this article: Jaime Allen , Maria Grazia Bellizzi , Laura Eboli , Carmen Forciniti & Gabriella
Mazzulla (2020): Identifying strategies for improving airport services: introduction of the Gap-IPA to
an Italian airport case study, Transportation Letters, DOI: 10.1080/19427867.2020.1861506

To link to this article: https://doi.org/10.1080/19427867.2020.1861506

Published online: 11 Dec 2020.

Submit your article to this journal

Article views: 7

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ytrl20
TRANSPORTATION LETTERS
https://doi.org/10.1080/19427867.2020.1861506

Identifying strategies for improving airport services: introduction of the Gap-IPA to an

Italian airport case study
aa bb
Jaime Allen , Maria Grazia Bellizzi , Laura Ebolibb, Carmen Forcinitibb and Gabriella Mazzulla bb

a
National Laboratory of Materials and Structural Models (Lanammeucr), Universidad De Costa Rica, San José, Costa Rica; bDepartment of Civil
Engineering, University of Calabria, Rende, Italy

ABSTRACT KEYWORDS
An effective strategy for offering services characterized by adequate levels of quality is to investigate on the Air transport; airport
single service aspects, to verify what needs to be improved or maintained, and what does not need particular services; Gap-IPA; IPA; SEM;
resources. In this paper, we want to provide a practical instrument for investigating airport transport service service quality
quality, based on the opinion expressed by a sample of passengers. The case study is the International
Airport of Lamezia Terme, a peripheral airport placed in the south of Italy; the data were derived from
Customer Satisfaction Surveys conducted during 2015–2016. We propose a modification of the well-known
Importance-Performance Analysis (IPA), by adopting as importance the weights determined through
a Structural Equation Modeling (SEM) approach, and by plotting the gap between importance and perfor­
mance on a visual graph, that we hold as easier to be read and more intuitive than the one used in the
traditional IPA. The results obtained by the proposed Gap Importance-Performance Analysis (Gap-IPA)
suggest that the attention of the airport operator should focus on the service aspects related to cleanliness.
Besides, the methodology can show which aspects should be kept under control for avoiding to become
critical issues of the service.

Introduction a certain number of features characterizing the service according to

a measurement scale. In the field of road and rail public transport,
People choosing to travel by public transport modes make use of
many studies dealt with the assessment of service quality based on
services provided during the time spent on board but also services
users’ perceptions (see, for example, de Oña et al. 2016; Allen et al.
offered at the public transport stops or stations where they take the
2020a; Eboli, Forciniti, and Mazzulla 2018b; Allen et al. 2019;
transport mean. Air transport can be considered as quite different
Bellizzi et al. 2020). Only in recent years, this topic has become of
from public transport systems. People traveling by air are con­
interest also in the field of air transport, due to the exponential
strained to arrive at the airport at least 40 minutes before the flight
increase in the number of passengers traveling by air (IATA,
for a series of reasons. In fact, within the terminal area, services
International Air Transport Association 2018). Assessing airport
such as check-in, passport and security controls, baggage drop,
service quality is fundamental both for users and airport manage­
customs, and baggage claim are provided to departing and arriving
ment companies. Improving the level of quality of the provided
travelers (Ashford, Stanton, and Moore 1997). Therefore, an airport
service is essential because travelers would undoubtedly be
has a prominent role in the complete travel experience as compared
delighted from a comfortable and well-functioning airport.
to a railway station or a bus stop, which are places where people stay
Among the various techniques developed for investigating on
for a relatively short time. Airport passenger terminals are impor­
transit service quality, there is the Importance-Performance
tant processing nodes where the continuous flow of air travelers
Analysis (IPA) (Martilla and James 1977). Numerous practitioners
arriving from the airport catchment area is transformed into dis­
and researchers have applied IPA to identify the critical perfor­
crete departures, i.e. the scheduled flights (Postorino et al. 2019).
mance factors in customer satisfaction survey data for products and
For this reason, providing airport services characterized by high
services (Deng 2007). Hansen and Bush (1999) pointed out that IPA
levels of quality is very important to make the travel more pleasant
is a simple and effective technique that can assist practitioners in
for the passengers, with the final objective to attract more users. As
identifying improvement priorities for customer attributes and
an example, in an airport such as our case study, characterized
direct quality-based marketing strategies. IPA is applied based on
predominantly by a traffic of national flights, it is imperative to
two dimensions of customer attributes: performance level (satisfac­
attract users to the detriment of railway or bus services, which are
tion) and importance. By combining the two dimensions, a graph
often chosen as alternative modes to reach the various national
with four quadrants is determined, where each quadrant defines
destinations, from the South of Italy to the Center or the North,
a certain level of priority of the service attributes falling within it.
and vice-versa. Moreover, an adequate level of service quality makes
Starting from the study of the literature about IPA and the applica­
the airport more attractive and can contribute to the development
tion of IPA to the various public transport modes, we propose
of the surrounding territorial context.
revisiting the technique, with the final aim to provide an even
In order to interpret and understand users’ desires and needs,
more practical instrument for public transport operators. Our
their perceptions about the service should be known. Users’ percep­
challenge is to propose a technique and a graphical representation
tions are generally collected through the Customer Satisfaction
that is more intuitive and immediate. It will suggest to the operator,
Surveys (CSS), where the interviewees express their judgments on

