Service Quality Measurement: Issues and Perspectives

CONTENTS

Acknowledgement

Synopsis

List of Figures

List of Tables

Contents

1 Introduction
1.1 The Background
1.2 The SERVQUAL Metric
1.3 The SERVPERF Metric
1.4 Criticisms on SERVQUAL and SERVPERF
1.5 The Problem Statement
1.6 Objectives of Research
1.7 Significance of this Research

2. Literature Review
2.1 Service Quality Research
2.2 The Service Quality Models
2.2.1 Gronroos' Service Quality Model -
2.2.2 The SERVQUAL Model
2.2.3 The SERVPERF Model
2.2.4 The Three-Component Model
2.2.5 The Multilevel Model
2.3 Discussion on Service Quality Models
2.4 Metric Development

3. Structural Models, Hypothesis & the Metric
3.1 The Structural Model
3.2 Research Hypothesis
3.3 Metric Preparation

4. Research Methodology
4.1 Nature of Research and the Variables
4.2 Research Framework
4.3 Sample Selection and its Rationale
4.4 Development of SERVQUAL/SERVPERF Metric
4.5 Organizational Profile and Demographics
4.6 Reliability, Validity and Practicality
4.6.1 Reliability of the Instrument
4.6.2 Validity of the Instrument
4.7 Practicality of the Instrument
4.8 Data Collection Strategies
4.9 Statistical Procedures
4.10 Types of Data Analysis

5. Analysis and Results
5.1 Descriptive Statistics
5.2 Reliability Analysis
5.3 Distribution Pattern
5.4 Factor Analysis
5.5 Comparisons of SERVQUAL and SERVPERF
5.5.1 Department-wise Comparison
5.5.1.1 Electronics Engineering Department
5.5.1.2 Computer Science & Engineering Dept.
5.5.1.3 Mechanical Engineering Department
5.5.2 Institution-wise Comparison
5.5.3 Dimension-wise Comparison

6. Findings, Implications & Conclusions
6.1 Findings
6.2 Implications
6.3 Scope for future work
6.4 Conclusions

References

Appendices

I Reliability of the Instruments

II SERVQUAL Multi-dimensional Scale

III Service Quality Metric

Acknowledgement

My thanks to my co-authors Dr. Girish Nair, Mrs. Anisa Hussain and Dr. Sayed Aktharsha who have given me timely support in bringing out this book. My thanks are due to the respondents of the survey on service quality. There are many who have helped in this venture of bringing out this book and all the names cannot be mentioned due to limited space but I want them to know that I am grateful to them.

Last but not the least, our sincere thanks to ©Anchor Academic Publishing ein Imprint der Diplomica® Verlag GmbH, Hamburg, Germany for their excellent printing of our manuscript in this form.

I am deeply indebted to:

- Dr. Gopalakrishna B., my guide, who was kind enough to take me as a research scholar and provide me with his invaluable guidance from- start to finish. His unstinting support and encouragement has enabled me to complete this work.
- All my colleagues who supported me and encouraged me during the course of this work.
- To all the respondents of this research without whose inputs the work would not have been possible.
- To all my earlier and contemporary researchers who have given enough arguments to agree or disagree with the two metrics available today for service quality evaluation.
- Finally, my sincere thanks to my wife Chandrika Raj who has helped me to keep myself awake late nights with refreshing cups of coffee thus enabling me to complete this project on time.

Dr. Lewlyn L.R. Rodrigues

Synopsis

1. Introduction

Service quality measurement is in the forefront of management literature since the past few decades, especially post liberalization and globalization. This is because of the tremendous pressure service providers are facing in the domestic and international market. This has lead to the development of several instruments to measure service quality and the two prominent ones are SERVQUAL and SERVPERF.

The use of a particular instrument in a given situation, or to a specific service industry is not clear and is subjective as well as context based. The research literature has no significant evidence on whether both the instruments totally differ in their outcomes, or concur with their each other in some cases. It is in this context this project work has made an attempt to determine the correlation and relationship between the two metrics.

Problem Statement

The SERVPERF and SERVQUAL are the two Service quality measurement instruments widely used in the measurement of service quality in various service sectors such as banking, hospitals, tourism, insurance etc. But the service quality literature indicates that there exists a significant difference in the philosophy of service quality measurement in these two metrics, and also, the results while these two metrics are used need not necessarily match.

