Evaluating TransJogja Service Quality based on User Perceptions and Expectations

Transcription

1 Evaluating TransJogja Servie Quality based on User Pereptions and Expetations Alesander PURBA a, Fumihio NAKAMURA b, Shinji TANAKA, Peamsoo SANIT d, Ryo ARIYOSHI e a Graduate Shool of Urban Innovation, Yoohama National University, Kanagawa , Japan/The University of Lampung, Bandar Lampung, 35145, Indonesia b,, d, e Graduate Shool of Urban Innovation, Yoohama National University, Kanagawa , Japan a purbaynu@gmail.om b f-naa@ynu.a.jp stanaa@ynu.a.jp d sanit-peamsoo-an@ynu.a.jp e ariyoshi-ryo-np@ynu.jp Abstrat: To tale the inreased motorization, Ministry of Transportation (MoT) of Indonesia has promoted pilot ities for land transport improvement. Unfortunately, TransJogja as a new transit system has failed to enourage number of passengers in spite of the inreasing of travel demand from population growth as its servie quality has ontinually delined after over five years of operations. This researh provides a omprehensive tool for measuring the overall transit servie quality, named Heterogeneous Customer Satisfation Index (HCSI), by onsidering different servie aspets. To measure TransJogja servie quality, users ompleted information about 8 servie aspets whih onsists of overall 27 servie attributes. The value of HCSI is 7.22 out of 10 indiating that the servie delivered is about 72 perent suessful in satisfying TransJogja ustomers. More efforts are needed to inrease the level of ustomer satisfation and attrat new users to establish TransJogja as an appropriate transit system within the ity. Keywords: Indonesia, TransJogja, Servie Quality, Servie Attribute, HCSI 1. INTRODUCTION Unlie most ities in developed ountries where over half of ommuting trips are made by ombined modes inluding yling, waling and publi transit, nearly all of ommuting trips in most ities of developing ountries are absolutely dominated by private vehiles even for a short distane. At a ertain point unontrolled private vehile using in this region beomes one of major auses of traffi ongestion and various externalities threatening the sustainability of the ities. These ities do not have a proper mass transportation system to suppress the inrease of motorization in urban areas. Moreover, the attitude of soiety is to use automobile ownership as one of the requirements for soial anowledgement. This has enouraged everyone to have their own private ar and disouraged them to travel by publi transport. In addition, urban growth and sprawl with a poor publi transport networ have heighten the trend of rapid motorization among urban residents in developing ities (Susilo et al., 2007). Almost all major and medium-sized ities of Indonesia, inluding Jogjaarta, are faing ritial onditions, espeially when urban buses are being abandoned beause of poor servie

2 quality and less omfort. Aiming to tale the inreased motorization in Indonesian ities, partiularly motoryles phenomenon, the Ministry of Transportation (MoT) of Indonesia enated deree No 51 of 2007, promoting pilot ities for land transport improvement. The deree mandates the pilot ity andidates to reflet on their ommitments by providing douments delaring their preparedness in terms of institutional apaity, funding apaity, human resoure availability and transportation master plan. From the target of thirty pilot ities by 2014, to date, twenty seven ities have signed a memorandum of understanding with MoT and launhed more than twenty new transit systems, inluding TransJogja of Jogjaarta. It is inevitable that one important aspet influening the sustainability of new urban buses is servie quality performane. Therefore, referring to Kersten (2010), it is neessary to determine whether their ustomers were satisfied. Indeed, the transit servie quality itself is an aspet maredly influening travel user hoies. Users who have a good experiene with transit will probably use transit servies again, whereas users who feel dissatisfied with transit may not use the next time. For this reason, improving servie quality is important for ustomizing habitual travelers and for attrating new users. Moreover, ustomer satisfation represents a measure of ompany performane aording to ustomer needs (Hill et al., 2003). Hene, the measure of ustomer satisfation provides a servie quality measure whih ustomers express their points of view onerning the servies by providing judgments on some servie aspets by means of ad ho experimental sample surveys, nown in the literature as ustomer satisfation surveys. The aspets generally desribing transit servies an be distinguished into two harateristis. First, the harateristi that more properly explore the servie (e.g., servie frequeny), and, the seond harateristi is the less easily measurable terms that depend more on ustomer tastes (e.g., omfort). In the literature, there are many researhes onerning transit servie quality. Examples of the most reent researh are reported in TRB (2003a, 2003b), Joewono and Kubota (2007), Tyrinopoulos and Antoniou (2008), Isei and Taylor (2008) and Eboli and Mazzula (2009). In these studies, different attributes determining transit servie quality are disussed; the main servie aspets haraterizing a transit servie inlude servie sheduling and reliability, servie overage, information, omfort, leanliness, and safety and seurity. Servie sheduling an be defined as servie frequeny (number of runs per hour or per day) and servie time (time during whih the servie is available). Servie reliability explores the regularity of runs that are on shedule and on time; an unreliable servie does not permit user travel times to be optimized. Servie overage examines servie availability in the spae and is expressed through line path harateristis, number of stops, distane between stops, and aessibility of stops. Information onsists of indiations about departure and arrival sheduled times of the runs, boarding/alighting stop loation, tiet osts, and so on. Comfort refers to passenger personal omfort while transit is used, inluding limate ontrol, seat omfort, ride omfort inluding the severity of aeleration and braing, odors, and vehile noise. Cleanliness speas about the internal and external leanliness of vehiles and leanliness of terminals and stops. Safety deals with the possibility that users an be involved in an aident, and seurity onerns personal seurity against rimes. Other servie aspets haraterizing transit servies analyze fares, personnel appearane and helpfulness, environmental protetion, and ustomer servies suh ease of purhasing tiets and administration of omplaints. This researh provides a omprehensive tool for measuring the overall transit servie quality, taing into aount user judgments about different servie aspets. The ity of Jogjaarta has been hosen for the ase study as its ity rans after Bali as the most visited tourist plae as well as it is dominated by university students. In Marh of 2008, TransJogja of Jogjaarta was launhed and quily attrated users beause it offered better servie quality

