Factors Determining Customer Satisfaction with Counter Services of Local Government Oces: On-Site Survey Examining Service Quality at Higashihiroshima City Hall, Japan

This study empirically investigated the major factors determining customer satisfaction with counter services in local government oces in Japan, comparing them with the results of similar surveys conducted by the author in 2020. This study considers visitors’ direct experience of counter services through an on-site survey conducted at a municipal government building, referring to (1) customer satisfaction research in various countries, especially the expectancy disconrmation model; (2) the SERVQUAL model; and (3) subsequent methods that emphasise the customer’s direct experience. SERVQUAL, as the name suggests, is a ve-group scale for evaluating service quality with respect to customer satisfaction. However, in actual organizational settings, ve groups of questions may be burdensome to respondents and may reduce the collection rate. The author’s research involves conducting on-site surveys on customer satisfaction or service quality of the local government oce using an original three-group questionnaire scale and tailoring the question items to a scale for Japanese local administration settings. The surveys comprise three categories of questions, referring to (1) customer satisfaction research in various countries, especially the expectancy disconrmation model; (2) the SERVQUAL model; and (3) subsequent methods that emphasise the customer’s direct experience. A total of 1,000 questionnaires were distributed over four weekdays at the City Hall. As of September 27, 2021, there were 537 respondents (response rate: 53.7%). The regression analysis indicated that the independent variables of group C (service delivery) were most inuential, followed by those of groups B (human-related factors) and A (oce hardware) (adjusted R 2 value: .58). Service delivery, human factors, and hardware aspects are important, in that order, to ensure customer satisfaction with the City Hall’s counter services. This is consistent with the results of the author’s Internet survey of users of Osaka City Government’s ward oces. Focusing on the second principal component of the service delivery category (C), promptness and short waiting times do not necessarily affect customer satisfaction positively.

question group of factors into synthetic variables using principal component analysis, and demonstrate how the most in uential factors by multiple regression analysis using these compound variables is examined. SERVQUAL, as the name suggests, is a ve-group scale for evaluating service quality with respect to customer satisfaction. However, in actual organizational settings, ve groups of questions may be burdensome to respondents and reduce the collection rate. The author's research involves conducting on-site surveys on customer satisfaction or service quality of the local government o ce using an original three-group questionnaire scale referring to the similar simpli ed framework of three groups of question items applied in Malaysia by Mansor and Razali (2010) and tailoring the question items to a scale for Japanese local administration settings.
The study of customer or citizen satisfaction in public administration in Japan was derived and developed from the research and practice of policy evaluations and government evaluations with an awareness of the issues of local government reform. Mie Prefecture, in central Japan, implemented a government evaluation system during the 1990s. Local governments across the country followed suit, implementing government evaluation systems. This movement in uenced the central government to enact a law mandating all ministries to carry out three types of policy evaluations-performance assessment, program evaluation for constructing infrastructure, and general evaluation with several perspectives-in 2002. Several studies using surveys have indicated that large municipalities such as cities are more assertive in incorporating policy evaluations into local governments' operational activities than small ones such as towns and villages (Author 2015;Tabuchi 2010). Evaluation activities that can be applied to individual concrete situations are now required, and these can be developed through the improvement of more speci c operations and introduction of more speci c methods of evaluation practice (for the current status of government evaluations in Japan, see Author 2020a). These circumstances necessitate that Japanese municipalities apply evaluation activities to speci c conditions through the improvement of more speci c operations and the introduction of new methods. More diversi ed methods are required for speci c organizational needs, such as program satisfaction surveys targeting citizens and the use of logic models. Sesaku Manzoku-do Chosa, that is, the Program Satisfaction Survey-also known as the 'citizen satisfaction survey'-focuses on the level of importance of and satisfaction with all programs and analyses them in four quadrants. Many local governments in Japan, including that of the city of Higashihiroshima, implement this type of survey regularly. As the survey targets municipal residents, the residents' perceptions and behaviours related to a wide range of local government programs can be collected by mail. Based on the respondents' answers, each program is placed in one of the four quadrants decided by the two elements of importance and performance.
