- Email: [email protected]
Air Transport Passengers’ Satisfaction: an Ordered Logit Model
Available online at www.sciencedirect.com Available online at www.sciencedirect.com
ScienceDirect ScienceDirect
Transportation Research Procedia 00 (2018) 000–000 Available online at www.sciencedirect.com Transportation Research Procedia 00 (2018) 000–000
ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Transportation Research Procedia 33 (2018) 147–154 www.elsevier.com/locate/procedia
XIII Conference on Transport Engineering, CIT2018 XIII Conference on Transport Engineering, CIT2018
Air Transport Passengers’ Satisfaction: an Ordered Logit Air Transport Passengers’Model Satisfaction: an Ordered Logit Model a a a, a Maria Grazia Bellizzi , Laura Eboli , Carmen Forciniti *, Gabriella Mazzulla Maria Grazia Bellizzia, Laura Ebolia, Carmen Forcinitia,*, Gabriella Mazzullaa a
University of Calabria, Department of Civil Engineering, Via Pietro Bucci 46B, Rende 87036, Italy
a
University of Calabria, Department of Civil Engineering, Via Pietro Bucci 46B, Rende 87036, Italy
Abstract Abstract Evaluating air transport service quality performance is of crucial importance both for countries’ economic development and airline industry, which has a vitalquality role inperformance countries’ development competitiveness. multicultural nature of this industry Evaluating air transport service is of crucialand importance both forThe countries’ economic development and requires strong effort offer good levels of quality. and competitiveness. The multicultural nature of this industry airline industry, whichtohas a services vital rolewith in countries’ development Airport and to services are the first a passenger receives upon arrival. Therefore, measuring the levels of requiresfacilities strong effort offer services with experiences good levels that of quality. airport services evaluating is essential to understand the needs of customers. Airport facilitiesbyand services passengers’ are the first satisfaction experienceswith that them a passenger receives upon arrival. Therefore, measuring the levels of In this services work, we want to propose a toolsatisfaction for measuring airport service quality starting judgements expressed by airport by evaluating passengers’ with them is essential to understand thefrom needsthe of customers. passengers about aspects. The data supporting airport the work refer to the services Lamezia Terme airport, In this work, wesome wantservice to propose a tool for measuring service quality startingoffered from by thethe judgements expressed by the most important airport of Calabria, in the South of Italy. The data were collected through face-to-face interviews addressed to passengers about some service aspects. The data supporting the work refer to the services offered by the Lamezia Terme airport, departing passengers. Passengers expressed an ordinal verbal scale about some service factors the most important airport of Calabria, in the their Southperceptions of Italy. Theaccording data weretocollected through face-to-face interviews addressed to concerningpassengers. aspects such as information, security, andaccording comfort. to an ordinal verbal scale about some service factors departing Passengers expressed theircleanliness perceptions We proposeaspects an ordered logit model with the aim to investigate on the influence of the various service factors on the overall concerning such as information, security, cleanliness and comfort. service quality, by considering passengers’ satisfaction with the service andofthe the overall model We propose an ordered logit model with the aim to investigate on the factors influence theoverall variousservice. serviceSpecifically, factors on the can be adopted for considering identifying the combinations of the with satisfaction levels of theand service factorsservice. that return the level the of overall service quality, by passengers’ satisfaction the service factors the overall Specifically, model satisfaction desired the operator. We propose some for different of users in order verifythe thelevel differences in can be adopted for by identifying the combinations of themodels satisfaction levels ofgroups the service factors that toreturn of overall perceptions service satisfaction of desired byquality. the operator. We propose some models for different groups of users in order to verify the differences in perceptions of service quality. © 2018 The Authors. Published by Elsevier Ltd. © 2018 The Authors. by Elsevier Ltd. This is an open accessPublished article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) © 2018 The Authors. by Elsevier Ltd. This is an open accessPublished article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. CIT2018. Keywords: Air transport; airport; service quality; passengers' satisfaction; ordered logit model. Keywords: Air transport; airport; service quality; passengers' satisfaction; ordered logit model.
* Corresponding author. Tel.: +39 0984 494020; fax: +39 0984 496787. E-mail address:author. [email protected] * Corresponding Tel.: +39 0984 494020; fax: +39 0984 496787. E-mail address: [email protected]
2352-1465 © 2018 The Authors. Published by Elsevier Ltd. This is an open access under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) 2352-1465 © 2018 Thearticle Authors. Published by Elsevier Ltd. Selection under responsibility of the scientific of the XIII Conference on Transport Engineering, This is an and openpeer-review access article under the CC BY-NC-ND licensecommittee (https://creativecommons.org/licenses/by-nc-nd/4.0/) CIT2018. Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. 2352-1465 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. 10.1016/j.trpro.2018.10.087
148 2
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
1. Introduction Today the role of airport Service Quality (SQ) is considered as a fundamental contribution to airport attractiveness (Pantouvakis and Renzi, 2016). The evaluation of SQ is also relevant for airport authorities for better allocating resources and organizing their investment strategies (Liou et al., 2011). Service business operators often assess SQ provided to their customers in order to improve their service, to quickly identify problems, and to better assess client satisfaction. For this reason, the need to explore the nature of airport SQ and its components is evident in the relevant transport and marketing literature (Fodness and Murray, 2007). SQ is defined as “an assessment of how well a delivered service conforms to the client’s expectations”; due to the complexity of the phenomenon, its measurement is not a trivial practice. Often, SQ assessment is made by capturing Customer Satisfaction, defined as “the degree of satisfaction provided by the goods or services of a company as measured by the number of repeat customers” (www.businessdictionary.com). SQ depends by many different variables that are not necessarily observed directly (Kennet and Salini, 2012). The choice of the crucial variables depends on the type of service analysed, because it is necessary to identify attributes characterizing the service and influencing user perceptions. Generally, these variables have to reflect the service environment investigated, so they are “context dependent” (Yeh and Kuo, 2003). In an airport, there is a lot of different activities dealing with aviation, security controls, shopping, etc.; as a consequence, many different and independent attributes have to be considered in order to measure the delivered SQ. Within the airport-related literature, there is not a single way to classify the terminal activities. Bogicevic et al. (2013), consider the transport service as a chain process, distinguish two major components of consumer’ air travel experience: airport ground service and in-flight service. In general, the “airport ground services” are also called “landside operations”. The airport terminal may be divided in three major areas: access interface, processing area, and flight interface (Horonjeff et al., 2010). In detail, Popovic et al. (2009) distinguish process activities (regarding passenger flow from check-in, security screening, and passport control to boarding gate) from discretionary activities (regarding passengers’ ways to engage their slack time such as getting a coffee, shopping, or exchanging money). Instead, Liu (2016) considers the service system of an airport company divided into two sub-processes: aeronautical service and commercial service. Despite these differences in activities’ classification, the attributes used to evaluate airport