CONTACT Maria Grazia Bellizzi mariagrazia.bellizzi@unical.it Department of Civil Engineering, University of Calabria, P. Bucci, Cubo 46/B, Rende 87036, Italy
© 2020 Informa UK Limited, trading as Taylor & Francis Group
2 J. ALLEN ET AL.
in a quick look at the graph, if the various service attributes are to be
considered as critical aspects, where the financial resources have to
be concentrated as performance does not reach users’ expectations
(representing the importance). Alternatively, if they have to be
considered as aspects that operators could disregard or consider
as not priority one because performance exceeds importance. More
specifically, we decided to consider the difference (or gap) between
importance and performance and to represent this gap in a graph
where the various service attributes are placed in two different
sectors depending on the value of the gap. We decided to name
the proposed technique as Gap-IPA. Definitively, we have only two
sectors to be observed against well four quadrants to be considered
for evaluating the priority of the attributes. We retain that just for
this reason the new method is more intuitive and immediate for
interpreting the results, without losing fundamental information.
Moreover, an important key is that the Gap-IPA brings the two-
dimensional matrix into a single indicator (i.e. the gap between
importance and performance) that allows the administration to
prioritize and also assess which indicators are most critical and
which ones are close to becoming critical. The same information
could not be so incisive according to IPA representation, because
the attributes included in the quadrant retained as less important
(i.e. low priority and possible overkill) could be overlooked and
become critical. In other words, having a single-metric is more
intuitive and may appeal more to the administrations.
As we applied the technique by adopting the data collected on
a sample of air passengers expressing judgments only on the per­
formance of the service, we opted to calculate importance through
a Structural Equation Modeling (SEM) approach because it can
consider the relationship between the service attributes, the overall
satisfaction, and latent constructs (Eboli, Forciniti, and Mazzulla
2018a).
We deem that the literature on this topic is still in its infancy,
especially concerning the use of the SEM approach applied to air­
port services. A Principal Component Analysis (PCA) was per­
formed as an exploratory approach for establishing the latent
variables of the SEM and the various relationships.
The rest of the paper is structured as follows. In the following
section, we report a literature review particularly focusing on the
works that adopted IPA for investigating on transit service quality.
Successively, we describe the proposed method: the Gap-IPA, and
the SEM approach used for determining the importance. Section 4
presents the case study, a small-sized airport (about two million
passengers per year) placed in the south of Italy, the survey, the
questionnaire designed for collecting the data, and the sample
characteristics. Finally, we present the results, and we propose
a discussion of the findings and the practical implications.
Literature review
Understanding which factors mainly affect overall passengers’ satis­
faction could help airport management companies to achieve better
financial resource administration.
In the literature, there is a series of papers dealing with the
assessment of airport service quality to identify the service attri­
butes mostly influencing airport service quality, representing the
aspects on which a company should focus the efforts for improving
the service and satisfying the users. In most of them, it appears that
there are several attributes to be taken into account for the analysis
of the airport service quality. According to this, over the years,
researchers have always tried to use methodologies capable of
synthesizing the phenomenon as much as possible. From the
exploratory factor analysis (EFA) conducted by Bezerra and
Gomes (2015), it emerged that the airport service quality can be
explained by seven factors named as check-in, security, conveni­
ence, ambience, basic facilities, prices, earliness of arrival and travel
frequency. Through a regression analysis, the authors found that
the dimension with the highest effect on airport service quality is
related to comfort and cleanliness inside the terminal (represented
by the factor ‘ambience’), and to prices at food facilities and stores.
Further analyses have been carried out by Bezerra and Gomes
(2016) in order to test the differences in the perceptions of domestic
and international passengers. By performing a PCA, Brida,
Moreno-Izquierdo, and Zapata-Aguirre (2016) obtained five differ­
ent components (image perception, airport information, terminal
servicescape, airport sound information system, and flight informa­
tion screen), and from the results of a Logit model, they concluded
that airports should improve mainly the way of communicating
flight information and the location of different airport’s utilities.
Pantouvakis and Renzi (2016) used a Rasch Modeling technique
and concluded that the service quality provided in an airport multi-
national context can be better described by three distinct, indepen­
dent, and invariant dimensions, namely servicescape and image,
signage and service. Moreover, the authors stated that the presence
of the dimension of servicescape and image seems to play the most
decisive role in satisfying passengers. In Ceccato and Masci (2017)
the focus is on passengers’ satisfaction with safety inside the airport,
and from the results of a binary logistic regression, it emerges that
the airport’s environment (e.g. cleanliness, maintenance, informa­
tion facilities, signage, acoustic and thermal comfort) plays an
important role on the perceived safety.
Other studies proved that the SEM approach could adequately
account for latent constructs affecting overall air service quality,
and to explore observed indicators for measuring the introduced
latent constructs themselves. Specifically, in Park and Jung (2011)
the SEM approach has been adopted to test the relationships
between the airport service quality, value, satisfaction, airport
image, and passenger behavior, by taking into account the percep­
tion of only transfer passengers. From the results, it emerges that
airport service quality has a significant positive effect on value,
satisfaction, and airport image. Moreover, they concluded that the
airport service quality influences transfer passengers’ reuse inten­
tions. The results of the study carried out by Nesset and Helgesen
(2014) presented the airport service quality as the most important
driver for the loyalty attitude, passengers’ satisfaction creation, and
airport’s image building. The study of Prentice and Kadan (2019)
examines through an SEM the relationship between airport service
quality, passengers’ satisfaction, and behavioral intentions includ­
ing airport reuse and destination revisit. In Allen et al. (2020b), the
authors proposed an SEM-MIMIC ordinal Probit for capturing the
heterogeneity in passengers’ perceptions and for identifying groups
of passengers with similar assessments of the services. Otherwise,
Hong, Choi, and Chae (2020) proposed an SEM by comparing the
results obtained for air travelers and service providers as two
different groups of airport users.
All these methodologies, albeit useful for identifying the most
influential service’s aspects and for analyzing the passengers’ het­
erogeneity, do not provide practical information about which pro­
vided service require prompt action. A specific technique that
represents a practical instrument for reaching such an objective is
surely the Importance-Performance Analysis (IPA), proposed initi­
ally by Martilla and James (1977). IPA identifies which product or
service attributes a firm should focus on to enhance customer
satisfaction (Matzler et al. 2004). Many other researchers have
subsequently revisited IPA. Some modified versions of the IPA
consisted of the proposal of calculating importance starting from
the satisfaction (or performance) ratings expressed by the users
about the various service attributes and the overall service. IPA

