INTRODUCTION The reasons of choosing the thesis In the process of economic integration, how the increasing competition will affect the banking sector depends in part on the adaptability and the quality of service (QoS) of banks themselves In recent years, aware of the raising role of QoS of banks, many scientists in the field of Banking in Vietnam has organized many researches on this issue These studies mentioned on identifying approaches of the content reflecting QoS of banks and evaluation methods of content that reflects the QoS Each study has a different identify of QoS, but they are based on quality models that has been studied such as a model of SERVQUAL (Service Quality), a model of Function and Technical Service Quality (FTSQ), a BANKSERV model (Bank Service) The assessment methods of QoS, that previous studies use, are mostly an analysis of the primary component, multiple regression analysis and a combination of this two method in structural modeling methods SEM (Structural Equation modeling) The studies have advantages that they have identified the content that reflects the quality and evaluate the correlation between the content reflecting QoS and customer satisfaction when using banking services However, these earlier studies did not realize that despite of the advantages, the models of QoS has some drawbacks as inconsistent measure contents of QoS, the comparison between perceived and expected QoS is very difficult to determine accurately, or many measurement contentsQoS are almost identical Moreover, in the study of QoS, sometimes, bank managers not only want to recognize the content reflecting QoS, determine the correlation between the content that reflects QoS and customer satisfaction, but also want to identify the correlation by visual image and recognize the orders of priority improvements for contents reflecting QoS Research methodology of the correlation with the visual image has benefits to easily recognize the relationship intuitively and the ability to add extra elements (being individuals as a new bank, new banking services or new variables) to purpose of helping the comparison between a new individual and an old individual in the correlation with the parameters of the quality of services This is the content that the thesis will target With the aim of enhancing the technical application of statistical methods for the quality of banking services, I have chosen the thesis: "Researchingapplications of statistical methods in evaluating the quality of services of Vietnamese commercial banks - Illustrations through a specific research at the Joint Stock Commercial Bank of Investment and Development of Vietnam (BIDV)" This thesis will provide an overview of the quality of banking services and the ability of application combining with statistical techniques to assess QoS of commercial banks in general and the Joint Stock Commercial Bank of Investment and Development of Vietnam in particular The aims of research - Overall aim: Researching techniques forapplying of statistical methods in evaluating QoS, then proposing a plan to investigate QoS of commercial banks in general and the Joint Stock Commercial Bank of Investment and Development of Vietnam (BIDV) in particular - Particular aim: + Constructing research models and building a process of assessment QoS for BIDV + Researching techniques for using a combination of statistical methods for a purpose of studyingthe multi-dimensional relationship of QoS and segmenting customer using banking services + Determining the contribution of service quality indicators to customer satisfaction and the orders of priority improvementQoS of BIDV Object and scope of the research - Object: the researching object of dissertation is the statistical methods in the assessment QoS of banks - Scope: The thesis only focuses on the quality of banking services following the view of customers (which does not mention the concept of QoS standing on the supplier perspective) The survey data illustrates the research limited to objects that are individual customers using the service at some branches of BIDV in Hanoi Research Methodology Thesis uses a combination of statistical methods below: - Methods of analyzing available information - Methods of expert - Methods of statistical surveys - Methods of descriptive statistics - Methods of quantitative analysis The contributions of the thesis The thesis has made the following contributions: - Develop models including appropriate indicators to measure QoS and customer satisfaction in the Joint Stock Commercial Banks in Vietnam in general and the BIDV in particular - Develop the measure process and assessmentthe quality of banking services - Purposed applications combining with methods of multidimensional statistical analysis in evaluationthe quality of banking services: (1) analyzing two-step cluster (applied to the mixed data including quantitative and qualitative data) to segment customers; (2) analyzing key components to find the basic component of QoS; (3) analyzing multiple correspondence to represent visually the relationship between QoS with the type of service and client groups - Apply the evaluation method of capacity of Provincial Competitiveness (PCI-2012) to determine the orders of priority improvements QoS Structure of the thesis Besides the introduction and conclusion, the main content of the thesis consists of four chapters: Chapter 1: Several common issues of measurement the quality of banking services Chapter 2: Several statistical methods of assessing the quality of banking services Chapter 3: The situation of assessing the quality of services in the Joint Stock Commercial Bank of Investment and Development of Vietnam Chapter 4: Application of some statistical methods in evaluating the quality of banking services in the Joint Stock Commercial Investment and Development of Vietnam CHAPTER SOME GENERAL ISSUES ON MEASUREMENT OF THE QUALITY OF BANKING SERVICES The content of this chapter