SQ are similar among the authors. In this context, the Airport Council International (ACI) introduced the Airport Service Quality (ASQ) program in order to better understand passengers’ expectations from an airport’s facilities and services, and to measure passenger’ experience in different airports. ACI introduces different SQ indicators, which focus on how passengers perceive the level of service and on objective measures of service delivery (ACI, 2012). Also the Italian civil aviation authority (ENAC) defined the Charter of Airport Standard Services that established some parameters related to SQ, among them trip security, overall waiting time, cleanliness, comfort, additional services, information to passengers, ticketing/check-in/passport control services, connections to the airport. Due to the complexity and variety of airport services, there is still absence of unique dimensions’ definition. Also for this reason is extremely important to investigate on the factors which describe the service offered at the airports, and to identify which are the most relevant factors on the overall satisfaction of the passengers. Not many researchers have studied the phenomenon of satisfaction/dissatisfaction of users at the airport, and the activities and offered services inside an airport have been treated in different ways. For example, Yeh and Kuo (2003) used different performance measures associated to several service factors for the airport service evaluation. Popovic et al. (2009) discussed how activities mediate people’s experiences in the airport through an observational analysis, instead Liu (2016) measured the aeronautical service efficiency and the commercial service efficiency of different airports. The objective of this paper is just to investigate on the perceptions of the airport passengers about different aspects of an airport service. Specifically, we propose Ordered Logit (OL) models for identifying the service attributes that passengers consider as the most important, and we calibrate the models for different groups of users in order to identify the differences in passengers’ perceptions. Similar studies have been conducted by Bogicevic et al. (2013), Bezerra and Gomez (2015, 2016) and by Pantouvakis and Renzi (2016). The former, through data mining techniques, identified which airport attributes are enhancers of passenger satisfaction without taking account the different characteristics of the users. Pantouvakis and Renzi (2016), with a multinomial logit model, explore only the
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
149 3
potential differences concerning resident versus visiting nationalities perceptions of airport SQ. Bezerra and Gomez (2015) also have modelled the phenomenon with an ordinal logistic regression but they considered the personal characteristics of the passengers like parameters of the model and not as a way to divide the sample of users to calibrate different models. As regards the state of the art, the contribution of the proposed work is to provide a further probabilistic approach for modelling SQ within the airport context and exploring the differences among passengers' profiles considering more than one characteristic. The case study is the Lamezia Terme airport, which is the most relevant airport of Calabria, a region sited in the south of Italy. We analyse the data collected by sample surveys addressed to passengers departing from the airport, who expressed their perceptions about some service aspects (e.g. modal integration, information, personnel, security, cleanliness, comfort). Specifically, 3,313 face-to-face interviews were effected from 2012 to 2014 by the agency managing the airport. The rest of the paper is organized as follows. The next section focuses on the presentation of the study case, and more specifically we propose a brief description of the airport and its services, a description of the survey and the adopted questionnaire, and the statistical analysis of the data, by showing the frequency distributions of the judgements expressed by the passengers. The third section is about the OL model and the results. Finally, in the last section of the paper the conclusions about the work are summarized. 2. The study case 2.1. The Lamezia Terme airport Our case study is the Lamezia Terme International airport, which is the principal airport of Calabria region (Italy), and it is among the first twenty Italian airports. The airport is located in the middle of Calabria, and it is connected to the Highway A2 and the Lamezia Terme Central railway station. Airport is reachable from many Calabrian cities by bus links. During 2016, Lamezia Terme airport registered more than 2,500,000 passengers and about 22,000 flights between landings and take-offs (S.A.CAL., 2017). Scheduled domestic traffic has been centred on the airports of the main Italian cities (e.g. Rome, Milan, Turin). International scheduled flights were primarily concentrated in the summer months with destination to many European cities, such as Amsterdam, Berlin, Brussels, London, Paris, Prague, Vienna, Zurich and the intercontinental link with Toronto. Additional links with foreign cities will be scheduled soon due to new agreements with some major airlines. The recently completed extension of the runway is now on use. With a length of over 3,000 meters, the runway allows non-stop routes to intercontinental destinations. 2.2. Data collection Data collection refers to the period from March 2012 to December 2014. The surveys were designed and realized by the S.A.CAL. S.p.A., the managing company of Lamezia Terme airport, which is under the direct control and surveillance of ENAC (National Agency for Civil Aviation). S.A.CAL. manages the airport infrastructure and coordinates and controls the activities of private operators working in the airport, trying to maintain adequate standards of service and airport safety. In order to promote the improvement and to prevent inefficiencies, S.A.CAL. constantly monitors many parameters (quality indicators) by measuring the actual events (assessment of provided SQ) and passengers’ perceptions (assessment of the perceived SQ). The first part of the questionnaire has the aim to collect some passengers’ characteristics such as country of origin and trip purpose. The second part has the aim to investigate on customer satisfaction, and contains six groups of service aspects, named: modal integration, information, airport staff, security, cleanliness, and comfort. Each of these six groups includes one or more SQ’ attributes. SQ attributes used in the questionnaire were selected based on the specifications of the Charter of Airport Standard Services of ENAC. Passengers were asked to express a judgement on each service attribute and on the overall service. The evaluation scale is an ordinal verbal scale, and it is composed of five judgements, from “Very poor” to “Excellent”, and of an additional response “Service not used”. The sample is made of 3,313 respondents.
150 4
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
2.3. Data analysis In order to consider the different perceptions among passengers concerning the service aspects, data were analyzed according to three different partitions of the passengers; for each division, two different passengers’ categories were identified. Specifically, they were divided by nationality, in “Italian” and “Other”, by trip purpose, in “Leisure” and “Other”, by earliness of arrival, in “Less than 2 hours” and “More than 2 hours”. Most of passengers comes from Italy (79%), 14% from Europe, and the remaining 7% from extra-Europe countries. The major part of passengers travels for “leisure” (63%), and this percentage highlights a strong tourist vocation of the airport. Other trip purposes are “work and business” (19%), “medical care” (8%), study and university (2%), and “other” (8%). Regarding arrival time of passengers at the airport, most of them (53%) arrives from one to two hours early, 10% arrives less than one hour before, and 37% more than two hours before. For each service attribute, the percentage of valid responses is for almost all the attributes higher than 90%. We calculated the relative frequency distributions of the judgements expressed by the passengers divided according to nationality, trip purpose and earliness of arrival. Concerning the opinion expressed for the overall service, the judgement “good” is the most frequently expressed by the passengers, and none of users pronounced “Very poor” judgment.
Fig. 1. Frequency distributions concerning the differences by “Earliness of arrival”.