According to Parasuraman et al. (1985) the direction of gap between customer Perception and Expectation as measured by SERVQUAL affects Service quality. Accordingly, customers will have ‘Perception’ of high service quality to the extent that their ‘Expectation’ is lower than perceived service performance, and if the converse is true they would perceive low service quality. The second string lead by Cronin & Taylor (1992) argue that unweighted performance based SERVPERF metric, which takes into consideration only customer ‘Perception’ as the basis is a better measure of Service quality.

Hence, the problem identified in this project is: do these two metrics concur in their results, or is there a significant difference in their outcomes as applicable to a given service sector. The study also extends to the correlation between the outcomes of these two metrics and looking into the possibility of drawing implications based on the combined outcome.

Research Methodology

The research is partly qualitative and partly quantitative in nature. Qualitative in the sense that it analyses existing metrics of service quality based on meta-analysis and, through the use of secondary data, discusses the relative importance of both the metrics in service sectors. The research becomes quantitative, as it deals with descriptive statistics and tests various hypotheses using standard statistical tools. Keviat diagrams have been used to identify the service quality gaps.

Significance of this Research

The results of this project work will add to the body of knowledge of service literature in the sense that the empirical study has proved that there is a significance difference in the outcomes of the SERVQUAL and SERVPERF metrics. Further, using the combined outcome of these two metrics and Gap analysis, the implications drawn have identified that Tangibles and Reliability are the two dimensions which have received higher level of Service quality satisfaction by the customers and Empathy and Assurance are the dimensions of least satisfaction. Responsiveness seems to be moderately satisfactory dimension among the five Service quality dimensions. These implications would help the service providers to strengthen the weaker dimensions.

Limitations and Scope for future Research

Even though the sample size is adequate in comparison to that of the universe, the study outcome cannot be generalized as it is based on a focused research in a particular higher educational institute. So, to generalize the results on a concrete basis several samples may have to be drawn at national and international level owing to the fact that educational services today are global in nature. However, as the selection of the respondents and service providers is such that they are from different parts of the country the results can be generalized to a considerably high level. The study can be extended at national level, and if System Dynamics approach is applied, simulation may be possible to study the influence of each of the Service quality dimensions on the overall service of the sector under consideration.

Conclusions

The research has clearly indicated that there is a significant difference in the outcomes of SERVQUAL and SERVPERF metrics. The reliability of the study was 0.8815, which is at adequately acceptable level. Tangibles and Reliability are highly scored, and Empathy and Assurance are least scored, whereas, Responsiveness is moderately scored Service quality dimension. It can be concluded that if meaningful outcome has to be obtained, both these metrics have to be applied to a service sector and based on the combined inference drawn, suggestions should be made for quality enhancement.

1. INTRODUCTION

1.1. The Background

Since the 90s service sectors have taken more stringent measures to enhance their performance and effectiveness in search of achieving ‘differentiation’ in the market. Quality consciousness is further enhanced by the Globalization and Liberalization due to fierce competition. As the quality consciousness among the service sectors increased, the need to measure the quality of service also increased and this prompted for the development of metrics for the measurement of service quality. There have been several attempts made by a group of researchers who have systematically identified the variables that quantify the service quality among which the two most popular metrics are SERVQUAL and SERVPERF.

Among the two popular metrics mentioned above, SERVQUAL is more popular and has a wide application in service quality measurement and has number of applications in a variety of settings. In service quality measurement, Health care applications are numerous (Babakus and Mangold, 1992; Bebko and Garg, 1995; Bowers et al., 1994; Headley and Miller, 1993; Licata et al., 1995; Lytle and Mokwa, 1992; O’Connor et al., 1994; Reidenbach and Sandifer-Smallwood, 1990; Woodside et al., 1989). Other settings include a dental school patient clinic, a business school placement center, a tire store, and acute care hospital (Carman, 1990); independent dental offices (McAlexander et al., 1994); at AIDS service agencies (Fusilier and Simpson, 1995); with physicians (Brown and Swartz , 1989; Walbridge and Delene, 1993); in large retail chains such as: kMart, WalMart, and Target (Teas, 1993); in General Insurance sectors (Gopalakrishna, Varambally and Rodrigues, 2007), and banking, pest control, dry cleaning, and fast-food restaurants (Cronin and Taylor, 1992).