focuses on introducing general theoretical issues related to measure the quality of banking services 1.1 General theory of banking services In Vietnam, the Law on Credit Institutions in 1997, the term of "banking services" were also mentioned in Clause and Clause of Article 20, but not having a specific definition and clearly distinction of what banking services is are no formal concept There are many suggestions that all the operations of joint stock commercial banks (credit operations, currency, forex ) are regarded as banking services This view is consistent with the view of the world that defines "banking services are understoodgenerally to include all the operations of the bank for the public and businesses." Banking services have the basic characteristics: (1) Invisibility; (2) Simultaneity; (3) Instability and the difficulty to determine the mass; (4) Products of banking services having highly complementary; (5) An average cost per transaction for personal customer service high; (6) Banking services for individual customers who want to develop good technical infrastructure requiring modern technology; (7) The need for topical points; (8) Banking serves for individual customers have the advantage of economies by scale and scope; (9) Banking serves for individual customers are simple and easy to implement; (10) The risks of banking services for personal customers are low Banking services are varied and suitable for many different types of customers (individual customers or corporate clients) In the scope of research, the thesis only mentioned on some personal banking services including: (1) Money Management; (2) Credit; (3) International payment: (4) Trading currency; (5) Modern Bank 1.2 The quality of banking services The quality banking services is the difference between the expectations of servicing products and the attitude of customers when using such services This can be considered as the most general concept, comprising of fully inclusive sense of service when considering the quality of banking services by the view of customers regarded to be the center The quality banking services includesfive characteristics: (1) Excellence, (2) Characteristic of the product, (3) Supply, (4) Satisfy of demands, and (5) Creation of value The all five characteristics of QoS aim to create customer satisfaction when using the service By the point of view that services are to server customers, considering customers as the center, the quality banking servicescan be measured on a number of basic criteria: (1) The perfection of the service provided; (2) The attitude and responsibility of staff providing banking services; (3) The competitive capacity of services and diversification of services provided; (4) The difference of the service compared with other banks Based on measurement criterions QoS, the researchers have developed models to measure the quality of service in general and its application in the banking sector in particular include: (1) SERVQUAL; (2) SERVPERF; (3 BANKSERV; (4) FSTQ Each model has advantages and disadvantages, so it should not apply one particular model, but selecting appropriately indicators for research in Viet Nam market In this section, the PhD student also mentioned on the relationship between QoS and customer satisfaction Though being closely related to each other, QoS and customer satisfaction have certain differences The purpose of understanding these differences will help reduce confusion between the identification of measurement criterions of QoS and the criterions reflecting customer satisfaction In summary, by analyzing the available documents on the measurement indicator models QoS, the measurement indicator of customer satisfaction and consulting experts, the PhD student proposed the measurement indicator model for the quality of banking servicesand the measurement indicator model of customer satisfaction when using banking services (The contents of the two models are represented in Chapter 3) CHAPTER SEVERAL STATISTICAL METHODS TO EVALUATE THE QUALITY OF BANKING SERVICES The content of Chapter two introduces statistical methods of measurement and evaluation QoS including: (1) A data collection method; (2) A data aggregation method; (3) A data analysis method 2.1 A data collection method In this section, the PhD student focuses on a general introduction to some methods to collect data (directly or indirectly) with the data collection technique (facing interview, writing interview, mobile interview, group interview, observation techniques and analysis of available data) Based on that, the PhD student introduced methods of collecting data that can be applied in researching the quality of banking services which includes collection methods with directly facing interview techniques standard, mobile interviews or group interviews combined with technical observation, taking notes 2.2 A data aggregation method After performing data collection, it needs to conduct processing data before putting in the analysis of that data A mechanism of processing statistical data comprises of data test, data correction, data encryption, data entry, data description that are called statistical aggregation First, checking the data to examine the reasonableness and completeness of data in terms of form, content or check the reliability of measurement Second, the data after being checked and cleaned (corrected data) will be encrypted Testing of data after coding is usually done through building diagonal tables or stem - leaf diagrams, box-plot graphs … Third, the data after being encrypted can be expressed in technical classifications, tables and graphs 2.3 A data analysis method There are many methods of analyzing data such as analyzing variance, regression In this part, the thesis focuses on introducing the content and significance of the multidimensional statistical analysis methods to evaluate QoS The method of multidimensional statistical analysis is the common name of a set of