Concerning the two categories of passengers divided by nationality, we registered that the judgment “good” is expressed more by the non-Italian passengers for all the service attributes, and that the non-Italian passengers expressed the judgment level “very poor”, while more than 20% of Italian passengers expressed this level. For certain attributes, even twice of the non-Italian users expressed the level “good” as regards the Italian ones, and
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
151 5
specifically for all the service attributes concerning information and comfort. In general, we can conclude that nonItalian passengers are most satisfied with the service. Concerning the groups of passengers divided by trip purpose, we noted no relevant differences. On the other hand, some interesting differences were registered concerning the groups of passengers divided by earliness of arrival (Figure 1). Specifically, passengers arriving more than two hours early are less satisfied with the attributes relating information as they expressed less “good” judgements and more “very poor” judgments as regards passengers arriving late. The same tendency can be observed for the attributes relating staff, security, cleanliness and comfort. From these results, it would seem that passengers arriving early discover more criticalities, because they spend more time in the airport and have more time to experience the various service aspects. 3. The Ordered Logit models The Ordered Logit (OL) model is an extension of the logistic regression applied when the dependent variable is categorical and has a meaningful order with more than two categories (or levels). The ordinal variable Y is a function of another variable Y,* that is continuous and not measured and has various threshold points. The value of the observed variable depends on whether or not it crossed a particular threshold, as showed by the following formulas: Yi 1 if Yi* k1
(1)
Yi j if ki Yi* ki 1
(2)
Yi M if Yi* kM 1
(3)
The continuous latent variable K
Yi* k 1 k X ki i
is equal to: (4)
in which there is a random disturbance term normally distributed. The error term reflects the fact that the variables may not be perfectly measured, and some relevant variables may be not introduced in the equation. The vector of β parameters is estimated by the Maximum Likelihood method and generally the goodness-of-fit of the OL model is verified by Nagelkerke R2 (Eboli and Mazzulla, 2009). The statistical impact of variables is based on the pvalues of the Wald tests (Eboli et al., 2016). In this case study, the OL model was adopted to estimate the weight of each service attribute on the overall service offered by the Lamezia Terme airport. Seven different OL models were calibrated by using data collected by the survey above described: one model for all the passengers and one for each of the six groups of passengers divided by nationality, trip purpose and earliness of arrival. In these models the independent variables are the judgments on the investigated service attributes, while the dependent variable is the satisfaction with the overall service (Table 1). The general model, calibrated considering the whole sample of passengers, is useful for identifying the most important service aspects. The models calibrated for each passengers’ category allow to compare users according to their own characteristics or inclinations. The calibration of the models was effected by SPSS statistical software through the PLUM procedure (PoLytomus Universal Model procedure). Three levels of quality were defined for the variables: a low level of quality (corresponding to the judgements “Very Poor” and “Poor”), an intermediate level (corresponding to the judgement “Fair”) and a high level (corresponding to the judgements “Good” and “Excellent”). Values from 0 to 2 were associated respectively to these three levels of quality. In this way, each independent variable and the dependent one are classified into three categories. The baseline category, or “reference case”, is represented by the variable when it takes the level 2. As a consequence of the choice of the reference case, the coefficients (β) of the statistically significant variables have a
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
152 6
negative sign. The negative sign means that the contribution to the overall dissatisfaction decreases when the independent variables assume higher values, or rather the users are more satisfied with the single service attribute. The statistics on the goodness of fit of all the models are adequate. The values of Nagelkerke R2 are all higher or close to 0.5. Table 1. OL Models. Global model Nationality Trip purpose Earliness of Arrival VARIABLES Italian Other Leisure Other < 2 hours > 2 hours Roadlinks=0 -0.422 (**) -0.941 (***) - (n.s.) -0.361 (*) -0.670 (**) -1.075 (***) - (n.s.) Roadlinks=1 - (n.s.) -0.546 (**) - (n.s.) - (n.s.) -0.701 (**) -0.575 1.021 (**) Display=0 -1.805 (***) -1.860 (***) -2.158 (***) -1.884 (***) -2.189 (***) -1.512 (***) -2.648 (***) Display=1 -0.417 (*) - (n.s.) - (n.s.) - (n.s.) -1.034 (**) - (n.s.) - (n.s.) Signposting=0 -1.402 (***) -1.547 (***) -0.934 (*) -1.645 (***) -0.961 (**) -1.730 (***) -1.376 (**) Signposting=1 0.438 (*) 1.023 (**) -0.958 (*) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Announcements=0 -2.096 (***) -2.572 (***) -1.487 (***) -1.948 (***) -2.901 (***) -2.369 (***) -1.664 (***) Announcements=1 -1.328 (***) -1.803 (***) - (n.s.) -1.194 (***) -1.661 (***) -1.481 (***) -0.959 (**) Staff=0 -1.695 (***) -1.986 (***) -1.126 (*) -2.410 (***) -1.264 (**) -1.901 (***) - (n.s.) Staff=1 - (n.s.) - (n.s.) - (n.s.) -0.778 (*) - (n.s.) - (n.s.) - (n.s.) Control=0 - (n.s.) - (n.s.) -1.587 (**) - (n.s.) - (n.s.) -0.848 (**) - (n.s.) Control=1 -0.496 (*) - (n.s.) -1.890 (**) -1.590 (***) 0.978 (**) - (n.s.) -1.298 (**) Security=0 -0.730 (*) -1.123 (**) - (n.s.) - (n.s.) -1.553 (**) -1.211 (**) - (n.s.) Security=1 0.521 (*) - (n.s.) 2.358 (**) 0.894 (**) - (n.s.) - (n.s.) - (n.s.) Cleanliness=0 -1.305 (***) -1.360 (***) -1.294 (**) -1.354 (***) -1.530 (**) -1.391 (***) -1.723 (***) Cleanliness=1 -0.836 (***) -0.848 (**) -0.942 (*) -1.444 (***) - (n.s.) - (n.s.) -2.367 (***) Toilettes=0 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Toilettes=1 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) -0.771 (**) - (n.s.) Airconditioning=0 - (n.s.) -0.448 (**) - (n.s.) - (n.s.) -0.671 (**) - (n.s.) -1.001 (**) Airconditioning=1 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Lighting=0 -1.758 (***) -1.528 (***) -1.650 (**) -1.981 (***) -0.958 (*) -1.355 (**) -2.445 (***) Lighting=1 -1.257 (***) -1.274 (***) - (n.s.) -2.071 (***) - (n.s.) -1.292 (***) -2.324 (***) Noise=0 -2.083 (***) -2.418 (***) -1.629 (***) -1.986 (***) -2.473 (***) -2.953 (***) -1.297 (**) Noise=1 -2.509 (***) -2.777 (***) -1.400 (**) -2.866 (***) -2.071 (***) -2.778 (***) -3.293 (***) k0 (threshold) -8.788 (***) -9.598 (***) -8.025 (***) -9.908 (***) -8.701 (***) -9.334 (***) -10.644 (***) k1 (threshold) -4.087 (***) -4.717 (***) -3.167 (***) -3.981 (***) -5.035 (***) -4.789 (***) -3.389 (***) LogLikelihood with 2402.471 1925.177 546.271 1519.976 963.612 1686.762 816.445 zero coefficients Final LogLikelihood 1144.656 840.844 315.141 688.556 462.219 751.569 399.466 R2 of Nagelkerke 0.544 0.597 0.458 0.565 0.584 0.611 0.518 Notes: (*) Significant at level of 90 per cent; (**) 95 per cent; (***) 99 per cent; (n.s.) Not Significant