was applied for assessing the quality of different services of public
transport. In the field of airport service quality, these techniques
were proposed by Lubbe, Douglas, and Zambellis (2011), Babu
(2015), Jiang and Zhang (2016), Pandey (2016) and Tseng (2020).
In almost all the mentioned studies, performance and importance
represented the ratings provided directly by the passengers.
Specifically, in Lubbe, Douglas, and Zambellis (2011) reports the
importance and performance findings with respect to purpose of
travel and to frequency of travel, in order to show the differences
between business and leisure travelers and between frequent and
infrequent travelers. Babu (2015) conducted a similar study but
with the aim to identify the differences among the ages’ groups of
passengers. In the work of Pandey (2016), a fuzzy analysis was
performed for deriving both performance and importance ratings.
A modified version of the traditional IPA was proposed by Tseng
(2020) to classify and diagnose the service attributes of an airport:
the IPA-Kano model. Both Jiang and Zhang (2016) and Tseng
(2020) added to their study a GAP analysis, but none of them
proposed a graphical representation of their results.
Methodology
As mentioned in the introduction, IPA is applied based on two
dimensions of customer attributes: performance level (satisfaction)
and importance. Data from customer satisfaction surveys are typi­
cally used to construct a two-dimensional matrix, where attribute
performance (satisfaction) is depicted along the x-axis, and attri­
bute importance (satisfaction) is depicted along the y-axis.
Attribute importance is measured using some forms of self-stated
importance (e.g. rating scales) or implicitly derived importance (e.g.
multiple regression weights, partial correlation weights). Generally,
the means of performance and importance divide the matrix into
four quadrants (Figure 1). Quadrant I, where both performance and
importance are high, contains the attributes that can be considered
as significant strengths; they represent opportunities for achieving
or maintaining competitive advantage. The management scheme
for this quadrant is ‘keep up the good work.’ Quadrant II, where
performance is low and importance is high, contains the major
Figure 1. IPA.
TRANSPORTATION LETTERS 3
weaknesses, which require immediate attention for improvement.
The management scheme for this quadrant is ‘concentrate here.’
Quadrant III, where performance and importance are low, contains
the minor weaknesses, which do not require additional effort. The
management scheme for this quadrant is ‘low priority.’ Finally,
Quadrant IV, where performance is high and importance is low,
contains the attributes that can be considered as minor strengths,
indicating that business resources committed to these attributes
would be overkill and should be deployed elsewhere. The manage­
ment scheme for this quadrant is ‘possible overkill.’ Researchers
commonly suggest that major weaknesses (Quadrant II) should be
a top priority and targeted for immediate improvement efforts
(Martilla and James 1977). On the other hand, major strengths
(Quadrant I) should be maintained and heavily promoted
(Lambert and Sharma 1990).
Some studies have modified and extended IPA; however, the
basic framework has primarily remained the same (Sampson and
Showalter 1999). We propose the Gap-IPA, an alternative repre­
sentation of the two dimensions, performance and importance, by
using the concept expressed in the SERVQUAL model where ser­
vice quality is defined and calculated as the gap between customer’s
expectations and perceptions (Parasuraman, Berry, and Zenithal
1991). Following the theory of Teas (1993), who stated that expec­
tations could be interpreted as attribute importance, we assume that
the gap between importance and performance can represent the gap
between expectations and perceptions. We consider the distance
(gap) between the importance and performance of each service
attribute and represent it on a circular graph composed of two
different sectors (Figure 2).
The external sector contains the attributes for which importance
is higher than performance; in other words, these are the attributes
for which the perceptions of the users do not reach the expectations.
The inside sector contains the attributes for which importance is
lower than performance, that is the attributes for which perfor­
mance exceeds the expectations. We can consider the service
aspects falling within the external sector as criticalities, on which
financial resources should be concentrated because the perfor­
mance is distant from the importance. On the other hand, the
4 J. ALLEN ET AL.
Figure 2. Gap-IPA.
aspects falling within the inside sector have to be considered as
aspects that operators could disregard or consider as not prior
because the performance exceeds the importance. The degree of
criticality and non-priority varies as a function of the value of the
distance between importance and performance. As an example, the
attributes located near the border dividing the two sectors could
require particular attention because they could easily jump from
a sector to the opposite one.
The application of Gap-IPA requires that importance and per­
formance have to be expressed according to the same scale. For this
reason, we propose a normalization of the values from 0 to 1.
Therefore, the external sector contains attributes with a gap from
0 to 1, where 0 indicates that performance is equal to importance,
and 1 the maximum gap between importance and performance, or
the maximum distance of performance from importance. On the
other hand, the inside sector contains the attributes with a gap from
−1 to 0, where −1 the maximum gap between performance and
importance, or the maximum distance of importance from perfor­
mance. The value of 0 delineates the border dividing the two
sectors.
We retain that the graphical representation of Gap-IPA can offer
a more straightforward reading of the data regarding the IPA. First
of all, IPA proposes a subdivision of the service attributes in well
four quadrants, where the reader could be disoriented, while
according to Gap-IPA there is a sharper division of the attributes
in only two groups. Secondly, for interpreting IPA, a continuous
comparison between importance and performance has to be
effected, while Gap-IPA adopts only one value, which is the gap
between importance and performance, a more immediate
approach. In this way, it acquires more relevance the gap between
importance and performance than the separate concepts. As an
example, we could have a first attribute with high performance
and higher importance and a second one with low performance
and lower importance. These two attributes could fall into the
external area, while according to IPA they would be in different
quadrants. If we analyze the condition of the two attributes, we
could conclude that both the attributes are not priorities and would
rightly fall into the internal area of non-priorities. Moreover, the
position of the attribute in the graph, which could be near or far the
border separating the two sectors would be the difference. This last
observation guarantees that applying Gap-IPA there is not a loss of
information as regards IPA.
We decided to adopt the SEM approach for calculating impor­
tance. The reason for our decision was suggested by the capabilities
of the SEM as comparing to other tools used for calculating impor­
tance, such as regression models. Specifically, an SEM approach
differs from a traditional regression model because it introduces
latent variables in addition to observed ones. Latent variables
represent theoretical concepts or unobservable constructs that can­
not be directly measured by the analyst. In contrast, observed
variables allow latent constructs to be measured by considering
latent measurement errors (Bowen and Guo 2012), for a more
precise measurement. The assessment of service quality is
a complex issue due to the numerousness of the aspects character­
izing a service and the relationship between them. SEM permits to
consider both observed and latent constructs taking part in the
vastness of service quality, and all the direct and indirect relation­
ship existing among them.
In order to give a clearer idea of the proposed methodology, we
summarized the various steps in Figure 3.
First, the data collection was made through the survey addressed
to a sample of air passengers. From the collected data we had the
possibility to derive the performance ratings by the simple calcula­
tion of the average satisfaction rates from the judgments expressed