Subsequently, more speci c surveys of customer satisfaction emerged in the form of on-site counterservice satisfaction surveys to improve the operations of government o ces. For example, Kumano Town in Hiroshima Prefecture has focused on customer satisfaction with counter services. They implemented an on-site survey in 2009 (Kumano Town 2009) and have been conducting similar surveys regularly.
Moreover, the city of Ogaki, in Shiga Prefecture, performed a questionnaire survey of 2,000 residents by mail to help consider the rebuilding of a City Hall (Ogaki City Government 2014). The response rate was 47.7%. The residents were questioned about their experience during a previous visit, the purpose of that visit, the means of transportation used, and their experiences with the counter service.
This study used an academic approach to examine the determinants of satisfaction with counter services. In other words, the focus is on the counter service among the services provided by local governments. This study involved a multi-group questionnaire survey of residents, and it examined the factors affecting the overall customer satisfaction level with the counter service statistically using selective multi-group principal component regression analysis (Takahashi and Kawasaki 2019), which combines principal component analysis and multiple regression analysis. Everitt (2005) mentions that in the method of applying principal component analysis, multiple question items can be grouped into several categories before the combined principal component scores of explanatory variables are used in the multiple regression analysis. Thus, the use of the principal component score can weaken the multicollinearity problem between each question item used in the regression analysis. In addition, this study compares the results of the survey conducted by the authors at Higashihiroshima City Hall with those implemented at the Kurose branch o ce in 2020 and the Internet survey of other cities. Based on Mansor and Razali (2010), the questionnaire includes three groups of factors: hardware, including physical aspects such as buildings and access; software, including staff response, among others; and service delivery. Hardware, which was relatively unimportant in relation to the cost explanatory variables in the regression analysis of the Kurose branch survey results, is examined in more detail in this survey.
Research on policy evaluation and administrative evaluation in Japan has focused mainly on the introduction of overseas evaluation theories and the examination of advanced examples of domestic practice, such as the evaluation system for administrative projects. However, research from the perspective of speci c methods for improving actual evaluation practices based on overseas evaluation theories has not progressed much. This study makes it possible to grasp the factors that determine the level of satisfaction with counter services and to examine the constituent items of each group to understand those aspects that lead to the improvement of customer satisfaction with government o ce counter services scienti cally and academically.
This research is unique in that it provides suggestions for practical use through on-site research at Higashihiroshima City Hall. For this purpose, this study applied the method adopted by research in the business eld investigating private corporations. Noda (2013Noda ( , 2014Noda ( , 2019  , which perform almost all of the same functions that the old towns did before the municipal merger, except for citywide functions such as urban planning. Citizens can carry out many administrative procedures at nearby branch o ces even after the merger. Accordingly, a survey was conducted in the Kurose branch o ce in August 2020, and another was performed at Higashihiroshima City Hall in August 2021 as a part of the present study. Author (2021) presents the results of an Internet survey of ward o ces in Osaka city. Similarly, this study attempts to discover the determinants of satisfaction of the local government's counter services by referring to the method of customer satisfaction surveys targeting private companies. There are many municipalities, including the city of Higashihiroshima, that survey citizen satisfaction with programs and focus on the level of importance and satisfaction with each program. By contrast, there are only a few municipalities in Hiroshima Prefecture, such as Kumano Town, that survey visitors to the o ce, focusing on customer satisfaction with counter services. This study seeks to identify the primal factors that in uence the level of satisfaction with counter services. To explore the determinants of customer satisfaction with the counter service and to compare it with the survey already conducted in 2020 at the Kurose branch o ce, Higashihiroshima, the current study was performed at Higashihiroshima City Hall, located at Saijo Sakae-Machi, in the Saijo area. This study assumes three groups of factors composed of multiple questions: (A) hardware, (B) software, and (C) service delivery at the City Hall, referring to Mansor and Razali (2010), who also apply three similar groups of question items.

Literature Review
Research publications that are closely related to the ndings at the City Hall reported in this article include Author (2020b) and Author (2021). This section rst discusses these outlines and how they relate to the results and implementations of the current study. Thereafter, it presents an overview of the customer satisfaction research in the public sector.