The results of the general model drive to the fact that the significant impact on the overall satisfaction is given by the service attributes regarding “Information” (Display, Signposting and Announcement), “Cleanliness of terminal” and “Comfort” (Lighting and Noise). The models calibrated for the passengers’ categories allow to establish the most important service aspects according to users’ characteristics and inclinations such as nationality, trip purpose and earliness of arrival at terminal. The comparison between the model “Italian” and the model “Other” shows significant differences. As an example, road links, announcements, staff and noise are very important for Italians and less for Others. An opposite result is obtained for baggage control; also the aspect concerning display is more relevant for the foreign passengers. Finally, signposting inside terminal, lighting and cleanliness are relevant for both categories of passengers. The most important service aspects for passengers whose trip purpose is “Leisure” are partly different from passengers whose trip purpose is “Other” (work and business, study and university, and medical care). For the first
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
153 7
ones the statistically significant service aspects are: announcements, courtesy and friendliness of staff, cleanliness of terminal, lightning inside terminal, and acoustic comfort. For the others are road links, display, announcements and noise. So, people travelling for leisure give more importance to the service aspects related to comfort, unlike the others whose priority are “technical” aspects such as modal integration and information inside terminal. Finally, there are some differences between the passengers whose earliness of arrival at terminal is less than 2 hours and the passengers whose earliness of arrival is more than 2 hours. More specifically, while announcements, lightning inside terminal and noise are statistically significant for both the categories of users, people arriving early give more importance to lighting than the others, who give more importance to the announcements and the acoustic comfort. In addition, for users whose earliness of arrival at terminal is less than 2 hours road links are important, instead cleanliness of terminal is important for the others, probably due to their longer stay at the airport. 4. Conclusions In this paper, OL models are proposed with the aim to identify which service attributes have an influence on passengers’ satisfaction with the overall service, and to investigate about the differences of perceptions among different groups of users; for this aim, some models were calibrated for different groups of passengers. More specifically, we decided to investigate on the differences in terms of nationality, in order to verify if Italian people have different perceptions of the services offered at the airport as regards foreign passengers. We investigated also the differences in terms of trip purpose which we retain as an important aspect that can influence passengers’ behavior. Finally, we decided to investigate on an aspect concerning the attitude of the passengers in arriving at the airport ahead of time or not. We found many interesting elements from the analysis of the data. We have to precise that the analysis through the OL models allows to find best and more useful results as regards the simple analysis performed through the frequency distribution of the data. As an example, if we observe the distribution of the frequencies of the judgments of passengers differentiated in terms of trip purpose we can note that this analysis does not highlight relevant differences, which can be discovered, on the other hand, from the models. More specifically, we discovered that people travelling for leisure retain as most important aspects related to comfort, while people travelling for purposes linked to work or study or other important reasons give priority to “technical” aspects such as modal integration and information inside terminal, maybe because they are less relaxed than people travelling for leisure. Other interesting findings regard people arriving more than two hours early, who give more importance to cleanliness of terminal probably because they stay too time at the airport, or the foreign passengers, who consider as more important the aspect concerning display because they know the airport information less than Italian passengers. Respect to previous literature studies developing similar models (e.g. Bogicevic et al. (2013), Bezerra and Gomez (2015) and Pantouvakis and Renzi (2016)), the present work has taken into account the different characteristics of the passengers and identified the SQ aspects that affect the perceptions of the several groups of passengers. In conclusion, the findings resulted from the proposed models demonstrate that it is fundamental to investigate on the passengers’ perceptions to discover which are the service aspects mostly influencing passengers’ satisfaction, and therefore most important to be monitored by the operators in order to be competitive and to offer services characterized by good levels of quality. In addition, the results highlighted the importance to understand the differences of perceptions among groups of passengers with the aim to identify marketing strategies based on the different categories of users. ACKNOWLEDGEMENTS The authors would like to thank the S.A.CAL S.p.A for making available the data collected through the customer satisfaction surveys. Support from Spanish Ministry of Economy and Competitiveness (Research Project TRA201566235-R) is gratefully acknowledged.
154 8
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
REFERENCES ACI-Airport Council International, 2012. Guide to Airport Performance Measures. Agresti, A., 2007. An Introduction to Categorical Data Analysis. Wiley, Hoboken. Bezerra, G.C.L., Gomes, C.F., 2015. The effects of service quality dimensions and passenger characteristics on passenger’s overall satisfaction with an airport. Journal of air transport management 44.45, 77-81. Bezerra, G.C.L., Gomes, C.F., 2016. Measuring airport service quality: a multidimensional approach. Journal of air transport management 53, 8593. Bogicevic, V., Yang, W., Bilgihan, A., Bujisic, M., 2013. Airport service quality drivers of passenger satisfaction. Tourism Review 68.4, 3-18. Eboli, L., Forciniti, C., Mazzulla, M., Calvo, F., 2016. Exploring the Factors that Impact on Transit Use through an Ordered Probit Model: the Case of Metro of Madrid. Transportation Research Procedia 18, 35-43. Eboli, L., Mazzulla, G., 2009. An ordinal logistic regression model for analysing airport passenger satisfaction. EuroMed Journal of Business 4.1, 40-57. ENAC - National Agency for Civil Aviation, 2014. Qualità dei servizi nel trasporto aereo: le Carte dei servizi standard per gestori aeroportuali e vettori aerei. Fodness, D., Murray, B., 2007. Passengers’ expectations of airport service quality. Journal of Services Marketing 21.7, 492-506. Horonjeff, R., McKelvey, F.X., Sproul, W.J., Young, S.B., 2010. Planning and Design of Airports. McGraw-Hill, New York. Kennet, R.S., Salini, S., 2012. Modern Analysis of Customer Surveys: with Applications using R. Wiley, Chichester. Liou, J.J., Tang, C.H., Yeh, W.C., Tsai, C.Y., 2011. A decision rules approach for improvement of airport service quality. Expert Systems with Applications 38, 13723-13730. Liu, D., 2016. Measuring aeronautical service efficiency and commercial service efficiency of East Asia airport companies: an application of network data envelopment analysis. Journal of air transport management 52, 11-22. Long, J.S., 1997. Regression Models for Categorical and Limited Dependent Variables. SAGE Publications, Thousand Oaks. Pantouvakis, A., Renzi, M.F., 2016. Exploring different nationality perceptions of airport service quality. Journal of air transport management 52, 90-98. Popovic, V., Kraal, B.J., KIRK, P.J., 2009. Passenger experience in an airport: an activity-centred approach. IASDR 2009, 18-22 October 2009. Korean Society of Design Service, Coex, Seoul. S.A.CAL. - Calabrian Airport Company, 2017. Service Charter Guide. Yeh, C.-H., Kuo, Y.-L., 2003. Evaluating passenger services of Asia-Pacific international airports. Transportation Research Part E 39, 35-48.