Figure 3. Methodological framework.
by users; while the importance ratings were derived thanks to the
calibration of the SEM model. Before applying the SEM approach,
a PCA was performed to determine how service attributes group
together into latent constructs. This statistical technique can help to
explore the composition of the factors and to analyze the relation­
ships among the measured variables. In other words, PCA permits
to convert a set of correlated variables into a set of uncorrelated
variables called principal components (Jolliffe 2014). Therefore, we
constructed the conceptual model introducing the hypotheses on
the relationships among latent endogenous and exogenous vari­
ables supported by the literature review, which was investigated
through the SEM approach. Performance and importance ratings
were calculated and used for applying IPA, and finally computing
the Gap-IPA.
Case study
The case study is the Lamezia Terme international airport, the most
important airport in the Calabria region, south of Italy. The S.A.
CAL. S.p.a. manages the airport infrastructures and all the activities
of private operators working in the airport. Over the years, the air
traffic at the Lamezia Terme airport has considerably developed.
During 2017, the passengers registered were more than 2,500,000
and the flights, considering both landings and take-offs, was about
22,000 (S.A.CAL. Calabrian Airport Company 2018). The terminal
is spread over three floors. At the ground and second level, there are
only offices, toilets, and restaurants. All the most important opera­
tions related to passengers’ movements take place on the first floor.
The S.A.CAL monitors the services provided to the passengers
inside the terminal, every year, through face-to-face CS surveys.
The survey is addressed to the departing passengers, who spend
more time in the terminal than the arriving passengers because
Lamezia Terme is not a hub airport. The employed questionnaire
consists of several questions regarding all the services offered by the
airport, such as signposting of the terminal, airport staff, waiting
time at the check-in, personal security, cleanliness, and comfort.
Interviewed passengers had to evaluate the selected service aspects
by expressing a rating through a verbal measurement scale ranging
from ‘very poor’ to ‘excellent.’
The data analyzed in this study were collected during the period
from January 2015 to December 2016. Two thousand eighty-seven
completed questionnaires were obtained. After a preliminary selec­
tion of valid data, a sample of 1,873 records was obtained. The
sample’s characterization is reported in Table 1.
The sample consists of more females (53%) than males; about
40% of the users are aged from 30 to 50 (Table 1). Most of the
passengers come from Italy (60%); however, about 50% of them are
going toward other European or non-European countries. The high
TRANSPORTATION LETTERS 5
Table 1. Sample characteristics.
Sample
characteristics Percentage
Gender Female (52.8%), Male (47.2%)
Age Less than 30 (13.2%), between 30 and 39 (23.7%),
between 40 and 49 (16.2%), between 50 and 59
(13.3%), more than 60 (10.1%), No answer (23.5%)
Traveling alone Yes (6.7%), No (93.3%)
Trip purpose Work/business (15.5%), Holiday (74.2%), Study (1.2%),
Medical care (4.0%), other (5.2%)
Nationality Italy (60.5%), other Europe Countries (28.9%), non-Europe
Countries (10.6%)
Flight destination Italy (47.8%), other European Countries (40.8%), non-
European Countries (11.4%)
Level of education Junior high school diploma (7.2%), High school diploma
(28.8%), Bachelor or Master’s degree (24.0%), No
answer (40.0%)
Mode for reaching Car as driver (16.0%), Car as passenger (46.8%), Taxi
the airport (3.7%), Rental car (15.5%), Rental bus (7.3%), Bus
(5.8%), Bus shuttle (4.9%)
Time of arrival Less than 1 hour before the flight (8.1%), From 1 to
2 hours before the flight (48.7%), More than 2 hours
before the flight (43.2%)
percentages of users who travel for a holiday (74%) and with other
people (93.3%) highlight a strong tourist vocation of the airport.
Most of the passengers arrive at the airport by car (82%), especially
by a car driven by someone else (47%). Almost half of the sample
arrives at the airport from 1 to 2 hours before the departure time of
the flight, 43% of the users arrive more than 2 hours early, and the
remaining part of passengers less than 1 hour before.
From the analysis of the judgments expressed by the passengers,
we note that the most frequent one is ‘good’ for all the service
attributes. However, certain aspects were less satisfactory for the
passengers. The highest percentages of negative judgments (ranging
from ‘very poor’ and ‘poor’) were registered for ‘cleanliness of the
toilets,’ ‘terminal air conditioning,’ and ‘cleanliness of terminal.’
The service quality attributes with the highest percentage of ‘excel­
lent’ judgments are ‘waiting time at check-in’ and ‘personal secur­
ity’ (see Table A.1 in Appendix A).
Results
SEM
Firstly, we introduced an unobserved endogenous variable repre­
senting the overall service quality (so-called OVSERVICE). The
theoretical concept of ‘overall airport service quality’ was already
introduced in many previous papers specifically focused on transit
service quality (Lai and Chen 2011; Sumaedi, Bakti, and Medi 2012;
de Oña et al. 2016; Allen et al. 2019, 2020a) and air transport service
6 J. ALLEN ET AL.
quality (Nesset and Helgesen 2014; Prentice and Kadan 2019; Allen
et al. 2020b). This abstract construct is not directly measurable by
the analyst; a preliminary analysis of the data drove us to two
service quality aspects which can be used for measuring the above
introduced latent construct, and which can be considered as
observed endogenous variables: ‘Terminal comfort’ and
‘Availability and efficiency of the airport services.’ PCA confirmed
this hypothesis.
From PCA, we discover three latent constructs that can be
considered as unobserved exogenous variables in the model. The
obtained results are reported in Table A.2 (Appendix A) where we
show with bold characters the scores corresponding to that specific
principal component. The first one (PC1), so-called ACCESS,
represents the accessibility to the services, and it is explained by
service aspects related to information and signposting. The second
one (PC2), so-called CONTROL, comprises all the factors linked to
the control operations in the terminal and also includes processes
related to check-in and baggage handling. The last one (PC3), so-
called ENVIRONMENT, represents the sense of passengers' well-
being in the terminal, and it is explained by service aspects related
to cleanliness and air conditioning.
We hypothesize direct and indirect effects among the latent
constructs. Specifically, five hypotheses have been introduced to
be tested by the SEM approach:
H1: ACCESS latent construct has a direct effect on OVSERVICE
latent construct.
H2: CONTROL latent construct has a direct effect on OVSERVICE
latent construct.
H3: ACCESS latent construct has a direct effect on CONTROL and
an indirect effect on OVSERVICE latent constructs.
H4: ENVIRONMENT latent construct has a direct effect on
OVSERVICE latent construct.
H5: ACCESS latent construct has a direct impact on
ENVIRONMENT and an indirect effect on OVSERVICE latent
constructs.
Figure 4. Conceptual model.
The conceptual model we aim to test adopting SEM approach
can be outlined as in Figure 4.
From the conceptual model, we can observe that the ACCESS
variable assumes the role of an antecedent exogenous construct,
which impacts OVSERVICE both directly and indirectly, with
CONTROL and ENVIRONMENT latent constructs acting as med­
iator variables.
The results obtained for the measurement model are shown in
Table 2, where the nomenclature is the same adopted in Bollen
(1989). Specifically, xi is the observed exogenous variables, and yi is
the observed endogenous variables. Otherwise, the latent exogenous
and endogenous variables are represented by ξi and ηi, respectively.
The measurement model defines the relationships among
hypothesized latent variables and the observed variables whose
scores they influence. In Table 2, (***) in the P-value column
indicates that the estimated parameter is significant at a level smal­
ler than 0.001. The standardized regression weights represent the
amount of change in the dependent variable that is attributable to
a single standard deviation unit’s worth of change in the predictor
variable. In Table 3, the results obtained for the structural model are
reported. The significance of the resulting statistics of the SEM
model confirmed all the formulated hypotheses.
The model consists of 13 observed variables, and 21 unobserved
variables, including four latent constructs and 17 error terms, one
for each observed variable and latent construct. The estimated
parameters were finally 52, consisting of 35 regression weights
and 17 variances. Chi-square Minimum is 900.816 (CMIN) with
60 Degrees of Freedom (DF). As reported in Hu and Bentler (1999),
CMIN/DF was calculated to indicate the magnitude of discrepancy
between the sample and fitted covariance’s matrix. The obtained
value (15.02) is significant at a 0.000 probability level, and it is
higher than the recommended value of 5.0 (Hooper, Coughlan,
and Mullen 2008). An acceptable fit is confirmed also by the Root
mean square error of approximation (RMSEA) equal to 0.08 (Hu
and Bentler 1999). The Goodness of Fit Index (GFI equal to 0.93)
indicates a well-fitting model, with a recommended cutoff point of
0.90 (Hooper, Coughlan, and Mullen 2008). However, the Adjusted
Goodness of Fit Index (AGFI) presents a lower value (0.90).
The model comparisons fit indices indicate that the hypothe­
sized model fits the observed variance-covariance matrix (Normed
Fit Index, NFI equal to 0.86) enough. This result is reinforced by the
Comparative Fit Index that represents a revised form of the NFI,
 TRANSPORTATION LETTERS 7