After distinguishing the differences in the administrative evaluation concept between Japan and the United States, Author (2020b) analyses the results of a questionnaire survey of towns and villages in Japan that has been widely administered by Japanese local governments since the 1990s. The results indicated that there is a large difference in the implementation status between the larger city governments and the small municipalities, which face the challenge of a lack of resources such as expertise and personnel because of scal constraints. Author (2021) summarises the results of an online survey of visitors to the ward o ces of Osaka, which was conducted to investigate the determinants of satisfaction of citizens with the o ces of local governments. This study compares its results with a similar survey conducted in the Kurose branch o ce, Higashihiroshima, in August 2020. Unlike the survey in Kurose, which emphasised the importance of human factors such as the response of staff, in Osaka, the service delivery factor had the strongest effect on the visitors' overall satisfaction with the counter service.
Customer satisfaction research in the public sector has referred to the accumulation of research on marketing by private corporations. Oliver and Winer's (1987) prominent study pioneers this research, focusing on the discrepancy between expectations and perceptions of performance afterwards as determining factors of customer satisfaction. Using their view as the foundation of analysis, Zeithaml (1988) treats the intrinsic and extrinsic attributes as in uential factors of perceived service quality (4). In this study, 'hardware' corresponds with Zeithaml's extrinsic attributes.
Later, in 2010, Oliver published his theories and perspectives on customer satisfaction in a book entitled Satisfaction, and Oliver (2015) updated the book's content in its second edition. Following Oliver's expectancy discon rmation theory, many public administration studies considering citizens as customers have been published in academic journals (Van Ryzin 2004, 2006Van Ryzin and Immerwahr 2007;Van Ryzin et al. 2004). Morgeson and Petrescu (2011) analyse citizen data on six federal government agencies and nd that the factors that contribute to satisfaction and trust in government include perception of the quality of service, information provided by the government, demographic factors, citizen expectations, and e-government adoption. Further, an empirical study by Morgeson (2012) adds respondents' political ideology, party identi cation, and overall trust in the federal government as other factors determining citizen satisfaction. Kelly (2003) examines the relationship between overall customer satisfaction and the re and police services' performance in local public administration. Kelly and Swindell (2002) also examine the relationship between the indicators of actual program performance and the overall customer satisfaction of local governments statistically.
The SERVQUAL model explained by Parasuraman, Zeithaml, and Berry (1988), following Oliver's theory, focuses on the exceptional gap between expectations and actual performance. This model had a strong in uence on subsequent research on measures of service quality and customer satisfaction. Shortly after this study on SERVQUAL, other research has been conducted on SERVQUAL, including Carman (1990), Reidenbach and Sandifer-Smallwood (1990), and Finn and Lamb (1991). Since then, many studies have continued to examine the ve elements of factors in uencing service quality. The identi ed proponents of SERVQUAL are tangibles, reliability, responsiveness, assurance, and empathy. Each element comprises multiple items. The 'tangibles' category of the model is equivalent to the hardware category in the current study. Assurance, empathy, and responsiveness-the multiple human factor-related elements-of the SERVQUAL scale are equivalent to the software category in this study. Wilson et al. (2021, 89-91)  Customer satisfaction studies after 2010 in the public sector shift toward examining the direct service experience. Studies on patient experience derived from patient satisfaction are an example in the medical eld (Manary et al. 2013;Wolf et al. 2014). The Patient Experience Journal, specialising in this eld, was launched in 2014. Further, the Japan Patient Experience Research Association developed a standardised questionnaire index on patient experience in the Japanese language, tailored for medical practice in Japanese hospitals. As for the private sector, 1 study discusses positive experiences and customer satisfaction in the Lithuanian hotel industry and applies the value co-creation-dialogue, access, risk assessment, transparency (DART) model in marketing research (Solakis et al. 2021). Solakis, Peña-Vinces, and Lopéz-Bonilla (2017) focus on the four aspects of DART in consumer behaviour in the Greek hotel industry, which is the basis for the aforementioned study. Referring to these recent studies on customer experience in the public sector, including the medical eld, the current study recognises that customers' direct experiences in uence their overall satisfaction with counter services and that expectations are an important factor of customer satisfaction. A series of studies by the author intends to examine whether there are differences in the primal factors of customer satisfaction with government counter services in different environmental conditions using a simpli ed three-factor model.