ScienceDirect ScienceDirect
Transportation Research Procedia 00 (2018) 000–000 Available online at www.sciencedirect.com Transportation Research Procedia 00 (2018) 000–000
ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Transportation Research Procedia 33 (2018) 147–154 www.elsevier.com/locate/procedia
XIII Conference on Transport Engineering, CIT2018 XIII Conference on Transport Engineering, CIT2018
Air Transport Passengers’ Satisfaction: an Ordered Logit Air Transport Passengers’Model Satisfaction: an Ordered Logit Model a a a, a Maria Grazia Bellizzi , Laura Eboli , Carmen Forciniti *, Gabriella Mazzulla Maria Grazia Bellizzia, Laura Ebolia, Carmen Forcinitia,*, Gabriella Mazzullaa a
University of Calabria, Department of Civil Engineering, Via Pietro Bucci 46B, Rende 87036, Italy
a
University of Calabria, Department of Civil Engineering, Via Pietro Bucci 46B, Rende 87036, Italy
Abstract Abstract Evaluating air transport service quality performance is of crucial importance both for countries’ economic development and airline industry, which has a vitalquality role inperformance countries’ development competitiveness. multicultural nature of this industry Evaluating air transport service is of crucialand importance both forThe countries’ economic development and requires strong effort offer good levels of quality. and competitiveness. The multicultural nature of this industry airline industry, whichtohas a services vital rolewith in countries’ development Airport and to services are the first a passenger receives upon arrival. Therefore, measuring the levels of requiresfacilities strong effort offer services with experiences good levels that of quality. airport services evaluating is essential to understand the needs of customers. Airport facilitiesbyand services passengers’ are the first satisfaction experienceswith that them a passenger receives upon arrival. Therefore, measuring the levels of In this services work, we want to propose a toolsatisfaction for measuring airport service quality starting judgements expressed by airport by evaluating passengers’ with them is essential to understand thefrom needsthe of customers. passengers about aspects. The data supporting airport the work refer to the services Lamezia Terme airport, In this work, wesome wantservice to propose a tool for measuring service quality startingoffered from by thethe judgements expressed by the most important airport of Calabria, in the South of Italy. The data were collected through face-to-face interviews addressed to passengers about some service aspects. The data supporting the work refer to the services offered by the Lamezia Terme airport, departing passengers. Passengers expressed an ordinal verbal scale about some service factors the most important airport of Calabria, in the their Southperceptions of Italy. Theaccording data weretocollected through face-to-face interviews addressed to concerningpassengers. aspects such as information, security, andaccording comfort. to an ordinal verbal scale about some service factors departing Passengers expressed theircleanliness perceptions We proposeaspects an ordered logit model with the aim to investigate on the influence of the various service factors on the overall concerning such as information, security, cleanliness and comfort. service quality, by considering passengers’ satisfaction with the service andofthe the overall model We propose an ordered logit model with the aim to investigate on the factors influence theoverall variousservice. serviceSpecifically, factors on the can be adopted for considering identifying the combinations of the with satisfaction levels of theand service factorsservice. that return the level the of overall service quality, by passengers’ satisfaction the service factors the overall Specifically, model satisfaction desired the operator. We propose some for different of users in order verifythe thelevel differences in can be adopted for by identifying the combinations of themodels satisfaction levels ofgroups the service factors that toreturn of overall perceptions service satisfaction of desired byquality. the operator. We propose some models for different groups of users in order to verify the differences in perceptions of service quality. © 2018 The Authors. Published by Elsevier Ltd. © 2018 The Authors. by Elsevier Ltd. This is an open accessPublished article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) © 2018 The Authors. by Elsevier Ltd. This is an open accessPublished article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. CIT2018. Keywords: Air transport; airport; service quality; passengers' satisfaction; ordered logit model. Keywords: Air transport; airport; service quality; passengers' satisfaction; ordered logit model.
* Corresponding author. Tel.: +39 0984 494020; fax: +39 0984 496787. E-mail address:author. [email protected] * Corresponding Tel.: +39 0984 494020; fax: +39 0984 496787. E-mail address: [email protected]
2352-1465 © 2018 The Authors. Published by Elsevier Ltd. This is an open access under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) 2352-1465 © 2018 Thearticle Authors. Published by Elsevier Ltd. Selection under responsibility of the scientific of the XIII Conference on Transport Engineering, This is an and openpeer-review access article under the CC BY-NC-ND licensecommittee (https://creativecommons.org/licenses/by-nc-nd/4.0/) CIT2018. Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. 2352-1465 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of the XIII Conference on Transport Engineering, CIT2018. 10.1016/j.trpro.2018.10.087
148 2
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
1. Introduction Today the role of airport Service Quality (SQ) is considered as a fundamental contribution to airport attractiveness (Pantouvakis and Renzi, 2016). The evaluation of SQ is also relevant for airport authorities for better allocating resources and organizing their investment strategies (Liou et al., 2011). Service business operators often assess SQ provided to their customers in order to improve their service, to quickly identify problems, and to better assess client satisfaction. For this reason, the need to explore the nature of airport SQ and its components is evident in the relevant transport and marketing literature (Fodness and Murray, 2007). SQ is defined as “an assessment of how well a delivered service conforms to the client’s expectations”; due to the complexity of the phenomenon, its measurement is not a trivial practice. Often, SQ assessment is made by capturing Customer Satisfaction, defined as “the degree of satisfaction provided by the goods or services of a company as measured by the number of repeat customers” (www.businessdictionary.com). SQ depends by many different variables that are not necessarily observed directly (Kennet and Salini, 2012). The choice of the crucial variables depends on the type of service analysed, because it is necessary to identify attributes characterizing the service and influencing user perceptions. Generally, these variables have to reflect the service environment investigated, so they are “context dependent” (Yeh and Kuo, 2003). In an airport, there is a lot of different activities dealing with aviation, security controls, shopping, etc.; as a consequence, many different and independent attributes have to be considered in order to measure the delivered SQ. Within the airport-related literature, there is not a single way to classify the terminal activities. Bogicevic et al. (2013), consider the transport service as a chain process, distinguish two major components of consumer’ air travel experience: airport ground service and in-flight service. In general, the “airport ground services” are also called “landside operations”. The airport terminal may be divided in three major areas: access interface, processing area, and flight interface (Horonjeff et al., 2010). In detail, Popovic et al. (2009) distinguish process activities (regarding passenger flow from check-in, security screening, and passport control to boarding gate) from discretionary activities (regarding passengers’ ways to engage their slack time such as getting a coffee, shopping, or exchanging money). Instead, Liu (2016) considers the service system of an airport company divided into two sub-processes: aeronautical service and commercial service. Despite these differences in activities’ classification, the attributes used to evaluate airport SQ are similar among the authors. In this context, the