Table 2. Measurement models. services (ACCESS) is better explained by indicators related to

st. information than signposting; specifically, the most significant stan­
RW* SE* P* RW* dardized weight is obtained by the indicator ‘accessibility of infor­
Road signposting (x1) ACCESS (ξ1) 1.000 0.477 mation.’ This result can be partly counterintuitive; however, in the
Flight information (x2) ACCESS (ξ1) 0.930 0.051 *** 0.653 authors’ opinion, this happens because Lamezia Terme is a small
Terminal signposting (x3) ACCESS (ξ1) 0.928 0.053 *** 0.610
Infopoint and security staff (x4) ACCESS (ξ1) 0.947 0.049 *** 0.740
airport where all the areas are close to each other. Also, Prentice and
Information accessibility (x5) ACCESS (ξ1) 0.934 0.047 *** 0.851 Kadan (2019) registered that the airport’s signs have a significant
Waiting time at check-in (x6) CONTROL (ξ2) 1.000 0.418 role in directing passengers to services/facilities. As expected,
Baggage and passenger CONTROL (ξ2) 1.807 0.120 *** 0.757 CONTROL latent construct (control operations in the terminal) is
control (x7) better explained by indicators related to passenger control and
Personal security (x8) CONTROL (ξ2) 1.541 0.103 *** 0.751
Cleanliness of terminal (x9) ENVIRONMENT 3.383 0.453 *** 0.791 personal security. The terminal environment gives a sense of well-
(ξ3) being to the passengers more if the cleanliness of the terminal and
Cleanliness of toilets (x10) ENVIRONMENT 3.486 0.462 *** 0.647 toilets is perceived as satisfactory. As in Norazah (2014), air-
(ξ3) conditioning in the airport results less critical for the passengers
Terminal air conditioning (x11) ENVIRONMENT 1.000 0.208
(ξ3)
staying in the terminal than the cleanliness of the airport toilets,
Terminal comfort (y1) OVSERVICE (η1) 6.438 0.996 *** 0.674 although in our sample we have a significant percentage of passen­
Availability and efficiency of OVSERVICE (η1) 1.000 0.203 gers who do not use this service. Finally, we found that ‘terminal
the airport services (y2) comfort’ is the indicator mainly affecting overall service
(*) RW (Regression Weights), SE (Standard error), P (Probability level), st.RW (stan­ (OVSERVICE), as confirmed by Sunran and Min-Su (2012),
dardized Regression Weights) whereas ‘availability and efficiency of the airport services’ has
a lower influence on it.
Tables 4 and 5 summarize the elements we used for deriving the
Table 3. Structural model.
importance of the application of IPA and Gap-IPA. In Table 4,
RW* SE* P* st.RW*
direct and indirect effects among the latent variables, calculated by
OVSERVICE (η1) ACCESS (ξ1) 0.055 0.012 *** 0.299 taking into account the antecedent and mediator variables, are
OVSERVICE (η1) CONTROL (ξ2) 0.059 0.014 *** 0.235
OVSERVICE (η1) ENVIRONMENT (ξ3) 0.194 0.040 *** 0.544 shown. Successively, the total effects reported in Table 5 were
CONTROL (ξ2) ACCESS (ξ1) 0.360 0.033 *** 0.493 distributed among the service attributes by considering the values
ENVIRONMENT (ξ3) ACCESS (ξ1) 0.179 0.029 *** 0.348 of the standardized regression weights, as shown in Table 5.
(*) RW (Regression Weights), SE (Standard error), P (Probability level), st.RW (stan­
dardized Regression Weights)
IPA and Gap-IPA
taking into account sample size (CFI equal to 0.87). Although the Before applying Gap-IPA, we applied IPA, intending to highlight
last indices showed values a little bit lower than the cutoff recom­ the analogies and the differences between the two techniques and to
mended by several authors (0.90), Bollen (1989) suggests that these better explain the advantages of Gap-IPA as regards IPA.
criteria are merely guidelines. As an example, some authors report The values derived multiplying the total effects of the standar­
that CFI ≥ 0.8 is good enough for the structural validity of the dized regression weights (Table 5) were adopted as the importance
model (Browne and Cudeck 1993; Hair et al. 2010). By considering of each attribute. On the other hand, the average value of the
the obtained results (Table 4), it is evident that passengers’ percep­ judgments directly expressed by the passengers about each service
tions about overall service OVSERVICE are directly affected mainly attribute was considered as performance. A numerical value was
by the latent construct related to terminal environment assigned to each rating of the verbal scale, ranging from 1 (very
ENVIRONMENT (0.544), and secondly by the accessibility to the poor) to 5 (excellent). By observing the results of IPA, we can see
airport services ACCESS (0.299). However, there are significant that importance range varies from 0.10 to 0.51 (with an average
indirect effects of the latent construct ACCESS on the value of 0.30), whereas Performance varies from 3.65 to 4.21 (with
OVSERVICE mediated by both the latent constructs an average value of 3.97) (Table 6).
ENVIRONMENT (0.348) and CONTROL (0.493). Accounting for Importance and performance values were normalized for the appli­
both direct and indirect effects allows obtaining a total effect of cation of Gap-IPA, as the calculation of the gap, a subtraction, requires
ACCESS on OVSERVICE equal to 0.604. In other words, having that the values are expressed following the same scale. More specifically,
precise information and signposting inside the terminal makes the the normalization was made by considering the minimum values
airport services more accessible and, at the same time, increases the registered by the attributes both for importance and performance.
sense of passengers' well-being in the terminal. In turn, passengers’ For this reason, we can easily observe that the normalized importance
satisfaction with the overall service is improved.
On the other hand, having precise information and signposting
inside the terminal makes control operations easier and check-in or Table 5. Importance calculation.
baggage handling faster, improving passengers’ satisfaction with Total effect St.rw Importance
CONTROL and OVSERVICE latent aspects. Evidence from the Road signposting (x1) 0.604 0.477 0.29
measurement model (Table 2) shows that accessibility to the airport Flight information (x2) 0.653 0.39
Terminal signposting (x3) 0.610 0.37
Infopoint and security staff (x4) 0.740 0.45
Information accessibility (x5) 0.851 0.51
Table 4. Direct, indirect, and total effects. Waiting time at check-in (x6) 0.235 0.418 0.10
Baggage and passenger control (x7) 0.757 0.18
Direct effect Indirect effect Total effect Personal security (x8) 0.751 0.18
ACCESS (ξ1) 0.299 0.305 0.604 Cleanliness of terminal (x9) 0.544 0.791 0.43
CONTROL (ξ2) 0.235 0.000 0.235 Cleanliness of toilet (x10) 0.647 0.35
ENVIRONMENT (ξ3) 0.544 0.000 0.544 Terminal air conditioning (x11) 0.208 0.11
8 J. ALLEN ET AL.