Cronin and Taylor (1992) found that their measure of service performance (SERVPERF) produced better results than SERVQUAL. SERVPERF instrument has also many applications in diversified areas. There are several studies to compare the two instruments and discuss on which one of the two measures Service quality in the true sense. In the light of this background this research is carried out to seek relationship between the two most popular metrics of Service Quality measurement.

1.2. The SERVQUAL Metric

Zeithaml, Parasuraman & Berry (1990) developed SERVQUAL, which was originally measured on 10 aspects of service quality: reliability, responsiveness, competence, access, courtesy, communication, credibility, security, understanding or knowing the customer, and tangibles. It measured the gap between customer expectations and experience. By the early nineties the authors had refined the model to the useful acronym RATER:

1. Reliability - ability to perform the promised service accurately and dependably.2. Assurance - knowledge and courtesy of employees and their ability to convey trust and confidence.3. Tangibles – physical facilities, equipment, and the appearance of personnel.4. Empathy - caring and individualized attention to customers.5. Responsiveness - willingness to help customers and to provide prompt service.

The SERVQUAL instrument consists of a 22-items for assessing service quality based on customer’s perceptions, which is, by his turn, the difference between the customer’s perceived quality and customer expectation. The perceived quality is assessed based on service quality dimensions that correspond to the criteria used by consumers when assessing service quality. There are 10 potentially overlapping dimensions: tangibles, reliability, responsiveness, communication, credibility, assurance, competence, courtesy, understanding/knowing the customer, and access. A more detailed description of those dimensions can be found in Zeithan et al. (1990). Afterwards, these dimensions were reduced to five, namely: tangibles, reliability, responsiveness, assurance, empathy. Using those 10 or 5 dimensions as the evaluation criteria the specification of service quality becomes the gap between customers’ Expectations and their Perceptions (Parasuraman et al, 1985). This performance-expectation model was also adopted by other authors (e.g. Brown and Swartz, 1989).

The SERVQUAL has its detractors and is considered overly complex, subjective and statistically unreliable. The simplified RATER model however is a simple and useful model for qualitatively exploring and assessing customers' service experiences and has been used widely by service delivery organizations. It is an efficient model in helping an organization shape up their efforts in bridging the gap between ‘perceived’ and ‘expected’ service. Nyeck, Morales, Ladhari, and Pons (2002) stated the SERVQUAL measuring tool, “remains the most complete attempt to conceptualize and measure service quality” (p. 101). The main benefit to the SERVQUAL measuring tool is the ability of researchers to examine numerous service industries such as healthcare, banking, financial services, and education (Nyeck, Morales, Ladhari, & Pons, 2002). The fact that SERVQUAL has critics does not render the measuring tool moot. Rather, the criticism received concerning SERVQUAL measuring tool may have more to do with how researchers use the tool. Nyeck, Morales, Ladhari, and Pons (2002) reviewed 40 articles that made use of the SERVQUAL measuring tool and discovered that “few researchers concern themselves with the validation of the measuring tool” (p. 106).

According to SERVQUAL model,

Service quality = Perception - Expectation

1.3. The SERVPERF Metric

Cronin and Taylor (1992) had re-examined and extended the SERVQUAL model with a more focused approach on the conceptualization of the model SERVPERF. This model is basically ‘Perception’ part of SERVQUAL model, which measures Service quality in terms of perceptions of customers based on the performance of service providers.

Hence,

Service quality = Perception

1.4. Criticisms on SERVQUAL and SERVPERF

Francis Buttle (1996) critiques SERVQUAL in the article "SERVQUAL: review, critique, research agenda" on a number of theoretical and operational bases. He particularly notes that SERVQUAL's five dimensions (Reliability, Assurance, Tangibility, Empathy, and Responsiveness) are not universals, and that the model fails to draw on established economic, statistical and psychological theory. Although SERVQUAL's face and construct validity are in doubt, it is widely used in published and modified forms to measure customer expectations and perceptions of service quality.