Airport Council International (ACI) introduced the Airport Service Quality (ASQ) program in order to better understand passengers’ expectations from an airport’s facilities and services, and to measure passenger’ experience in different airports. ACI introduces different SQ indicators, which focus on how passengers perceive the level of service and on objective measures of service delivery (ACI, 2012). Also the Italian civil aviation authority (ENAC) defined the Charter of Airport Standard Services that established some parameters related to SQ, among them trip security, overall waiting time, cleanliness, comfort, additional services, information to passengers, ticketing/check-in/passport control services, connections to the airport. Due to the complexity and variety of airport services, there is still absence of unique dimensions’ definition. Also for this reason is extremely important to investigate on the factors which describe the service offered at the airports, and to identify which are the most relevant factors on the overall satisfaction of the passengers. Not many researchers have studied the phenomenon of satisfaction/dissatisfaction of users at the airport, and the activities and offered services inside an airport have been treated in different ways. For example, Yeh and Kuo (2003) used different performance measures associated to several service factors for the airport service evaluation. Popovic et al. (2009) discussed how activities mediate people’s experiences in the airport through an observational analysis, instead Liu (2016) measured the aeronautical service efficiency and the commercial service efficiency of different airports. The objective of this paper is just to investigate on the perceptions of the airport passengers about different aspects of an airport service. Specifically, we propose Ordered Logit (OL) models for identifying the service attributes that passengers consider as the most important, and we calibrate the models for different groups of users in order to identify the differences in passengers’ perceptions. Similar studies have been conducted by Bogicevic et al. (2013), Bezerra and Gomez (2015, 2016) and by Pantouvakis and Renzi (2016). The former, through data mining techniques, identified which airport attributes are enhancers of passenger satisfaction without taking account the different characteristics of the users. Pantouvakis and Renzi (2016), with a multinomial logit model, explore only the
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
149 3
potential differences concerning resident versus visiting nationalities perceptions of airport SQ. Bezerra and Gomez (2015) also have modelled the phenomenon with an ordinal logistic regression but they considered the personal characteristics of the passengers like parameters of the model and not as a way to divide the sample of users to calibrate different models. As regards the state of the art, the contribution of the proposed work is to provide a further probabilistic approach for modelling SQ within the airport context and exploring the differences among passengers' profiles considering more than one characteristic. The case study is the Lamezia Terme airport, which is the most relevant airport of Calabria, a region sited in the south of Italy. We analyse the data collected by sample surveys addressed to passengers departing from the airport, who expressed their perceptions about some service aspects (e.g. modal integration, information, personnel, security, cleanliness, comfort). Specifically, 3,313 face-to-face interviews were effected from 2012 to 2014 by the agency managing the airport. The rest of the paper is organized as follows. The next section focuses on the presentation of the study case, and more specifically we propose a brief description of the airport and its services, a description of the survey and the adopted questionnaire, and the statistical analysis of the data, by showing the frequency distributions of the judgements expressed by the passengers. The third section is about the OL model and the results. Finally, in the last section of the paper the conclusions about the work are summarized. 2. The study case 2.1. The Lamezia Terme airport Our case study is the Lamezia Terme International airport, which is the principal airport of Calabria region (Italy), and it is among the first twenty Italian airports. The airport is located in the middle of Calabria, and it is connected to the Highway A2 and the Lamezia Terme Central railway station. Airport is reachable from many Calabrian cities by bus links. During 2016, Lamezia Terme airport registered more than 2,500,000 passengers and about 22,000 flights between landings and take-offs (S.A.CAL., 2017). Scheduled domestic traffic has been centred on the airports of the main Italian cities (e.g. Rome, Milan, Turin). International scheduled flights were primarily concentrated in the summer months with destination to many European cities, such as Amsterdam, Berlin, Brussels, London, Paris, Prague, Vienna, Zurich and the intercontinental link with Toronto. Additional links with foreign cities will be scheduled soon due to new agreements with some major airlines. The recently completed extension of the runway is now on use. With a length of over 3,000 meters, the runway allows non-stop routes to intercontinental destinations. 2.2. Data collection Data collection refers to the period from March 2012 to December 2014. The surveys were designed and realized by the S.A.CAL. S.p.A., the managing company of Lamezia Terme airport, which is under the direct control and surveillance of ENAC (National Agency for Civil Aviation). S.A.CAL. manages the airport infrastructure and coordinates and controls the activities of private operators working in the airport, trying to maintain adequate standards of service and airport safety. In order to promote the improvement and to prevent inefficiencies, S.A.CAL. constantly monitors many parameters (quality indicators) by measuring the actual events (assessment of provided SQ) and passengers’ perceptions (assessment of the perceived SQ). The first part of the questionnaire has the aim to collect some passengers’ characteristics such as country of origin and trip purpose. The second part has the aim to investigate on customer satisfaction, and contains six groups of service aspects, named: modal integration, information, airport staff, security, cleanliness, and comfort. Each of these six groups includes one or more SQ’ attributes. SQ attributes used in the questionnaire were selected based on the specifications of the Charter of Airport Standard Services of ENAC. Passengers were asked to express a judgement on each service attribute and on the overall service. The evaluation scale is an ordinal verbal scale, and it is composed of five judgements, from “Very poor” to “Excellent”, and of an additional response “Service not used”. The sample is made of 3,313 respondents.
150 4
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
2.3. Data analysis In order to consider the different perceptions among passengers concerning the service aspects, data were analyzed according to three different partitions of the passengers; for each division, two different passengers’ categories were identified. Specifically, they were divided by nationality, in “Italian” and “Other”, by trip purpose, in “Leisure” and “Other”, by earliness of arrival, in “Less than 2 hours” and “More than 2 hours”. Most of passengers comes from Italy (79%), 14% from Europe, and the remaining 7% from extra-Europe countries. The major part of passengers travels for “leisure” (63%), and this percentage highlights a strong tourist vocation of the airport. Other trip purposes are “work and business” (19%), “medical care” (8%), study and university (2%), and “other” (8%). Regarding arrival time of passengers at the airport, most of them (53%) arrives from one to two hours early, 10% arrives less than one hour before, and 37% more than two hours before. For each service attribute, the percentage of valid responses is for almost all the attributes higher than 90%. We calculated the relative frequency distributions of the judgements expressed by the passengers divided according to nationality, trip purpose and earliness of arrival. Concerning the opinion expressed for the overall service, the judgement “good” is the most frequently expressed by the passengers, and none of users pronounced “Very poor” judgment.
Fig. 1. Frequency distributions concerning the differences by “Earliness of arrival”.