3 than the regular buses. After a few years of operation, number of passengers has gradually dereased as the servie quality also has simultaneously delined. For example, the frequeny of bus arrival, whih promised every 5 to 10 minutes, now strethes to more than 30 minutes at pea hours. In point of fat, authors propose a syntheti index of overall satisfation, whih an be used easily by both transit ageny and regulatory body for evaluating servie performane of TransJogja. In the next setion, a ritial review of indies for measuring servie quality from a user perspetive is made; observations and remars emerge from the omparison among the indies analyzed. Beause of the disadvantages of the indies reported in the literature, a new index is proposed. The proposed methodology is applied by using experimental data olleted by a ustomer satisfation survey of passengers of a new transit servie alled TransJogja. The results and findings are disussed at the end of the paper. 2. REVIEW OF TRANSIT EVALUATION METHODS It is inevitable, the onept of ustomer satisfation as a measure of pereived servie quality was mostly introdued in maret researh. In this field, many ustomer satisfation tehniques have been developed. The best nown and most widely applied tehnique is ServQual method, proposed by Parasuraman et al. (1985). The ServQual method introdued the onept of ustomer satisfation as a funtion of ustomer expetations (what ustomers expet from the servie) and pereptions (what ustomers reeive). The method was developed to assess ustomer pereptions of servie quality in retail and servie organizations. In the method, five servie quality dimensions and twenty two items for measuring servie quality are defined. Servie quality dimensions are tangibles, reliability, responsiveness, assurane, and empathy. The method is in the form of a questionnaire that uses a Liert sale on the seven levels of agreement/disagreement (from strongly disagree to strongly agree). The ServQual provides an index alulated through the differene between pereption and expetation rates expressed for the items, weighted as a funtion of the five servie quality dimensions embedding the items. Some variations of this method were introdued in subsequent years. For example, Cronin and Taylor (1994) introdued the ServPerf method, and Teas (1993) proposed a model named Normed Quality (NQ). Although the ServQual represents the most widely adopted method for measuring servie quality, the adopted sale of measurement for apturing ustomer judgments has some disadvantages in obtaining an overall numerial measure of servie quality. In fat, to alulate an index, the analyst is fored to assign a numerial ode to eah level of judgment. In this way, equidistant numbers are assigned to eah qualitative point of the sale; this operation assumes that the distanes between two onseutive levels of judgment expressed by the ustomers have the same size. A number of both national and international indies also based on ustomer pereptions and expetations have been introdued in the last deade. For the most part, these satisfation indies are embedded within a system of ause and effet relationships or satisfation models. The models also ontain latent or unobservable variables and provide a reliable satisfation index (Johnson et al., 2001). The Swedish Customer Satisfation Barometer (SCSB) was established in 1989 and is the first national ustomer satisfation index for domestially purhased and onsumed produts and servies (Fornel, 1992). The Amerian Customer Satisfation Index (ACSI) was introdued in the fall of 1994 (Fornel et al., 1996). The Norwegian Customer Satisfation Barometer (NCSB) was introdued in 1996 (Andreassen and Lervi, 1999; Andreassen and Lindestad, 1998). The most reent development among these indies is the European Customer Satisfation Index (ECSI) (Elof, 2000). The original SCSB model is based on ustomer pereptions and expetations regarding produts or