Lages and Fernandes (2005) in the article "The SERPVAL (Service Personal Values) scale: A multi-item instrument for measuring service personal values" suggests that consumer final decisions are taken at a higher-level of abstraction. Similarly to the SERVQUAL scale, the SERPVAL scale is also multi-dimensional. It presents three dimensions of service value to peaceful life, social recognition, and social integration. All three SERPVAL dimensions are associated with consumer satisfaction. While service value to social integration is related only with loyalty, service value to peaceful life is associated with both loyalty and repurchase intent.

Brown, Churchill and Peter (1993) had expressed their primary focus of critique on the difference score (i.e. perception minus expectation) with psychometric concerns about conceptualization. They come to a general agreement that the SERVQUAL 22 items are reasonably good predictors of service quality in its wholeness, however, they observe that careful scrutiny of the 22 items by and large deal with the element of human interaction intervention in the service delivery.

Sureshchandar et al., (2001), comment in general on both SERVQUAL and SERVPERF ; instruments and claim that these metrics have overlooked some other important factors of service quality like core service, systematization, standardization of service delivery (the non-human element), and the social responsibility of the service organization.

1.5. The Problem Statement

SERVQUAL and SERVPERF metrics are widely in use for the measurement of Service quality, but it is not clear whether their results match. There are a group of researchers who argue that the measure of Service quality is through ‘Perception minus Expectation’, as used in SERVQUAL but another group of researchers consider only ‘Perception’ as the measure. The main problem identified in this research is to carry out an empirical investigation on whether the two metrics concur in their measurement or they differ significantly. The problem statement gives rise to the following research questions:

1. Do both the metrics i.e. SERVQUAL and SERVPERF produce the same outcome while measuring the service quality w.r.t. the five dimensions: Reliability, Assurance, Tangibility, Empathy, and Responsiveness?2. Is there a significant difference in the results obtained by SERVQUAL and SERVPERF?3. If they differ in their measurement, are there some dimensions in which they agree or do they differ w.r.t. all the five dimensions?4. Can the inferences be drawn based on the interpretation of the results on combined evaluation of SERVQUAL and SERVPERF?

1.6. Objectives of the research

With the above problem statement, it is clear that the cardinal objective of this research is to study whether the SERVPERF and SERVQUAL metrics concur in their outcome of service quality measurement, or if there is a significance difference in their outcomes. Following sub-objectives have been formulated to accomplish this main objective:

1. Prepare a SERVQUAL and SERVPERF metric for the measurement of service quality in a chosen service organization.2. Validate the metric for use in the chosen service sector.3. Apply the metric to collect data for Service quality measurement.4. Build hypothesis to test if there is a significant difference in the measurement of service quality as measured by the two metrics and analyze the same.5. Identify the dimensions, if any, w.r.t. which the two instruments concur and attempt to find reasons for the difference.6. Identify weaker dimensions of service quality and make suggestions to improve the same.

1.7. Significance of this Research

The service quality literature is shrouded in mystery regarding which instrument to be used for measurement of service quality, as there are a number of models which describe service quality. The choice of a particular model to a great extent is subjective as the researcher chooses a model which he feels covers most of the measurable parameters. A good number of researchers have applied the available metrics and empirically evaluated the quality of service in various organizations. There are also a sizable number of articles available on criticizing the SERVQUAL and SERVPERF metrics, which are popularly used by service quality researchers. This research throws light on, do these two instruments concur in their results or significantly differ in their results. Hence, the outcome of this research adds to the body of knowledge of service literature. Education sector is considered to be one of the largest service organization post liberalization, and as these two metrics are tested in an educational set-up, the outcomes of this research not only identify the pitfalls but also make suggestions to improve upon the same.

2. LITERATURE REVIEW

Service quality literature is rich in models, hypotheses and criticisms owing to the fact that active research has been in progress since the past two decades. The chronological order of Service quality research has been tabulated in the beginning and the focus of this chapter is on the review of various models proposed to measure service quality and discussions on the two models viz., SERVQUAL and SERVPERF.