Concerning the two categories of passengers divided by nationality, we registered that the judgment “good” is expressed more by the non-Italian passengers for all the service attributes, and that the non-Italian passengers expressed the judgment level “very poor”, while more than 20% of Italian passengers expressed this level. For certain attributes, even twice of the non-Italian users expressed the level “good” as regards the Italian ones, and
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
151 5
specifically for all the service attributes concerning information and comfort. In general, we can conclude that nonItalian passengers are most satisfied with the service. Concerning the groups of passengers divided by trip purpose, we noted no relevant differences. On the other hand, some interesting differences were registered concerning the groups of passengers divided by earliness of arrival (Figure 1). Specifically, passengers arriving more than two hours early are less satisfied with the attributes relating information as they expressed less “good” judgements and more “very poor” judgments as regards passengers arriving late. The same tendency can be observed for the attributes relating staff, security, cleanliness and comfort. From these results, it would seem that passengers arriving early discover more criticalities, because they spend more time in the airport and have more time to experience the various service aspects. 3. The Ordered Logit models The Ordered Logit (OL) model is an extension of the logistic regression applied when the dependent variable is categorical and has a meaningful order with more than two categories (or levels). The ordinal variable Y is a function of another variable Y,* that is continuous and not measured and has various threshold points. The value of the observed variable depends on whether or not it crossed a particular threshold, as showed by the following formulas: Yi 1 if Yi* k1
(1)
Yi j if ki Yi* ki 1
(2)
Yi M if Yi* kM 1
(3)
The continuous latent variable K
Yi* k 1 k X ki i
is equal to: (4)
in which there is a random disturbance term normally distributed. The error term reflects the fact that the variables may not be perfectly measured, and some relevant variables may be not introduced in the equation. The vector of β parameters is estimated by the Maximum Likelihood method and generally the goodness-of-fit of the OL model is verified by Nagelkerke R2 (Eboli and Mazzulla, 2009). The statistical impact of variables is based on the pvalues of the Wald tests (Eboli et al., 2016). In this case study, the OL model was adopted to estimate the weight of each service attribute on the overall service offered by the Lamezia Terme airport. Seven different OL models were calibrated by using data collected by the survey above described: one model for all the passengers and one for each of the six groups of passengers divided by nationality, trip purpose and earliness of arrival. In these models the independent variables are the judgments on the investigated service attributes, while the dependent variable is the satisfaction with the overall service (Table 1). The general model, calibrated considering the whole sample of passengers, is useful for identifying the most important service aspects. The models calibrated for each passengers’ category allow to compare users according to their own characteristics or inclinations. The calibration of the models was effected by SPSS statistical software through the PLUM procedure (PoLytomus Universal Model procedure). Three levels of quality were defined for the variables: a low level of quality (corresponding to the judgements “Very Poor” and “Poor”), an intermediate level (corresponding to the judgement “Fair”) and a high level (corresponding to the judgements “Good” and “Excellent”). Values from 0 to 2 were associated respectively to these three levels of quality. In this way, each independent variable and the dependent one are classified into three categories. The baseline category, or “reference case”, is represented by the variable when it takes the level 2. As a consequence of the choice of the reference case, the coefficients (β) of the statistically significant variables have a
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
152 6
negative sign. The negative sign means that the contribution to the overall dissatisfaction decreases when the independent variables assume higher values, or rather the users are more satisfied with the single service attribute. The statistics on the goodness of fit of all the models are adequate. The values of Nagelkerke R2 are all higher or close to 0.5. Table 1. OL Models. Global model Nationality Trip purpose Earliness of Arrival VARIABLES Italian Other Leisure Other < 2 hours > 2 hours Roadlinks=0 -0.422 (**) -0.941 (***) - (n.s.) -0.361 (*) -0.670 (**) -1.075 (***) - (n.s.) Roadlinks=1 - (n.s.) -0.546 (**) - (n.s.) - (n.s.) -0.701 (**) -0.575 1.021 (**) Display=0 -1.805 (***) -1.860 (***) -2.158 (***) -1.884 (***) -2.189 (***) -1.512 (***) -2.648 (***) Display=1 -0.417 (*) - (n.s.) - (n.s.) - (n.s.) -1.034 (**) - (n.s.) - (n.s.) Signposting=0 -1.402 (***) -1.547 (***) -0.934 (*) -1.645 (***) -0.961 (**) -1.730 (***) -1.376 (**) Signposting=1 0.438 (*) 1.023 (**) -0.958 (*) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Announcements=0 -2.096 (***) -2.572 (***) -1.487 (***) -1.948 (***) -2.901 (***) -2.369 (***) -1.664 (***) Announcements=1 -1.328 (***) -1.803 (***) - (n.s.) -1.194 (***) -1.661 (***) -1.481 (***) -0.959 (**) Staff=0 -1.695 (***) -1.986 (***) -1.126 (*) -2.410 (***) -1.264 (**) -1.901 (***) - (n.s.) Staff=1 - (n.s.) - (n.s.) - (n.s.) -0.778 (*) - (n.s.) - (n.s.) - (n.s.) Control=0 - (n.s.) - (n.s.) -1.587 (**) - (n.s.) - (n.s.) -0.848 (**) - (n.s.) Control=1 -0.496 (*) - (n.s.) -1.890 (**) -1.590 (***) 0.978 (**) - (n.s.) -1.298 (**) Security=0 -0.730 (*) -1.123 (**) - (n.s.) - (n.s.) -1.553 (**) -1.211 (**) - (n.s.) Security=1 0.521 (*) - (n.s.) 2.358 (**) 0.894 (**) - (n.s.) - (n.s.) - (n.s.) Cleanliness=0 -1.305 (***) -1.360 (***) -1.294 (**) -1.354 (***) -1.530 (**) -1.391 (***) -1.723 (***) Cleanliness=1 -0.836 (***) -0.848 (**) -0.942 (*) -1.444 (***) - (n.s.) - (n.s.) -2.367 (***) Toilettes=0 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Toilettes=1 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) -0.771 (**) - (n.s.) Airconditioning=0 - (n.s.) -0.448 (**) - (n.s.) - (n.s.) -0.671 (**) - (n.s.) -1.001 (**) Airconditioning=1 - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) - (n.s.) Lighting=0 -1.758 (***) -1.528 (***) -1.650 (**) -1.981 (***) -0.958 (*) -1.355 (**) -2.445 (***) Lighting=1 -1.257 (***) -1.274 (***) - (n.s.) -2.071 (***) - (n.s.) -1.292 (***) -2.324 (***) Noise=0 -2.083 (***) -2.418 (***) -1.629 (***) -1.986 (***) -2.473 (***) -2.953 (***) -1.297 (**) Noise=1 -2.509 (***) -2.777 (***) -1.400 (**) -2.866 (***) -2.071 (***) -2.778 (***) -3.293 (***) k0 (threshold) -8.788 (***) -9.598 (***) -8.025 (***) -9.908 (***) -8.701 (***) -9.334 (***) -10.644 (***) k1 (threshold) -4.087 (***) -4.717 (***) -3.167 (***) -3.981 (***) -5.035 (***) -4.789 (***) -3.389 (***) LogLikelihood with 2402.471 1925.177 546.271 1519.976 963.612 1686.762 816.445 zero coefficients Final LogLikelihood 1144.656 840.844 315.141 688.556 462.219 751.569 399.466 R2 of Nagelkerke 0.544 0.597 0.458 0.565 0.584 0.611 0.518 Notes: (*) Significant at level of 90 per cent; (**) 95 per cent; (***) 99 per cent; (n.s.) Not Significant