4 servies. All the other models are based on the same onepts, but they differ from the original regarding the variables onsidered and the ause and effet relationships introdued. The models from whih these indies are derived have a very omplex struture. In addition, model oeffiient estimation needs of large quantities of experimental data and the alibration proedure is not easily worable. For this reason, this method is not very usable by transit agenies, partiularly for monitoring servie quality. More reently, an index based on disrete hoie models and random utility theory has been introdued. The index, named Servie Quality Index (SQI), is alulated by the utility funtion of a hoie alternative representing a servie (Hensher and Prioni, 2002). The user maes a hoie between the servie habitually used and hypothetial servies. Hypothetial servies are defined through Stated Preferenes (SP) tehniques by varying the level of quality aspets haraterizing the servie. Habitual servie is desribed by the user by assigning a value to eah servie aspet. The design of this type of SP experiments is generally very omplex; an example of an SP experimental design was introdued by Eboli and Mazzulla (2008b). Afterwards, Fu and Xin (2007) introdued Transit Servie Indiator (TSI) as an alternative measure for the servie quality of a transit system. This indiator inorporated spatial and temporal variations in travel demand and integrated various measures suh as servie headway, servie hours, route overage, and travel time omponents. From extant literature, Chooharuul and Sriroongvirai (2013) analyzed ustomer satisfation of Bango mass rapid transit passengers by applied fator analysis and strutural equation modeling. A more diret measure for servie quality evaluation is provided by an overall index, often alled Customer Satisfation Index (CSI) (Hill et al., 2003). The CSI represents a measure of servie quality on the basis of the user or onsumer pereptions on servie aspets expressed in terms of importane rates, ompared with user or onsumer expetations expressed in terms of satisfation rates. The CSI plugs the gap of ServQual beause it is based on judgments expressed by ustomer aording to a numerial sale. Compared to all the desribed indies, CSI is based on a simple proedure, fully desribed in the next setion, whih allows the index to be easily alulated by transit operators. The CSI does not tae into aount the heterogeneities among user judgments. To the ontrary, the index proposed by the authors provides an overall servie quality measure introduing the dispersion of the importane and satisfation rates among users. 3. METHODOLOGY 3.1 Customer Satisfation Index The methodology adopted in this researh aims to obtain a onise indiator that provides overall measures of servie quality by onsidering different servie aspets. The indiator an be alulated on the basis of user judgments expressed by a numerial sale; this ind of sale has some advantages as ompared to the sales with points desribed by means of words (e.g., Liert and verbal sale) beause it allows quantitative tehniques of analysis to be applied. To measure ustomer satisfation, different numerial values an be used, generally from 1 to 3, from 1 to 5, from 1 to 7, from 1 to 9, et. The adopted sale an also have an even number of levels, for example, the traditional numeri sholasti sale omposed of points from 1 to 10. As mentioned above, this researh fouses on CSI, whih is alulated by means of the satisfation rates expressed by users, weighted on the basis of the importane rates, aording

5 to the following formula: N CSI= [ ] = 1 S. W (1) where S : is the mean of the satisfation rates expressed by users on the servie quality attribute W : (importane weight) is a weight of the attribute, alulated on the basis of the importane rates expressed by the user. Speifially, the importane weight is the ratio between the mean of importane rates expressed by users on the attribute and the sum of the average importane rates of all the servie quality attributes: W = N I = 1 I (2) Ideally, CSI portrays good measure of overall satisfation beause it summarizes the judgments expressed by users in respet of various servie attributes in a single sore. Therefore, the more aurate the seletion of the attributes is the more aurate the measure of the overall satisfation will obtain. For this reason, the seleted attributes should haraterize the servie aspets deeply as well as reflet the urrent serving situations as lear as possible. However, when all the importane sores are lose to a ertain value, the importane weights are similar, and then the CSI value is lose to the average of all the satisfation sores. In this eventuality, CSI does not give any additional information ompared to the indiator alulated by onsidering only the satisfation sores. In addition, the average importane sores result from the rates expressed by a sample of ustomers, whih an be very heterogeneous; the dispersion of the rates an be represented by the variane or the standard deviation from the mean. In the same way, the satisfation rates an be very heterogeneous among users. These heterogeneities annot be taen into aount in the CSI alulation. To overome this la, importane weights an be orreted aording to the dispersion of the importane rates from the average value. Analogously, satisfation sores an be orreted aording to the dispersion of the satisfation rates from the average value. These adjustments have been introdued for alulating a new indiator, named Heterogeneous Customer Satisfation Index (HCSI). From a mathematial point of view, HCSI is alulated by the following formula: N HCSI = [ S W ] = 1. (3) where, S : is the mean of the satisfation rates expressed by users on the attribute, orreted aording to the deviation of the rates from the average value W : is the weight of the attribute, alulated on the basis of the importane rates expressed by users, orreted aording to the dispersion of the rates