2.1. ...Service Quality Research

;The literature review indicates that Service quality research is in the forefront of many countries owing to the fact that today’s global economy is basically a service economy. It is to be noted that almost all the organizations compete to some degree on the basis of service and it is difficult to name even one industry for which service matters are unimportant (Zeithmal et al., 1990). A lot of research is in progress round the globe on issues of Service quality enhancement in service industries to name a few: banking, health-care, transportation, hospitality, hospitals, health care, insurance, education, and tourism. The chronological order of some significant research and the outcome has been listed in Table 2.1.

illustration not visible in this excerpt

Table 2.1: Chronology of Service quality research

Hence, there is a clear evidence that Service quality research is a potential area of research and there has been enough work undertaken so far and the research is still in progress. The above chronological research also indicates that there have been different approaches to Service quality and several tools are available to measure Service quality. There is also a diversified view on measurement issues and there exists evidence to prove that generalization of quality dimensions is not possible and inconsistencies exist when it comes to the relationship between quality management and business performance. The focus of this research is to test whether a significance difference exists in the measurement of service quality when the two basic models namely SERVQUAL and SERVPERF are used. Hence, following key models have been chosen for discussion from the chronological outcomes of service quality research.

2.2. ...The Service Quality Models

;Even though the Service quality research is flooded with models, the focus of this chapter is to discuss the most relevant models which significantly influence service quality and contributes to the parameters of this research. The chosen models have been discussed in the following sections.

2.2.1. Gronroos' Service Quality Model

One of the oldest models in service quality measurement is the model developed by Gronroos developed in 1984. According to this model, the customer's perception of service has two components:

1. Technical quality - What the consumer receives, i.e., the technical outcome of the process2. Functional quality - How the consumer receives the technical outcome; or "expressive performance of a service"

Functional quality is generally perceived to be more important than technical quality according to Gronroos assuming. The Technical quality and Functional quality lead to Expected service and perceived service as shown in Fig. 2.1.

illustration not visible in this excerpt

Figure 2.1: Gronroos' Service Quality Model (Nordic Model)

(Source: Gronroos, 1984)

2.2.2. SERVQUAL Model

Parasuraman, Zeithaml and Berry developed the SERVQUAL model in the 1988. Service quality results from a comparison of perceived with expected performance (Fig. 2.2) and is based on the disconfirmation paradigm, which creeps in from the discrepancies between prior expectation and actual performance.

illustration not visible in this excerpt

Figure 2.2: The SERVQUAL Model

(Source: Parasuraman, Zeithaml & Berry, 1988)

;The ‘Perception’ of service by the customer is based on the ‘Performance’ level of Service quality. Parasuraman et al.’s (1985; 1988) basic model was that consumer perceptions of quality emerge from the gap between performance and expectations, as performance exceeds expectations, quality increases; and as performance decreases relative to expectations, quality decreases (Parasuraman et al., 1985; 1988). Thus, performance-to-expectations “gaps” on attributes that consumers use to evaluate the quality of a service form the theoretical foundation of SERVQUAL.

2.2.3. The SERVPERF Model

;Cronin and Taylor (1992) proposed that perceptions of performance are the only criteria to measure & define Service quality and brought out SERVPERF model (Fig. 2.3).

illustration not visible in this excerpt

Figure 2.3: The SERVPERF Model

(Source: Cronin & Taylor, 1992)

2.2.4.The Three-Component Model

Rust and Oliver (1994) suggested that Service quality is a function of three components viz., Service Product, Service Delivery and Service Environment (Fig. 2.4). The service product is the outcome and the consumer’s perception of the service. The service delivery is the consumption process and any relevant events that occur during the service act. The service environment is the internal and external atmosphere. The service environment is important because it is viewed as an integral role in consumer service perception development (Bitner, 1992).

illustration not visible in this excerpt

Figure 2.4: The Three-Component Model

(Source: Rust and Oliver, 1994)

2.2.5. The Multilevel Model

Dabholkar, Thorpe and Rentz (1996) developed a model which suggests that Service quality follows hierarchical conceptualization (Fig. 2.5) and it follows three levels viz.,

1. Customer’s overall perceptions of Service quality
2. Primary dimensions
3. Sub – dimensions.

illustration not visible in this excerpt

Figure 2.5: The Multilevel Model

(Source: Brady and Cronin, 2001)

2.3. ...Discussion on Service Quality Models

Adoption of a particular Service quality model into a situation is to a great extent subjective as each of the above models given in the preceding sections defines service quality in its own set of parameters. Zhao et al., (2002) state that Service quality is difficult to measure objectively, since services have been described as intangible, heterogeneous and inseparable. So, a group of researchers started working on the comparison of the results obtained by these models. Based on the results they obtained, researchers also started identifying new dimensions, which were not accounted for in a particular model. So, there has been a wealth of knowledge based on the research of service quality literature.