The results of the general model drive to the fact that the significant impact on the overall satisfaction is given by the service attributes regarding “Information” (Display, Signposting and Announcement), “Cleanliness of terminal” and “Comfort” (Lighting and Noise). The models calibrated for the passengers’ categories allow to establish the most important service aspects according to users’ characteristics and inclinations such as nationality, trip purpose and earliness of arrival at terminal. The comparison between the model “Italian” and the model “Other” shows significant differences. As an example, road links, announcements, staff and noise are very important for Italians and less for Others. An opposite result is obtained for baggage control; also the aspect concerning display is more relevant for the foreign passengers. Finally, signposting inside terminal, lighting and cleanliness are relevant for both categories of passengers. The most important service aspects for passengers whose trip purpose is “Leisure” are partly different from passengers whose trip purpose is “Other” (work and business, study and university, and medical care). For the first
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
153 7
ones the statistically significant service aspects are: announcements, courtesy and friendliness of staff, cleanliness of terminal, lightning inside terminal, and acoustic comfort. For the others are road links, display, announcements and noise. So, people travelling for leisure give more importance to the service aspects related to comfort, unlike the others whose priority are “technical” aspects such as modal integration and information inside terminal. Finally, there are some differences between the passengers whose earliness of arrival at terminal is less than 2 hours and the passengers whose earliness of arrival is more than 2 hours. More specifically, while announcements, lightning inside terminal and noise are statistically significant for both the categories of users, people arriving early give more importance to lighting than the others, who give more importance to the announcements and the acoustic comfort. In addition, for users whose earliness of arrival at terminal is less than 2 hours road links are important, instead cleanliness of terminal is important for the others, probably due to their longer stay at the airport. 4. Conclusions In this paper, OL models are proposed with the aim to identify which service attributes have an influence on passengers’ satisfaction with the overall service, and to investigate about the differences of perceptions among different groups of users; for this aim, some models were calibrated for different groups of passengers. More specifically, we decided to investigate on the differences in terms of nationality, in order to verify if Italian people have different perceptions of the services offered at the airport as regards foreign passengers. We investigated also the differences in terms of trip purpose which we retain as an important aspect that can influence passengers’ behavior. Finally, we decided to investigate on an aspect concerning the attitude of the passengers in arriving at the airport ahead of time or not. We found many interesting elements from the analysis of the data. We have to precise that the analysis through the OL models allows to find best and more useful results as regards the simple analysis performed through the frequency distribution of the data. As an example, if we observe the distribution of the frequencies of the judgments of passengers differentiated in terms of trip purpose we can note that this analysis does not highlight relevant differences, which can be discovered, on the other hand, from the models. More specifically, we discovered that people travelling for leisure retain as most important aspects related to comfort, while people travelling for purposes linked to work or study or other important reasons give priority to “technical” aspects such as modal integration and information inside terminal, maybe because they are less relaxed than people travelling for leisure. Other interesting findings regard people arriving more than two hours early, who give more importance to cleanliness of terminal probably because they stay too time at the airport, or the foreign passengers, who consider as more important the aspect concerning display because they know the airport information less than Italian passengers. Respect to previous literature studies developing similar models (e.g. Bogicevic et al. (2013), Bezerra and Gomez (2015) and Pantouvakis and Renzi (2016)), the present work has taken into account the different characteristics of the passengers and identified the SQ aspects that affect the perceptions of the several groups of passengers. In conclusion, the findings resulted from the proposed models demonstrate that it is fundamental to investigate on the passengers’ perceptions to discover which are the service aspects mostly influencing passengers’ satisfaction, and therefore most important to be monitored by the operators in order to be competitive and to offer services characterized by good levels of quality. In addition, the results highlighted the importance to understand the differences of perceptions among groups of passengers with the aim to identify marketing strategies based on the different categories of users. ACKNOWLEDGEMENTS The authors would like to thank the S.A.CAL S.p.A for making available the data collected through the customer satisfaction surveys. Support from Spanish Ministry of Economy and Competitiveness (Research Project TRA201566235-R) is gratefully acknowledged.
154 8
Maria Grazia Bellizzi et al. / Transportation Research Procedia 33 (2018) 147–154 Author name / Transportation Research Procedia 00 (2018) 000–000
REFERENCES ACI-Airport Council International, 2012. Guide to Airport Performance Measures. Agresti, A., 2007. An Introduction to Categorical Data Analysis. Wiley, Hoboken. Bezerra, G.C.L., Gomes, C.F., 2015. The effects of service quality dimensions and passenger characteristics on passenger’s overall satisfaction with an airport. Journal of air transport management 44.45, 77-81. Bezerra, G.C.L., Gomes, C.F., 2016. Measuring airport service quality: a multidimensional approach. Journal of air transport management 53, 8593. Bogicevic, V., Yang, W., Bilgihan, A., Bujisic, M., 2013. Airport service quality drivers of passenger satisfaction. Tourism Review 68.4, 3-18. Eboli, L., Forciniti, C., Mazzulla, M., Calvo, F., 2016. Exploring the Factors that Impact on Transit Use through an Ordered Probit Model: the Case of Metro of Madrid. Transportation Research Procedia 18, 35-43. Eboli, L., Mazzulla, G., 2009. An ordinal logistic regression model for analysing airport passenger satisfaction. EuroMed Journal of Business 4.1, 40-57. ENAC - National Agency for Civil Aviation, 2014. Qualità dei servizi nel trasporto aereo: le Carte dei servizi standard per gestori aeroportuali e vettori aerei. Fodness, D., Murray, B., 2007. Passengers’ expectations of airport service quality. Journal of Services Marketing 21.7, 492-506. Horonjeff, R., McKelvey, F.X., Sproul, W.J., Young, S.B., 2010. Planning and Design of Airports. McGraw-Hill, New York. Kennet, R.S., Salini, S., 2012. Modern Analysis of Customer Surveys: with Applications using R. Wiley, Chichester. Liou, J.J., Tang, C.H., Yeh, W.C., Tsai, C.Y., 2011. A decision rules approach for improvement of airport service quality. Expert Systems with Applications 38, 13723-13730. Liu, D., 2016. Measuring aeronautical service efficiency and commercial service efficiency of East Asia airport companies: an application of network data envelopment analysis. Journal of air transport management 52, 11-22. Long, J.S., 1997. Regression Models for Categorical and Limited Dependent Variables. SAGE Publications, Thousand Oaks. Pantouvakis, A., Renzi, M.F., 2016. Exploring different nationality perceptions of airport service quality. Journal of air transport management 52, 90-98. Popovic, V., Kraal, B.J., KIRK, P.J., 2009. Passenger experience in an airport: an activity-centred approach. IASDR 2009, 18-22 October 2009. Korean Society of Design Service, Coex, Seoul. S.A.CAL. - Calabrian Airport Company, 2017. Service Charter Guide. Yeh, C.-H., Kuo, Y.-L., 2003. Evaluating passenger services of Asia-Pacific international airports. Transportation Research Part E 39, 35-48.