Table 6. Importance, performance, and gap. performance. Finally, all the other service attributes, included in the
Normalized Normalized last two quadrants, can be considered as service aspects with a low
Importance Performance Importance Performance Gap priority, on which the agency could not focus the efforts. Definitively,
Road 0.29 3.82 0.46 0.31 0.15 for a small-sized airport as Lamezia Terme airport, the areas where
signposting policy-based actions will likely result in the most significant improve­
Flight 0.39 3.98 0.71 0.59 0.12
information
ment were identified in such aspects linked to cleanliness and comfort.
Terminal 0.37 3.97 0.65 0.57 0.08 ‘Cleanliness of terminal’ is the best candidate for policymaking because
signposting it has the highest importance, and performance lower than the average
Infopoint and 0.45 4.06 0.84 0.73 0.11 value obtained for all the service aspects.
security staff In Figure 6, the results of Gap-IPA can be easily observed.
Information 0.51 4.06 1.00 0.73 0.27
accessibility Accurately, we reported on a circular graph composed of two
Waiting time at 0.10 4.21 0.00 1.00 −1.00 different sectors the difference (gap) between the importance and
check-in performance of each service attribute. In the external sector, includ­
Baggage and 0.18 4.20 0.19 0.98 −0.78 ing the aspects for which importance is higher than performance,
passenger
control
most of the service attributes are located. More specifically, the
Personal 0.18 4.15 0.19 0.90 −0.71 significant criticalities are surely represented by the attributes
security regarding cleanliness, as also discovered through the IPA. The
Cleanliness of 0.43 3.92 0.80 0.48 0.32 attributes concerning information result as criticalities by Gap-
terminal IPA, while according to IPA, they were considered as strengths of
Cleanliness of 0.35 3.65 0.61 0.00 0.61
toilets the service because they registered high values of both importance
Terminal air 0.11 3.77 0.04 0.21 −0.17 and performance. But thanks to the calculation of the gap between
conditioning importance and performance, we can verify that even if perfor­
mance is high, it does not achieve the importance, and for this
reason, the attributes have to be considered as criticalities according
assumes a minimum value of 0.00 for the attribute ‘waiting time at to the concepts at the basis of Gap-IPA. It should be however
check-in,’ which presented the minimum value of 0.10 in the IPA, and highlighted that these attributes are located near the border divid­
a maximum value of 1.00 for the attribute ‘information accessibility,’ ing the two sectors; therefore, with a minimum effort from the
which assumed the maximum value of 0.51 in the IPA. Analogously, company, they could easily pass in the inside sector and be con­
concerning performance, ‘cleanliness of toilets’ has a normalized mini­ sidered as non-priorities.
mum value of 0.00 (3.65 in IPA) and ‘waiting time at check-in’ max­ The most performant attribute is surely ‘waiting time at check-in,’
imum value of 1.00 (4.21 in IPA). which registered a gap of −1, assuming the highest value of perfor­
The results from IPA can be observed in Figure 5. The first quadrant mance and the lowest value of importance. Other quite performant
contains the service attributes regarding information, which are the attributes are ‘personal security’ and ‘baggage and passenger control,’
strengths of the service; therefore, the agency should maintain high the which were possible overkill according to IPA because they registered
level of quality for these attributes for achieving competitive advantage. high values of both performance and importance. Finally, also the
On the contrary, the service attributes particularly in need of improve­ attribute regarding terminal air conditioning can be considered a non-
ments are cleanliness of terminal, cleanliness of toilets and terminal priority, but it is located near the border, and therefore it could pass in
signposting, because they have high importance but relatively low the external sector if performance decreases.

Figure 5. An application of IPA.


Figure 6. An application of Gap-IPA.
Discussion and practical implications
Evidence emerged from IPA and Gap-IPA demonstrating that the
Gap-IPA succeeds in highlighting useful information that IPA is
not able to capture because this latter considers the values of
importance and performance but without directly comparing
them as it happens in Gap-IPA, where importance and performance
are made in comparison through the calculation of the gap.
Moreover, the graphical representation of Gap-IPA is undoubtedly
more immediate and easily interpretable than IPA. In fact, IPA
requires to observe what happens in four quadrants, in Gap-IPA
a simple overlook at two sectors is sufficient for having a clear idea
of the criticalities of the service, without missing the most essential
information.
We discovered that the service provided by the Lamezia Terme
airport presents two main criticalities, which are the attributes linked
to cleanliness. Another attribute for which performance is relatively
distant to the expectation is ‘information accessibility.’ Finally, the
other four criticalities (i.e. the other two attributes regarding infor­
mation, and the attributes linked to signposting both in the terminal
and on-road for reaching the airport) are located near the border, and
for this reason, they could be considered as less critical.
Starting from the results of Gap-IPA, the right strategies for
effectively managing the service can be more conveniently iden­
tified. As an example, significant efforts should be concentrated
on the cleanliness of the terminal and toilets, to satisfy passen­
gers and make their stay in the airport as more comfortable.
The aspects concerning information represent less urgent criti­
calities, but a particular effort should also be reserved for this
aspect, in order to make the performance equal to the expecta­
tion and make the passengers fully satisfied. A certain level of
attention should also be addressed to terminal air conditioning
because in this case, the performance is a little more than
importance, and if the company does not take care of this
aspect, it could easily become a criticality of the service.
Finally, the policies of the company could be less focused on
aspects such as ‘waiting time at check-in,’ ‘personal security,’
TRANSPORTATION LETTERS 9
and ‘baggage and passenger control.’ This kind of result could
be due to the peculiarity of the analyzed airport, which is
a small-sized airport offering, for the most part, national flights,
while international scheduled flights were primarily concen­
trated in the summer months. Aspects linked to check-in, or
baggage and passenger control, or personal security perform
well probably because of the small dimensions of the airport
and the relatively contained number of passengers. Particularly
impressive is the result concerning the attributes linked to
information and signposting, which should be improved in
order to make it a strength of the service, as having precise
information and signposting inside the terminal make control
operations easier and check-in or baggage handling faster,
improving passengers’ satisfaction on overall service. Different
findings could be registered in studies analyzing large-sized
airports characterized by relevant traffic data. As an example,
Pandey (2016) attempted to measure the service quality of the
two gateway airports of Thailand Suvarnabhumi (BKK) and
Don Mueang (DMK) by utilizing the Fuzzy MCDM Analysis
and also conducting IPA using the Fuzzy expert system. He
discovered that for both the analyzed airports, there is a need to
improve aspects linked to check-in, security, and speed of
baggage delivery service, which in our study they can be con­
sidered as non-priorities.
We can conclude that aside from the case study, the pro­
posed methodology represents a useful and practical tool for
supporting the operators to identify the most beneficial strate­
gies for improving the service and adequately investing the
financial resources. Gap-IPA could be considered an alternative
to the well-known IPA, due to the immediateness and easiness
of representing and interpreting the results. The key of the work
is that the Gap-IPA brings the two-dimensional matrix of IPA
into a single indicator (i.e. the gap between importance and
performance) that can more conveniently help the administra­
tion to prioritize the service aspects resulting as the most
critical, but also the ones close to becoming critical.
10 J. ALLEN ET AL.