Churchill and Suprenant (1982) established the fact that Service quality was an attitude. ; Gronroos (1982) brought out another model by emphasizing the significance of processes and outcomes in defining Service quality. In this Nordic model, Gronroos proposed the concept of Expected service and Perceived service as a measure of service quality. Holbrook and Corfman (1985) expanded on the concept of an act being performed and defined perceived quality as a global value judgment and they indicated that quality does by its nature seem to express general approval. They brought out three specific dimensions of quality viz., implicit features, humanistic features, and operational features in nature. Further, Maynes (1985) quantified service quality by placing a number on the level of satisfaction.

The literature review on Service quality indicates that SERVQUAL metric dominates the literature and is most widely used (Lai et al., 2007), and has been widely tested for its validity and reliability (Babakus and Boller, 1992; Bolton and Drew, 1991; Brown and Swartz, 1989; Carman, 1990; Cronin and Taylor, 1992, 1994). Although some of these studies failed to support the five dimensional factor structure, Parasuraman et al. (1994) defended the framework on conceptual and practical grounds. Further, Researchers have criticized the SERVQUAL scale for its use of gap scores, measurement of expectations, positively and negatively worded items, the generalizability of its dimensions, and the defining of a baseline standard for good quality (Cronin and Taylor, 1992; Brown et al., 1993; Oliver, 1993).

;Cronin and Taylor (1992, 1994) suggested that service quality be measured using a performance-only index (SERVPERF) instead of the gap-based SERVQUAL scale. They reported that the use of the SERVPERF scale containing only performance items explained more of the variation in service quality than did the entire 44-item SERVQUAL instrument. The SERVPERF instrument contained 21 of the original SERVQUAL performance items. Cronin and Taylor (1992) argued that SERVQUAL confounds satisfaction and attitude. They stated that service quality can be conceptualized as “similar to an attitude”, and can be operationalized by the “adequacy-importance” model. In particular, they maintained that “performance” instead of “performance – expectation” determines service quality and that developed an alternative measurement tool, SERVPERF, which concerns only performance. In their empirical study, SERVQUAL appeared to have a good fit in only two of the four industries examined, whereas SERVPERF had an excellent fit in all four industries. A similar result was obtained from regression analyses.

Cronin and Taylor (1994) argue that SERVQUAL:

1. does not measure either Customer satisfaction or Service quality instead it “appears at best an operationalization of only one of the many forms of expectancy disconfirmation”.
2. does not exhibit construct validity.
3. does not ensure that the dependant measure is performance based.
4. has little empirical and conceptual research support.

On the above grounds they opine that SERVPERF can provide reliable, valid and useful tool for measuring overall service quality levels or attitudes. A group of researchers including Churchill et al. (1993), Carman (1990), Babacus ad Boller (1992) support this argument.

Hence, in consideration of the above discussions it can be very clearly concluded that both of these two instruments (SERVQUAL and SERVPERF) are is use across service quality literature and their relative importance is an issue of analysis.

2.4. ...Metric Development

The SRVQUAL and SERVPERF metrics have Service quality perception influenced by the five dimensions viz. Reliability, Assurance, Tangibility, Empathy, and Responsiveness as per the literature review. The metric used in this research consisted of 22 variables which measured the five dimensions just mentioned (Parasuraman et al., 1988). As the instrument was basically developed for a general service sector, it was to be modified to suit to the requirements of Higher Education settings. The details of the instrument development is given in Chapter 4: Research Methodology.

3. STRUCTURAL MODELS, HYPOTHESIS & THE METRIC

3.1. Structural Model: Difference in SERVQUAL and SERVPERF based measurement of Service quality

illustration not visible in this excerpt

Fig. 3.1: Structural Model: Service Quality Dimensions

The structural model provides the following aim to this research work:

Aim:; To estimate the relationship between the Service quality perception, as measured using SERVQUAL and SERVPERF metrics.

In other words, the Aim is to test the structural model, which leads to the formation of the following hypothesis.