In the public sector context, Wagenheim and Reurink's (1991) Figure 1 presents a conceptual diagram on the relationships among various elements from a customer service perspective in public administration. Speci cally, the four elements involved in the government's efforts and perceptions related to customer satisfaction lead to satisfaction with internal and external customer service needs (customer service), which in turn creates organizational e ciency and effectiveness. Many studies following this model examine customer or citizen satisfaction in the public sector worldwide. This study applies the simpli ed framework of Mansor and Razali (2010), who conduct research in Malaysia using three question-item categories related to the determinative factors of customer satisfaction with the municipal government's counter service. For service quality evaluation, the current research uses a three-group scale that is simpler than SERVQUAL's ve-group scale. Agus, Barker, and Kandampully (2007) utilise a similar study concept in the Malaysian public sector. They conduct two different surveys targeting customers and managers in government branches and present the co-relationships among service quality dimensions, service performance, and customer satisfaction.
Similar studies have been conducted in neighbouring countries in Southeast Asia. In Indonesia, Surapto (2014) conducts a survey targeting 200 residents from urban villages in South Tangerang. McMahon (2004) carries out a study on citizens' customer satisfaction with three local governments in Australia, implementing a mail-based survey among 1,500 residents in the city of Perth. Hsiao and Lin (2008) in Taiwan and Huque and Hayllar (1999) in Hong Kong conduct other prominent studies on customer satisfaction with government services in Asia. Akinboade, Kinfack, and Mokwena (2012) conduct a similar study in South Africa. As explained above, research on customer satisfaction in the eld of public administration-sometimes referred to as citizen satisfaction-began in Europe and the United States and has spread to Asia and Africa.
[ Figure 1 here] Considering the related studies conducted to date, this study draws implications for research on customer satisfaction by considering visitors' direct experience through an on-site survey held at a government building. Statistical examination of the survey results can reveal the possible determinants for counter services provided at branch o ces of the municipal government.

Research Methods
After research at the Kurose branch o ce and an Internet survey on the Osaka City Government in 2020, a self-administered on-site questionnaire survey was conducted in August 2021 at Higashihiroshima City Hall for citizens who visited and completed their business at this government facility. The questionnaire was four pages long. The survey was conducted over four weekdays, and 1,000 sets of questionnaires were distributed at two places at the City Hall: the main entrance hall and the area around the side of the service entrance.
The researcher and research assistants (third-and fourth-year students of the Public Administration seminar, Faculty of Law) waited in the o ce and guided visitors from the counter. There were ve research assistants in total, with a maximum of three working simultaneously.

Kurose Branch O ce, August 2020
For the questionnaire survey conducted at the Kurose branch o ce of Higashihiroshima, in August 2020, Author (2020a) provides a detailed description of the survey methodology. Customers of the Kurose branch o ce who completed their business received a four-page-long self-administered questionnaire. Researchers were waiting at the desks on the opposite side of the o ce counter. A total of 240 participants completed the questionnaires on the spot. The survey lasted for six business days (August 24-31, 2020). Based on the number of distributed yers, the survey participation rate was approximately 60%. The questionnaire included items based on four concept groups, A-C, and a group of questions related to the dependent variable Y. This framework of the research design is similar to the design of the study at the City Hall presented in this article.