Acknowledgments Hansen, E., and R. J. Bush. 1999. “Understanding Customer Quality

Requirements: Model and Application.” Industrial Marketing Management
S.A.CAL S.p.A (management company of Lamezia Terme airport) supported 28 (2): 119–130.
this research making available the data collected through Customer Satisfaction Hong, S. J., D. Choi, and J. Chae. 2020. “Exploring Different Airport Users’
Surveys. Service Quality Satisfaction between Service Providers and Air Travelers.”
Journal of Retailing and Consumer Services 52: 101917. doi:10.1016/j.
jretconser.2019.101917.
Disclosure statement Hooper, D., J. Coughlan, and M. R. Mullen. 2008. “Structural Equation
No potential conflict of interest was reported by the authors. Modelling: Guideline for Determining Model Fit.” Electronic Journal of
Business Research Methods 6 (1): 53–60.
Hu, L., and P. M. Bentler. 1999. “Cutoff Criteria for Fit Indexes in Covariance
Structure Analysis: Conventional Criteria versus New Alternatives.”
ORCID Structural Equation Modeling 6: 1–55.
Jaime Allen http://orcid.org/0000-0002-7616-009X IATA, International Air Transport Association. 2018. World Air Transport
Maria Grazia Bellizzi http://orcid.org/0000-0002-1650-1729 Statistics.
Gabriella Mazzulla http://orcid.org/0000-0003-0993-454X Jiang, H., and Y. Zhang. 2016. “An Assessment of Passenger Experience at
Melbourne Airport.” Journal of Air Transport Management 54: 88–92.
Jolliffe, I. 2014. “Principal Component Analysis.” In Wiley StatsRef:
Statistics Reference Online, edited by N. Balakrishnan, T. Colton,
References B. Everitt, W. Piegorsch, F. Ruggeri, and J. L. Teugels. Hoboken, New
Jersey: Wiley.
Allen, J., M. G. Bellizzi, L. Eboli, C. Forciniti, and G. Mazzulla. 2020b. “Service
Lai, W. T., and C. F. Chen. 2011. “Behavioral Intentions of Public Transit
Quality in A Mid-sized Air Terminal: A SEM-MIMIC Ordinal Probit
Passenger. The Roles of Service Quality, Perceived Value, Satisfaction and
Accounting for Travel, Sociodemographic, and User-type Heterogeneity.”
Involvement.” Transport Policy 18: 318–325.
Journal of Air Transport Management 84. doi:10.1016/j.
Lambert, D. M., and A. Sharma. 1990. “A Customer-based Competitive Analysis
jairtraman.2020.101780.
for Logistics Decisions.” International Journal of Physical Distribution &
Allen, J., L. Eboli, C. Forciniti, G. Mazzulla, and J. D. Ortúzar. 2019. “The Role of
Logistics Management 20 (1): 17–24.
Critical Incidents and Involvement in Transit Satisfaction and Loyalty.”
Lubbe, B., A. Douglas, and J. Zambellis. 2011. “An Application of the Airport
Transport Policy 75: 57–69. doi:10.1016/j.tranpol.2019.01.005.
Service Quality Model in South Africa.” Journal of Air Transport
Allen, J., L. Eboli, G. Mazzulla, and J. D. Ortúzar. 2020a. “Effect of Critical
Management 17: 224–227.
Incidents on Public Transport Satisfaction and Loyalty: An Ordinal Probit
Martilla, J. A., and J. C. James. 1977. “Importance-performance Analysis.” The
SEM-MIMIC Approach.” Transportation 47: 827–863. doi:10.1007/s11116-
Journal of Marketing 77–79.
018-9921-4.
Matzler, K., F. Bailom, H. H. Hinterhuber, B. Renzl, and J. Pichler. 2004. “The
Ashford, N., H. P. M. Stanton, and C. A. Moore. 1997. Airport Operations.
Asymmetric Relationship between Attribute-level Performance and Overall
New York: McGraw-Hill.
Customer Satisfaction: A Reconsideration of the Importance–performance
Babu, J. A. 2015. “Airport Service Improvement Strategies Using
Analysis.” Industrial Marketing Management 33: 271–277.
Importance-Performance Analysis.” International Journal for Research in
Nesset, E., and Ø. Helgesen. 2014. “Effects of Switching Costs on Customer
Applied Science & Engineering Technology 3 (4): 381–392.
Attitude Loyalty to an Airport in a Multi-airport Region.” Transportation
Bellizzi, M. G., L. dell’Olio, L. Eboli, and G. Mazzulla. 2020. “Heterogeneity in
Research Part A 67: 240–253.
Desired Bus Service Quality from Users’ and Potential Users’ Perspective.”
Norazah, M. S. 2014. “Passenger Satisfaction with Airline Service Quality in
Transportation Research Part A 132: 365–377.
Malaysia: A Structural Equation Modeling Approach.” Research in
Bezerra, C. L., and C. F. Gomes. 2015. “The Effects of Service Quality Dimensions
Transportation Business & Management 10: 26–32.
and Passenger Characteristics on Passenger’s Overall Satisfaction with an
Pandey, M. M. 2016. “Evaluating the Service Quality of Airports in Thailand
Airport.” Journal of Air Transport Management 44-45: 77–81.
Using Fuzzy Multi-criteria Decision Making Method.” Journal of Air
Bezerra, C. L., and C. F. Gomes. 2016. “Measuring Airport Service Quality:
Transport Management 57: 241–249.
A Multidimensional Approach.” Journal of Air Transport Management 53: 85–93.
Pantouvakis, A., and M. F. Renzi. 2016. “Exploring Different Nationality
Bollen, K. 1989. Structural Equations with Latent Variables. New York: John
Perceptions of Airport Service Quality.” Journal of Air Transport
Wiley.
Management 52: 90–98.
Bowen, N. K., and S. Guo. 2012. Structural Equation Modelling. Oxford,
Parasuraman, A., L. L. Berry, and V. A. Zenithal. 1991. “Understanding
England: Oxford University Press.
Customer Expectations of Service.” Sloan Management Review Spring 39–48.
Brida, J. G., L. Moreno-Izquierdo, and S. Zapata-Aguirre. 2016. “Customer
Park, J. W., and S. Y. Jung. 2011. “Transfer Passengers’ Perceptions of Airport
Perception of Service Quality: The Role of Information and
Service Quality: A Case Study of Incheon International Airport.”
Communication Technologies (Icts) at Airport Functional Areas.” Tourism
International Business Research 4 (3): 75–82.
Management Perspectives 20: 209–216.
Postorino, M. N., L. Mantecchini, C. Malandri, and F. Paganelli. 2019. “Airport
Browne, M. W., and R. Cudeck. 1993. “Alternative Ways of Assessing Model
Passenger Arrival Process: Estimation of Earliness Arrival Functions.”
Fit.” In Testing Structural Equation Models, edited by K. A. Bollen and
Transportation Research Procedia 37: 338–345.
J. S. Long, 136–162. Newbury Park, USA: SAGE Publication.
Prentice, C., and M. Kadan. 2019. “The Role of Airport Service Quality in
Ceccato, V., and S. Masci. 2017. “Airport Environment and Passengers’
Airport and Destination Choice.” Journal of Retailing and Consumer
Satisfaction with Safety.” Journal of Applied Security Research 12.3: 356–373.
Services 47: 40–48.
de Oña, J., R. de Oña, L. Eboli, C. Forciniti, and G. Mazzulla. 2016. “Transit
S.A.CAL. Calabrian Airport Company. 2018. Service Charter Guide.
Passengers’ Behavioural Intentions: The Influence of Service Quality and
Sampson, S. E., and M. J. Showalter. 1999. “The Performance-importance
Customer Satisfaction. Transportmetrica A.” Transport Science 12: 385–412.
Response Function: Observations and Implications.” The Service Industries
Deng, W. 2007. “Using a Revised Importance-performance Analysis Approach:
Journal 19 (3): 1–25.
The Case of Taiwanese Hot Springs Tourism.” Tourism Management 28:
Sumaedi, S., I. G. M. Y. Bakti, and Y. Medi. 2012. “The Empirical Study of Public
1274–1284.
Transport Passengers’Behavioral Intentions: The Roles of Service Quality,
Eboli, L., C. Forciniti, and G. Mazzulla. 2018a. “Formative and Reflective
Perceived Sacrifice, Perceived Value, and Satisfaction (Case Study:
Measurement Models for Analysing Transit Service Quality.” Public
Paratransit Passengers in Jakarta, Indonesia).” International Journal for
Transport 10: 107–127.
Traffic and Transport Engineering 2 (1): 83–97.
Eboli, L., C. Forciniti, and G. Mazzulla. 2018b. “Spatial Variation of the
Sunran, J., and K. Min-Su. 2012. “The Effect of the Servicescape on Customers’
Perceived Transit Service Quality at Rail Stations.” Transportation Research
Behavioural Intentions in an International Airport Service Environment.”
Part A: Policy and Practice 114: 67–83.
Service Business 6: 279–295.
Hair, J. F., R. E. Anderson, R. L. Tatham, and W. C. Black. 2010. Multivariate
Teas, R. K. 1993. “Expectations, Performance Evaluation and Consumer’s
Data Analysis. Englewood Cliffs: Prentice-Hall.
Perception of Quality.” Journal of Marketing 57 (4): 18–34.
 TRANSPORTATION LETTERS 11