6 from the average value. Whereas S is alulated by the following formula: S S var( S ) = S. N (4) N S var( S ). = 1 The adjustment fator is alulated as the mean of the satisfation rates expressed by users on the attribute divided by the mean of the average satisfation rates of all the servie quality attributes, weighted on the variane of the satisfation rates. Similarly, W is alulated as the mean of the importane rates expressed by users on the attribute divided by the sum of the average importane rates of all the servie quality attributes, weighted on the variane of the importane rates, aording to the following formula: W = N = 1 I var( I ) I var( I ) (5) The introdution of the variane for adjusting the importane and satisfation rates allows the attributes haraterized by more homogeneous user judgments to be onsidered more signifiant; to the ontrary, the attributes with heterogeneous judgments are onsidered less signifiant. Moreover, the mathematial basis of the HCSI formula is demonstrated by assuming that all the ustomers surveyed gave satisfation sores of 10 out of 10 for every servie harateristi, and the average satisfation sores would all be 10. When the variane of the satisfation judgments expressed by the ustomers tends to zero for all servie harateristis, the mean of the satisfation rates divided by the deviation from the mean of eah attribute would tend to the maximum value of 10, and S would tend to S. Therefore, total ustomer satisfation on all their attributes would produe a satisfation index of perent. 3.2 Data Colletion In order to evaluate TransJogja servie quality, the field survey was onduted in the end of Deember Eight surveyors distributed the questionnaire on board. Surveyors approahed passengers personally to as them to fill in the questionnaire list. Moreover, some passengers were most willing to fill out the questionnaire, but others required more detailed explanation. Approximately, it too around 15 to 20 minutes to fill out both soio-eonomi harateristis information and their opinion onerning TransJogja servie quality. Through fae-to-fae interviews during nine days of servie time as many as passenger points of views were olleted.

7 4. EVALUATION OF TRANSJOGJA SERVICE QUALITY 4.1 Transit Servie Charateristis Authors applied the proposed methodology by onsidering a seleted ase study onerning new transit system in a medium-sized ity of Indonesia. Jogjaarta is loated on the entral island of Java and rans after Bali as the most visited tourist plae and most people visit the ity for the temples of Borobudur and Prambanan. Greater Jogjaarta Speial Region itself is an urban agglomeration within Jogjaarta as a provinial apital. It onsists of three loal governments within the region: one muniipality, that is, the City of Jogjaarta, and two distrits, namely Sleman in northerly diretion and Bantul in southern apital. The area overing 234 m 2 was inhabited by 2.3 million people in Jogjaarta ity's population alone is around 510,108 people within m 2 area (Figure 1). Furthermore, many university students live in Jogjaarta and surrounding areas; approximately 51,000 students and 2,400 leturers attend the University of Gadjah Mada alone. Other big universities with more than 10,000 students are Jogjaarta State University, Islami Indonesia University, and Atma Jaya University, respetively. The analyzed transit system is a new urban bus servie initiated by MoT and Jogjaarta loal government to promote a smart bus-based urban transport system employing features suh as air-onditioned busses, sheduled servies, designated shelters and a smart ard tieting system. Figure 1. Map of Jogjaarta ity