statistical methods applied to sets of data consisting of a large number of original variables and the large number of individuals (the observer) Multidimensional analysis methods include methods of group factor analysis and methods of classification (1) A group of factor analysis methods include:  Principal component analysis: the number of measurement criterions QoS is large, so using analytical methods to extract key components of data is necessary  Multiple Correspondence Analysis: this method is used to locate the service products in the correlation with QoS components and customer segments (2) The cluster analysis method: to analyzing the quality of banking services, the cluster analysis method is used to segment customers and search targeting customer groups In the cluster analysis method, two-step clustering method (Two steps Cluster) is the method having a high applicability in bank because it is able to offer automatic clustering results for the mixed data (including data qualitative and quantitative data) Thus, the method of exploiting data, the exploitation of information and showing different basic outcomes of these two groups of methods, the PhD student aims toward to the ability of application that combines these two methods to research the quality of banking servicesin particular and to describe a set of complex data generally in society and economy -CHAPTER THE SITUATION OF ASSESSING THE QUALITY OF SERVICES IN THE JOINT STOCK COMMERCIAL BANK OF INVESTMENT AND DEVELOPMENT OF VIETNAM (BIDV) Chapter three presents two mainparts: 3.1 The situation of researching the quality of services in the Joint Stock Commercial Bank of Investment and Development of Vietnam (BIDV) In recent years, BIDV has carried out many customer surveys including the investigation for each service product and for each customer group By researches on each product or each customer group, BIDV can organize itself; but by researches integrated customer group and service products or investigating the trademark, the BIDV must employ outside researching organizations to 3.1.1 For the customer survey held by the Bank itself The personal customer survey performed by the BIDV self-organize use the following methods:  Method of information collection:the BIDV uses direct interviews through questionnaires (for VIP customer survey group) or telephone interviews (for the mass consumer groups and investigations of each of the service)  Method of samplechoice: the BIDV implementsa method of survey to chooserandom samples for the investigations of each type of service products, and make a method of convenient sample choice for the investigations of each client group  Method of data processing: A work of processing data for the surveys organized by the BIDV includes gathering votes, checking and correcting questionnaire, encryption, entering the information, testing and aggregating data Data entry and coding at the Bank is done automatically  Method of analyzing data about QoS of individual customers at current BIDV mainly isdescriptive statistical methods (as frequency and percentage) with tables and graphs 3.1.2 For customer surveys implemented by outsources that BIDV employs: With the customer survey project on a large scale as a trademark investigation, the investigation combined customers with service groups, the BIDV often employ outside organizations to perform With these projects,the BIDV will conduct monitoring the implementation process For this investigation, the organization of surveys will provide final results for the BIDV Specifically, in the survey of 2014, the BIDV hired Nielsen to investigate on the evaluation of the BIDV's customers with QoS and compare with other banks This investigation has the advantage of using different attributes to measure the quality of different types of services This helps to assess the details of each type of services However, the disadvantage is to be unable to compare with other types of services at the same quality characteristic Therefore: (1) when to compare the type of service you should use identical the measurement characteristics of QoS; (2) to evaluate details of customer comments for each specific service, it can use particular characteristicsof each service 3.2 General comments on the current situation of the eluvation in the BIDV’S QoS Through studying the status of the QoS evaluation process of BIDV now, this process still exist the following limitations: - First, restrictions on information gathering methods: - Second, the restrictions on the scale: - Third, the restrictions on the questionnaire: - Fourth, restrictions on methods of analysis In summary, through field surveys and consultation with experts in the field of banking and statistics, this thesis build new processes to assess service quality at BIDV 3.3 Develop new processes for BIDV in evaluating QoS Based on assessment of the advantages and disadvantages of the research process on QoS at the BIDV now, the thesis proposed to construct a new research process on QoS include the following steps: + Qualitative analysis: to find models of QoS measurement indicators and models of measurement indicators for customer satisfaction (Table 3.1) + Design the questionnaire: to include the content for clustering customers; QoS measurement indicators, the indicators that measure customer satisfaction + Select of sampling methods and collecting information: a method of information collection can be used is to collect directly through questionnaire Sampling method should be based on the structure of the sample similarity with the general structure form follow a certain criteria Specific, the thesis uses the overall structure with anoccupational criterion to build the structure of the sample + Select the method of information analysis: using multidimensional statistical analysis This process will help the BIDV to undertake research QoS actively and ensure the reliability of data and methods of analysis Table 3.1 - Models QoS measurement indicators and customer satisfaction Model Component Convenience The quality of services Customer satisfaction QoS measurement indicators TT1- Banks have convenient locations of transactions for TT2- Banks have the wide network of agencies TT3- Timing services of banks are suitable and TT4- Systems to access information are easy to use TT5- Transaction procedures are easy and quick HH1- Banks has equipment and modern machines Tangle HH2- Banks has a very nice layout HH3- Banks has the documents, books and photos introducing this service very attractive HH4- Banks has the explicit transaction documents with no mistakes HH5- Bank employees dress polite and impressive HH6- Banks have full facilities PC1- Bank employees resolve satisfactorily for customer complaints Stylish services of PC2- Bank employees has professional qualifications PC3- Bank employees perform services accurately and timely staff PC4- Bank employees are ready to server customers Nhân viên NH sẵn sàng phục vụ khách hàng PC5- Bank employees are very courteous and attentive to customers TXKH1- Bank employees often contacts with customers Interaction TXKH2- Banks organize banquet thank or reward programs for customers every year with TXKH3- Banks has hotline to service customers for 24 hours customers TXKH4- Banks always listens to opinions of customers TN1- Banks protects customer information and transactions The Trust TN2- Banks performs services right from the first time TN3- Banks sends account statement regularly and timely Time for customers to use banking services The number of services that customer are usingat banks Whether customers are willing to go to the bank in the future (continuing to perform transactions at banks) Customers introduce banksto others Levels of general satisfaction with the quality of banking services On the basis of this study on QoS, chapter introduces a studying sample that the thesis has made the investigation according to the right process and this sample will be tested to apply some statistical methods in the next chapter -Chapter APPLICATIONS SOME STATISTICAL METHODS IN EVALUATION OF QUALITY SERVICES OF THE JOINT STOCK COMMERCIAL BANK OF INVESTMENT AND DEVELOPMENT OF VIETNAM The content of chapter four presents a potential application of statistical methods in evaluating QoS with a sample data that is investigated at the BIDV Scope of research only refers to assess QoS for individual banks and properties of QoS measure built on the evaluating point of clients (not evaluate QoS based on the perspective of the provider service) In terms of analysis, chapter four will focus on introducing the ability to apply statistical methods of multidimensional analysis to evaluate QoS 4.1 About the sample According to the comments of the experts, authors choose to define form-based study "sample structures are similarity to the structure of the overall following a criterion." However, according to the statistical results of the BIDV bank, customer structures are generally classified by many different criterions such as occupation, education and income In the scope of the thesis, the author selected distribution sample structure by theoccupational criterion (because this is the criteria helping to create a pattern that matches the overall follow the remaining criteria of education and income) The thesis uses convenient sampling method with individual customers using the services of the BIDV After sending 300 questionnaires, the thesis earned 250 valid votes In the 250 customers surveyed, there are customers who use more than one service By the sample of 250 customers, the thesis checks the satisfaction of each customer for each type of service and storing data of a client for a service type as an independent individual The results are from 250 customers distributed by type of using service to become 363 independent individual 4.2 Assessing the quality of service at the BIDV The orders of assessment QoS at the BIDV is performed as following: 4.2.1 Applying techniques clustering of Twosteps-Cluster to segment customers using the services of the BIDV Use this clustering technique for a data set including qualitative and quantitative variables Results from clustering customers are to divide clients in the sample into three groups: + Group1: customer group with low status + Group 2: customer group with high status + Group 3: group of loyal customers 4.2.2 DeterminingQoS components To determine the QoS components, the thesis conducts in two steps: (1) Check the reliability of the indicator measurement of QoS (by the method of calculating the coefficient Cronbach's alpha) and (2) identify the major components of QoS (by method of main component analysis -PCA) The indicator model of QoS bank bases on originally analysis results composed of 23 indicator parameters After examination the reliability by Cronbach alpha, indicators are disqualified Therefore, there is remaining of 21 indicators being put in the PCA main component analysis to assess the degree of convergence of indicator parameters according to the components Analyzing main components for the first time leads to results that remove variables unsatisfied The model of measurement QoS remains 17 indicators and these 17 indicators will be used to analyze main components for the second time The results showed 17 indicator parameters that are extracted into main components with the total variance extracted for main components (with eigenvalues greater Eigenvalue 1) is 59.802% of satisfaction: * Style component and customer contacting capabilities includes: TXKH2; TXKH1; TXKH3; PC5; PC1; PC3 * Tangible component includes HH4; HH3; HH1; HH5; HH6 * Trust component includes TN2; TN3; TN1 * Convenient component includes TT2; TT3; TT4 4.2.3 Re-encrypt the data for service quality components To encode the Factorscore values of main quality components corresponding with values: satisfaction (Factorscore> 0) and dissatisfied (Factorscore