Tseng, C. C. 2020. “An IPA-Kano Model for Classifying and Diagnosing Airport Appendix A

Table A1. Judgments about each airport service quality aspect.

Service quality aspect Valid data Very poor Poor Fair Good Excellent
Road signposting 1765 32 (1.8%) 103 (5.8%) 161 (9.1%) 1319 (74.7%) 150 (8.5%)
Flight information 1743 9 (0.5%) 32 (1.8%) 86 (4.9%) 1468 (84.2%) 148 (8.5%)
Terminal signposting 1814 3 (0.2%) 67 (3.7%) 61 (3.4%) 1533 (84.5%) 150 (8.3%)
Infopoint and security staff 1742 3 (0.2%) 14 (0.8%) 70 (4.0%) 1445 (83.0%) 210 (12.1%)
Information accessibility 1702 0 (0.0%) 4 (0.2%) 72 (4.2%) 1446 (85.0%) 180 (10.6%)
Waiting time at check-in 1345 13 (1.0%) 26 (1.9%) 57 (4.2%) 817 (60.7%) 432 (32.1%)
Baggage and passenger control 1840 11 (0.6%) 14 (0.8%) 77 (4.2%) 1237 (67.2%) 501 (27.2%)
Personal security 1768 3 (0.2%) 14 (0.8%) 69 (3.9%) 1305 (73.8%) 377 (21.3%)
Cleanliness of terminal 1791 43 (2.4%) 74 (4.1%) 127 (7.1%) 1288 (71.9%) 259 (14.5%)
Cleanliness of toilets 1429 124 (8.7%) 107 (7.5%) 91 (6.4%) 930 (65.1%) 177 (12.4%)
Terminal air conditioning 1780 34 (1.9%) 204 (11.5%) 105 (5.9%) 1237 (69.5%) 200 (11.2%)
Terminal comfort 1782 12 (07%) 66 (4.4%) 167 (9.4%) 1380 (77.4%) 157 (8.8%)
Availability and efficiency of the airport services 1453 0 (0.0%) 6 (0.4%) 100 (6.9%) 1290 (88.8%) 57 (3.9%)

Table A2. Exploratory PCA.

PC1 PC2 PC3
Road signposting 0.578 0.155 −0.010
Flight information 0.766 0.065 0.087
Terminal signposting 0.711 0.070 0.127
Infopoint and security staff 0.719 0.145 0.167
Information accessibility 0.814 0.187 0.120
Waiting time at check-in 0.076 0.573 0.187
Baggage and passenger control 0.216 0.830 −0.041
Personal security 0.161 0.798 0.166
Cleanliness of terminal 0.095 0.262 0.801
Cleanliness of toilets 0.004 0.066 0.871
Terminal air conditioning 0.138 0.021 0.321

Service Attributes.” Research in Transportation Business & Management.

doi:10.1016/j.rtbm.2020.100499.