8 A new feature of TransJogja is the introdution of buy the servie liensing sheme, or well-nown as gross-ost ontrat, as an attempt to abolish the daily sublet revenue sharing system between the bus owner and bus operators that is ommonly applied. In this new sheme, the government sets up servie standards to be fulfilled by the operator. The servie provided by the operator will be ompensated based on ilometer-traveled as agreed in the initial ontrat. A brief ity and transit servie harateristis are presented in Table 1. Table 1. City and Transit Servie Charateristis Urban Area Charateristis Area (m²) 32.5 Population (people-2013) 510,108 Provine Jogjaarta Speial Region Provinial apital Jogjaarta Physial Measures Regulatory Framewor Operational Performane *Dirgahayani (2012) 4.2 Users Charateristis PT: Para-transit Year of implementation Number of fleets Number of routes Bus apaity Average length/route Number of bus stop/route Dediated lane available Regulator Bus operator Bus provider Approah to ompetition Other modes within the ity Way of payment Multimodal integration Servie time Daily riderships Load fator (%) Headway (minutes) Average speed (m/h) % Fare subsidy % Fare box revenue* No Prov. Transportation Offie Consortium MoT, provine, onsortium gross ost bus, PT, rishaw ash/ard at bus stop Airport 6 a.m. to 9 p.m. 16, Though the population is not signifiantly spread between male and female, the majority of the habitual transit users is female (57% of the sample). Most of the interviewed users are singles (62%) and age is between 21 and 40 years (38%). As previously predited, student is the largest group of interviewed respondents (57%); in addition, more than a third of transit users live outside muniipality (37%) indiating a high urban sprawl of Jogjaarta ity. Then, most of the respondents have preferene to use the new transit to support their mobility (48%). Moreover, about 59% of the sample are belongs to a middle lass of family inome and about 24% are a lower lass; the lasses of inome refer to the net monthly inome of the family unit, expressed in Rupiahs. The more detailed harateristis of soio eonomi are reported in Table 2.

9 Gender Marital status Age Plae of living Family members Employment Reason for maing use of TransJogja The way to reah bus stop Family inome level (IDR) Table 2. Soio Eonomi Charateristis Numbers male 105 female 141 married single up to 20 years from 21 to 40 years from 41 to 60 years over 61 muniipality area outside the muniipality up to two persons from three to four persons over five persons student ivil servant private ompany entrepreneur others (pensioner, housewife, et.) did not own any ar prefer to mae use of new transit unable to drive waling par and ride others (drops by family member, et.) up to 1 million from 1 million to 2.5 million from 2.5 million to 5 million over 5 million Perentage Users Pereptions and Expetations Aiming to get the measure of TransJogja servie quality, users ompleted information about 8 servie aspets whih overall onsists of as many as 27 servie attributes initially developed in the previous study onduted by authors (Purba et al., 2014). Subsequently, the respondents desribed a rating of importane and a rating of satisfation on eah attribute on a 10-point sale. Firstly, respondents are ased to rate the satisfation of all servie attributes, ranging from 1 totally dissatisfied to 10 totally satisfied (2 very dissatisfied; 3 dissatisfied; 4 somewhat dissatisfied; 5 less dissatisfied; 6 less satisfied; 7 somewhat satisfied; 8 satisfied; 9 very satisfied). Seond, respondents are ased to rate the importane of all servie attributes, ranging from 1 no important at all to 10 extremely important (2 very unimportant; 3 unimportant; 4 somewhat unimportant; 5 less unimportant; 6 less important; 7 somewhat important; 8 important; 9 very important). Furthermore, an evaluation of TransJogja servie quality is arried out by examining the rate of satisfation and importane by means of the alulation of the average satisfation and importane sores (Table 3).

10 Table 3. Importane and Satisfation Statistis Importane Satisfation Servie Aspet Servie Attribute No Mean Var. Std. Mean Var. Std. Av bus stop near home and 1 Route destination Charateristis Route harateristis Operation hours Servie Servie frequeny Charateristis Av of shelter and benhes at bus 5 stop Reliability of buses that ome on 6 the speified range Servie Vehile reliability & ompetene of 7 Reliability drivers Length of staying on board Av of map/route at bus stops Av of servie information by phone, 10 mail, internet Information Av of information on buses 11 regarding bus stops, transfer points The ease to submit omplaint, 12 request, opinion Follow up of the omplaint, opinion Av of paring at terminal & ost The ease of payment Quality of air onditioning on bus Comfort Cleanliness of interior, seats & 17 windows Bus overrowding Safety against rimes on buses Safety and Seurity at the bus stops while 20 Seurity waiting for the bus Helpfulness of personnel Fare Tiet ost TransJogja effet on emission TransJogja effet on ongestion Road aident aused by TransJogja Environment Road deterioration aused by 26 TransJogja Effet of TransJogja to the eonomis, soial, ultural & tourism Normally, the attributes with an average satisfation sore lower than 6.0 an be onsidered as ritial servie aspets. Aording to analyzed servies, as many as 17 of 27 attributes had an unsatisfatory average sore; this indiates that most of respondents felt unsatisfied with the TransJogja transit servie. Those seven attributes with the lowest average satisfation sore are the follow up of the omplaint and opinion, road aident aused by TransJogja, availability of paring at terminal and ost, the effet TransJogja on ongestion, road deterioration aused by TransJogja, length of staying on board, and the effet of TransJogja on the eonomis, soial, ultural and tourism. The attributes with the highest average satisfation sores are safety against rimes on bus, seurity at the bus stops while waiting for the bus, and availability of shelter and benhes at bus stop. Further, by exploring the importane rates, the most important attributes for the passengers an be identified. By onerning the average importane sores, all the servie attributes are onsidered important by the TransJogja passengers as they are pratially represented by an average importane sore lose to or higher than 8.0. The attributes with the