Survey at Higashihiroshima City Hall, August 2021
A self-administered questionnaire survey was conducted at Higashihiroshima City Hall for visitors who had completed their business. The questionnaire was four pages long. The researcher and two research assistants positioned themselves at two places-the entrance hall and the area around the service exit-at the City Hall and asked visitors to cooperate with the questionnaire survey of Hiroshima University, offering them a gift to do so. The authors handed out (1) a letter requesting survey participation, explaining the cooperation agreement between the Higashihiroshima City Government and Hiroshima University, which were working together to create an international research centre in Higashihiroshima; (2) a questionnaire; (3) an envelope for returning the questionnaire; and (4) a ballpoint pen in a paper box, all sealed in an A4-sized paper envelope. Visitors who accepted the questionnaire lled out the form after returning home and mailed it to Hiroshima University with a pre-stamped envelope. The survey was conducted over four weekdays from Monday to Thursday, and 1,000 envelope sets were distributed to citizens visiting the o ce. A list of these questions is presented in Table 1. Figure 1 illustrates an analytical model with the question items for the explanatory variables classi ed into three groups. Table   2 presents the total number of copies distributed by day of the week, and Table 3 presents the total number of copies distributed by time slot in a day (excluding Thursday). The original research plan was to distribute the survey over ve days from Monday to Friday, but distribution went smoothly, and the distribution of 1,000 copies nished a little after 1:00 p.m. on Thursday, August 26. During the period, Higashihiroshima was a target area for priority measures to prevent the spread of COVID-19 based on the Act on Special Measures concerning In uenza Pandemic (Novel In uenza or Re-Emerging In uenza). From Friday, the day after the actual survey period, Hiroshima Prefecture became an area subject to the declaration of a state of emergency as stipulated by the Act on Special Measures against In uenza Pandemic. As described below, the questionnaire also includes questions on measures to prevent the transmission of COVID-19 in the City Hall. Table 1 Outline of the survey question items related to customer satisfaction   [ Table 3 here] The letter requesting survey participation included in the envelope detailed the scope and signi cance of the survey. The enclosed request letter included an explanation of the signi cance of the research and the agreement between Higashihiroshima city and Hiroshima University on the formation of an international research centre in Higashihiroshima. The letter also included the telephone number and e-mail address of the author as the principal investigator, so that the visitors could inquire about any points that they might have been unclear about.
To compare the results of the survey at the Kurose branch o ce and the survey of users at the ward o ce in Osaka, both in 2020, with the results of this survey, the part of the questionnaire regarding the satisfaction level was left unchanged. Moreover, in the survey considered in this article, the three question items for the overall satisfaction of the dependent variable are convenience of the service, satisfaction with the experience, and others' recommendation of the service, based on similar surveys in the private sector. The rst change made was to add a question on countermeasures against COVID-19 infections in the hardware of the government building (Group B) in response to a request from the City Hall. Next, in

Results
A total of 1,000 sets of questionnaires were distributed during the four-day survey period. Table 2, displaying the number of questionnaire sets distributed each day, shows that the fewest questionnaires were distributed on Thursday, August 26, 2021, whereas the most were distributed on Monday, August 23. The distribution of the scheduled number of questionnaires was completed after 1:00 p.m. on August 26, a day earlier than planned. By September 27, 2021, 537 surveys had been collected, with a response rate of 53.7%. One of the questionnaires was largely un lled and contained only attribute data. The rest of the questionnaires contained valid answers to many of the questions.
To avoid multicollinearity in the regression analysis, the author rst summarises the explanatory variables into composite variables using principal component analysis for each group. Figure 2 depicts the results of the principal component analysis for questions in category Y. As a new column, the principal component scores for the rst and second components were appended to the dataset as variables ZY1 and ZY2. Figure 2 also shows that YQ1 (Q8_1), which is depicted at the top of the gure, had considerable effects on the second principal component-'convenience and recommendation to others'-whereas the rst component was interpreted as 'experience satisfaction'.