11 highest average importane sores are reliability of vehile and ompetene of drivers, reliability of buses on the speified range (servie frequeny), and safety against rimes on buses. These three attributes revealed average importane sores ranging from 9.52 to 9.10, pointing out that the servie attributes are onsidered very important by respondents. Afterwards, both satisfation and importane rates were analyzed also by means of the variane. This type of measures enables the heterogeneity of passengers in the evaluation of servie quality to be verified. By this way, the passenger judgments on expeted quality (rate of importane) are muh more homogeneous than the judgments on the pereived quality (rate of satisfation). In fat, the value of average variane, alulated by onsidering the rates expressed on all the attributes, is 0.72 for the importane and 1.95 for the satisfation. Then, the oeffiients of variation are 9.7 and 23.4 perent, respetively. These figures also indiate that the passenger judgments on the pereived quality are around three times more heterogeneous than the judgments on the expeted quality. Eventually, both satisfation and importane rates expressed by the bus passengers were used for the alulation of the CSI and HCSI as reported in Table 4. In the third and sixth olumn, the weighted sores are reported, whih represent the ontribution of eah attribute to the final value of CSI and HCSI, respetively. Attribute Table 4. Figuring CSI and HCSI Correted Importane Weighted Importane Correted Weighted Weight Sore Weight Satisfation Sore CS Index 5.87 Heterogeneous CS Index 7.22

12 5. IMPLICATIONS OF CUSTOMER SATISFACTION INDEX Based on the findings in the third and sixth olumns of Table 4, the CSI and HCSI values are 5.87 and 7.22, respetively. These figures indiate that mean satisfation rate of TransJogja users vary from less satisfied to somewhat satisfied. In the seventh olumn of Table 3, passenger judgments on pereived quality show that the only 7 of 27 servie attributes are onsidered somewhat satisfied with the transit servie as they are pratially represented by an average satisfation sore lose to or higher than 7.0. The attributes with the highest average satisfation sores are safety against rimes on buses (7.91), seurity at the bus stops while waiting for the bus (7.82), availability of shelter and benhes at bus stop (7.61), tiet ost (7.48), ease of payment (7.35), vehile reliability and ompetene of drivers (7.07), and availability of map/route at bus stops (7.06). Furthermore, there is no an average satisfation sore lose to or higher than 8.0 indiating that, none of the respondents is satisfied with the quality of transit servie provided. On the ontrary, as many as 20 of 27 servie attributes reveal that average satisfation sores ranging from the highest 6.93 (availability of information on buses regarding bus stops, transfer points) to the lowest 3.99 (follow up of the omplaint and opinion). This empirial evidene has demonstrated that most of the servie attributes are onsidered as neither less satisfied nor dissatisfied by respondents. As mentioned previously, users who have a good experiene with transit will probably use transit servies again, whereas users who feel dissatisfied with transit may not use the next time. The projets in terms of finanial and organizational reforms on publi transport industry aiming to reform the existing regulatory poliies and operational praties, are faing multiple obstales to ahieving organizational goals in both operational and management levels. The omprehensive assessment and evaluation required to inlude not only the related operation, finaning, regulations and enforement but also other poliies negleting business and ommerial elements on publi transport industry. These should be done through a series of omprehensive studies overing development of a modern, ommerial and ustomer oriented strategy followed by training and worshops to loal government staffs as well as bus operators on effetive route and servie planning, operational and maintenane system, followed by a demo projet as a showase for better operational system. Analyzing root auses of problems suh as unstable servie frequeny, and other disomfort aspets should be looed at to assist the operator with a finanial sheme in the prourement of new buses and spare parts, for example through redit programs and ooperatives, or through some sort of tax inentives. An established of a poliy reform and deregulation in urban publi transport finaning, should be followed by an improved standard for TransJogja servie and building institutional apability of the provinial government in administering urban transport system and in developing an appropriate urban transport strategy. After more than five years to fous more on the expansion of servie, it is the suitable time for the MoT and loal government to fous solely on sustaining the attrativeness of TransJogja as before Considerable Weight Contribute Satisfation The CSI whih represent a measure of servie on the basis of user pereptions on servie aspets expressed in terms of important rates, ompared with user expetations expressed in terms of satisfation rates. This index is simply alulated as a weighted funtion of the importane weight of eah servie attribute and the mean of the satisfation rate expressed by users. Yet, this index does not into aount the heterogeneity amongst user judgments. To overome this la, some researhers proposed a variation of this index, alled HCSI that

13 orreted the obtained users evaluations aording to the dispersion of the rates from the average value. By onerning the experimental results and findings, the value of CSI is By weighting satisfation and importane sores on the variane, authors obtain a value of HCSI equal to The differene between CSI and HCSI values is as a result of the different ontributions of eah servie attribute to eah index. Obviously, if the variane of importane rates is the same for all the attributes and, ontemporaneously, the variane of satisfation rates has the same value for all the attributes, CSI and HCSI value are equal. In the meanwhile, by examining the weighted sores regarding CSI, it emerges that the attributes giving the highest ontribution to overall satisfation are safety against rimes on buses, seurity at the bus stop while waiting for the bus, vehile reliability and ompetene of driver, availability of shelter and benhes at bus stop, and tiet ost. However, the values of the weighted sores range from 0.15 to 0.31 as shown in Table 4. Of ourse, the other attributes also mae a onsiderable ontribution to the overall value of CSI. Although of the gap between the two indies is 1.35, the weighted sores of CSI are very different ompared to those of HCSI, whih range from 0.05 to The two attributes with the highest weighted sores for HCSI are vehile reliability and ompetene of drivers, and safety against rimes on buses, respetively. Findings indiate the four most relevant servie attributes for CSI are also relevant for HCSI. They are safety against rimes on buses, vehile reliability and ompetene of drivers, seurity at the bus stops while waiting for the bus, and tiet ost, respetively. Furthermore, the availability of shelter and benhes at bus stop and the availability of map/route at bus stops show the onsiderable weights. From the weighted sore results, HCSI an be used as a worthwhile ontribution for quantifying TransJogja servie quality to oversee servie provider performane and fulfill ustomer requirements. Moreover, they an identify their own strengths and weanesses, where they stand in omparison to their ompetitors Environmental Issues The integration of environmental onerns into urban transport poliy has beome important issues over the last few deades. Providing new transit system without being followed by tightening the use of private vehiles may not obtained optimal results, as experiened by the ity of Jogjaarta. As the ity is dominated by university students and as the students have beome wealthier, most of them have bought motoryles as the main mode to support their daily ativities. Ease of redit system and low ost of vehile ownership also ontribute to the rapid growth of vehile ownership. Nowadays, motoryles mae up more than 80 perent of the ity s vehile fleet of 257,000 and have largely displaed other forms of travel inluding for short trips, although these trips are more effiiently done by foot. The presene of TransJogja as learly expressed by respondents is shown a less ontribution to environmental aspet; from 7.22 of HCSI value, the weighted sores of effets on emission, ongestion, and road aident are only 0.10, 0.09, and 0.07, respetively, indiating those servie attributes ontribute less than 5 perent of HCSI value. In other words, after over five years of operation, the new transit system has not been able to attrat more private vehile users and redue the level of emission, ongestion and aidents. Sine Jogjaarta ity is better well-eduated than other medium-sized ities aross ountry, the environmental issues are supposed to be easier ampaigned and implemented by all staeholders, inluding the ative partiipation of the universities.