relation to the previous point, questions about lighting in the building and the non-smoking/smoking environment from the hardware group were removed. In addition, the question about 'time (shortness) to nish errands', which was included in group B in the previous survey, was moved to one of the question items of group C about service delivery (CQ2). CQ5, on the smoothness of service delivery, was also added in this survey. The other three questions in Group C are the same as those in the previous two surveys. In addition to the three groups of questions related to causal factors of customer satisfaction as well as question items corresponding to components of overall customer satisfaction, which are common to the 2020 survey at the Kurose branch o ce and the Internet survey for Osaka and have been surveyed continuously, questions on visitors' attributes, errands to be done at the o ce, and transportation access to the o ce were also included. In particular, questions about the sogo madoguchi (general counter), which has become a keyword for improving administrative counter services in Japan, were added. In Japanese public administration, the problem of taraimawashii, or the need to visit multiple counters instead of just one to complete errands, has often been pointed out as a major problem in bureaucratic work (oyakusyo shigoto). Saga City and Matsuyama City are experimenting with a general counter system that allows people to complete all procedures from the birth of a child to moving to a local government by simply visiting one counter. The reason that question 10 asks about the use of the general information desk (reception desk) and question 4 asks about the number of counters used before running errands is to examine the necessity of introducing the general counter approach mentioned above.  Table 4 presents the correlation coe cients between the question items in groups A to C and ZY1 of Y category. Question items with correlation e ciencies higher than .40 were selected for use in the principal component analysis for each category. Figure    if we focus on the contents of BQ3 (Q3_9) and BQ4 (Q3_10).
[ Figure 4 here] Similarly, Figure 5 presents the principal component analysis results for the question items in category C. New columns were created for variables ZC1 and ZC2 from the principal component scores for the rst and second components. For the second principal component, CQ2 (Q7_2) and CQ4 (Q7_4) showed opposing characteristics. Further, CQ2 (Q7_2) and CQ5 (Q7_5), related to waiting time and processing time, were in the same direction. For the rst principal component, CQ3 (Q7_3), a question item about problem solving, is the most to the right. Component 1 is interpreted as 'solving the customer's problems and concerns'. Component 2 could be interpreted as 'waiting and processing time in the building' by focusing on the meaning of CQ3 (Q7_3).
[ Figure 5 here] Using these variables, multiple regression analysis was conducted using the variable increase/decrease method in a stepwise approach, with ZY1 as the explained variable ( Table 5). The synthetic variables used were ZA1, ZA2, ZB1, ZB2, ZC1, and ZC2, which were generated from the categories of explanatory question groups A-C. was signi cant at the 5% level and had a negative coe cient. Focusing on the second principal component of the service delivery category it was found that promptness of service delivery and short waiting time do not necessarily affect customer satisfaction positively. As each variance in ation factor was under 2.00, independence among the explanatory variables in the model was maintained.

Discussion
This study conducted at Higashihiroshima City Hall aimed to identify the determinants of customer satisfaction with counter service in the context of municipalities in Japan. There were differences among the results of the surveys conducted at the Kurose branch o ce, ward o ces of Osaka, and Higashihiroshima City Hall. The latter two surveys' results indicated that service delivery of ZC1 was more important than that of ZB1 based on the principal component scores of counter staff responses or human-related factors. The differences between the results of the three studies may be due to differences in the services provided by the three types of government o ces of these ordinancedesignated cities: branch o ce, ward o ce, and the City Hall. Further, the differences between these government o ces are linked to each visitor's attributes and characteristics.

Conclusion
This study has examined citizens' direct experience of counter services through an on-site survey held with visitors of a municipal government building. The regression analysis revealed that service delivery, human factors, and hardware aspects are important, in that order, concerning customer satisfaction with the services of the City Hall. The service-delivery-related component, ZC1, was more important than the hardware of the building (tangibles in SERVQUAL) and human-related factors. Regarding the second principal component of service delivery, the results showed that promptness of service delivery and short waiting time do not necessarily affect customer satisfaction positively. Among studies in the healthcare eld, which is close to the eld of government, some show that waiting time in hospitals has an effect, especially a negative effect, on patient satisfaction (Alrasheedi et al. 2019). By contrast, previous studies have shown that it is not the waiting time itself-or 'objective waiting time'-that affects patient satisfaction but the way the patient waits or their 'waiting environment' (Pruyn and Smidts 1998  Factor loadings related to Y (overall customer satisfaction with counter services) Figure 3 Factor loadings related to A Page 23/24

Figure 4
Factor loadings related to B Figure 5