14 6. CONCLUSIONS In this paper, loal resident pereption and expetation were explored by respondent partiipation in rating the TransJogja servie quality referring to the urrent level of servie. This researh employs overall measures of servie quality by onsidering different servie aspet method to reveal and examine the data. Aording to the findings, HCSI an be onsidered as useful tool for measuring transit servie quality to oversee transit ageny performanes and fulfil ustomer requirements. The index enables the auses generating ustomer satisfation/dissatisfation to be identified and the strategies for improving the servie quality to be defined. HCSI explores heterogeneity into user judgments beause importane and satisfation rates are orreted aording to dispersion from the average value. The value of HCSI is 7.22 out of 10. By onverting this sore into a perentage, the satisfation index shows that the servie is about 72 perent suessful in satisfying TransJogja ustomers. More efforts are needed to inrease the level of ustomer satisfation and attrat new users to use TransJogja. Both regulator and operator should more initiate orretive measures referring to the user's perspetive, more than the government projets in general whih are usually top down setting. For example, length of staying on board s attribute is the lowest weighted sore (0.05) whih is also the least satisfatory servie attribute expressed by respondent. Reduing travel time by the appliation of ITS in advane an be onsidered as the highest priority measure to tale the servie quality problems more than expanding the servie area. Other servie attributes that should be onsidered by both regulator and operator are the follow up of omplaint and availability of paring at terminal, respetively. The weighted sores are 0.08 and 0.07, respetively, indiating that they less ontribute HCSI value as well as the attributes whih are mostly omplained by respondents. These things have beome evident that the funtions, roles and regulatory framewor of urban transport are needed to reform in order to overome the inreasingly ompliated problems. REFERENCES Andreassen, T.W., Lervi, L. (1999) Pereived relative attrativeness today and tomorrow as a preditors of future repurhase intention. Journal of Servie Researh 2(2), Andreassen, T.W., Lindestad, B. (1998) Customer loyalty and omplex servies: the impat of orporate image on quality, ustomer satisfation and loyalty for ustomers with varying degrees of servie expertise. International Journal of Servie Management 9(1), Chooharuul, K., Sriroongvirai, K. (2013) Multivariate analysis of ustomer satisfation: a ase study of Bango s mass rapid transit (MRT) passengers. Journal of the Eastern Asia Soiety for Transportation Studies, 10, Cronin, J.J., Taylor, S.A. (1994) SERVPERF versus SERVQUAL: Reoniling performane-based and pereptions-minus-expetations measurement of servie quality. Journal of Mareting 58(1), Dirgahayani, P. (2012) Poliy elements to upsale the ontribution of urban transit initiative on sustainable urban transport: the ase of bus improvement initiatives in Indonesia. UNU-IAS Woring Paper No. 168, 23pp. Eboli, L., Mazzulla, G. (2007) Servie quality attributes affeting ustomer satisfation for bus transit. Journal of Publi Transportation 10(3),

15 Eboli, L., Mazzulla, G. (2009) A new ustomer satisfation index for evaluating transit servie quality. Journal of Publi Transportation 12(3), Elof, J.A. (2000) European ustomer satisfation index pan-european teleommuniation setor report based on the pilot studies European Organization of Quality and European Foundation for Quality Management, Stoholm, Sweden. Fornell, C., Johnson, M.D., Anderson, E.W., Cha, J., Everitt Bryant, B. (1996) The Amerian Customer Satisfation Index: Nature, purpose, and findings. Journal of Mareting 60, Fornell, C. (1992) A national ustomer satisfation barometer: The Swedish experiene. Journal of Mareting 56, Fu L., Xin, Y. (2007) A new performane index for evaluating transit servie quality. Journal of Publi Transportation 10(3), Hensher, D.A., Prioni, P. (2002) A servie quality index for an area-wide ontrat performane assessment regime. Journal of Transport Eonomis and Poliy 36(1), Hill N., Brierley, G., MaDougall, R. (2003) How to Measure Customer Satisfation. Gower Publishing, Hampshire. Isei, H., Taylor. B.D. (2008) Style versus servie? An analysis of user pereptions of transit stops and stations in Los Angeles -, Paper presented at 87 th Annual Meeting of the TRB, Washington, D.C., January Joewono, T.B., Kubota, H. (2007) User pereption of private paratransit operation in Indonesia. Journal of Publi Transportation 10(4), Johnson, M.D, A. Gustafsson, T.W. Andreassen, L. Lervi, Cha, J. (2001) The evolution and future of national ustomer satisfation index models. Journal of Eonomi Psyhology 22, Kersten W, Koh J (2010) The effet of quality management on the servie quality and business suess of logistis servie providers. Int. J. Qual. Reliab. Manage., 27(2), Parasuraman, A., Zeithaml, V.A., Berry, L.L. (1985) A oneptual model of servie quality and its impliation for future researh. Journal of Mareting 49, Purba, A., Naamura, F., Tanaa, S. (2014) Servie Delivered on New Transit System from users viewpoint (ase study: TransJogja and TransMusi-Indonesia) -, Paper presented at the International Symposium on City Planning, Hanoi, Vietnam, November 6-8. Susilo, Y.O., Joewono, T. B., Santosa, W., Pariesit, D. (2007) A refletion of motorization and publi transport in Jaarta metropolitan area. Journal of IATSS Researh 31(1), Teas, R.K. (1993) Expetations, performane evaluation, and onsumers pereptions of quality. Journal of Mareting 57(4), Transportation Researh Board. (2003a) A Guideboo for Developing a Transit Performane-Measurement System. Transit Cooperative Researh Program, Report 88, Washington, D.C, National Aademy Press. Transportation Researh Board. 2003b. Transit Capaity and Quality of Servie Manual. Transit Cooperative Researh Program, Report, Washington, D.C, National Aademy Press. Tyrinopoulos, Y., Antoniou, C. (2008) Publi transit user satisfation: Variability and poliy impliations. Transport Poliy 15(4),