3.2. Research Hypothesis

The purpose of the research hypothesis is to test whether the two instruments under study (SERVQUAL & SERVPERF) produce the same outcome while used for the measurement of Service quality, or is there a significant difference in their outcomes. The hypothesis is tested at two levels viz., Department level and then Institutional level (Overall). There are again three classes at the department level, and hence, the following main and sub-hypotheses have been formulated.

Main Hypothesis:

Research Hypothesis H1a:

“There is a significant difference in Service quality measurement, as measured by SERVQUAL and SERVPERF metrics.”

Null Hypothesis H1o:

“There is no significant difference in Service quality measurement, as measured by SERVQUAL and SERVPERF metrics.”

Sub-hypotheses:

Research Hypothesis H1a-1:

“There is a significant difference in Service quality measurement department wise, as measured by SERVQUAL and SERVPERF metrics.”

Null Hypothesis H1o-1:

“There is no significant difference in Service quality measurement department wise, as measured by SERVQUAL and SERVPERF metrics.”

Research Hypothesis H1a-2:

“There is a significant difference in Service quality measurement institute wise, as measured by SERVQUAL and SERVPERF metrics.”

Null Hypothesis H1o-2:

“There is no significant difference in Service quality measurement institute wise, as measured by SERVQUAL and SERVPERF metrics.”

Research Hypothesis H2a:

“There is a significant difference in individual Service quality dimension measurement, as measured by SERVQUAL and SERVPERF metrics.”

Null Hypothesis H2o:

“There is no significant difference in individual Service quality measurement, as measured by SERVQUAL and SERVPERF metrics.”

3.3. Metric Preparation

The metric used in this research for the collection of quantitative data is a self-administered questionnaire. The reason for choosing this instrument is that it is a relatively systematic and standardized method of collecting data, which lays emphasis on measurement and conversion of data from qualitative to quantitative form. Further, it is evolved from studying sampling population to probability sampling and provides means for simple counting to statistical description and inferential analysis. Finally, this method is considered to be economical and convenient for this kind of research.

The questionnaire has been designed to measure Service quality as defined by SERVQUAL and SERVPERF instruments. They obtain the answers to the research questions and provide the necessary data to test hypothesis.

;Though the literature review and interviews with the teaching faculties of various institutes identified several issues related to quality of service in educational institutes, only those areas specific to this research were selectively chosen to modify the SERVQUAL/SERVPERF metric, so as to fit into educational sector. The problem areas were categorized, simplified and redundancy was eliminated to develop a set of questions for the research questionnaire. Further, while developing the questionnaire the following points were taken into consideration:

- Are the categories of respondents competent enough to provide the necessary information?
- Do the chosen items of each questionnaire truly measure the dimension to which they correspond?

Questions were frames to be uniformly understood by all respondents belonging to different disciplines of engineering. A pilot study was conducted for about 25 students and was reviewed for problems with bias and it was confirmed that no particular question caused any problem of understanding. Simple language was used throughout the questionnaire and all possible ambiguity was eliminated. Moreover, it was ensured that all the questions were effective in obtaining information relevant to hypothesis testing in all the cases. The questionnaire was also subjected to Construct, Content and Criterion validity which has been explained in chapter 4: Research Methodology.

4. RESEARCH METHODOLOGY

This chapter deals with the research methods and the methodology adopted in this research. The validation and factor analysis of the instrument used to collect data has been explained. The chapter also outlines the research framework. The nature of research and the variables involved therein are highlighted in the beginning of this chapter. The methods used in this research are listed. Identification and rationale for the sample selection has been given. Organizational profile and respondents’ profile have also been given. The systematic procedure for performing reliability, validity and practicality test has been described. The best practices incorporated in developing the questionnaire, data collection strategies, statistical procedures, data analysis and limitations of these methods are discussed. Finally, methodological limitations have been discussed and the methods adopted to minimize the same have been listed.

4.1 Nature of Research and the Variables

This is basically a correlational type of research, from the study results of which, inferences are drawn and implications are made. The following dependent variables are the focal areas of research interest:

Service quality in this research is considered to be an independent variable and the dependent variables are: Reliability, Assurance, Tangibility, Empathy and Responsiveness.

Thus, ;

Service quality = f (Reliability, Assurance, Tangibility, Empathy and Responsiveness).

This research focuses on the service quality offered in a higher educational institution, to be more specific an engineering institute. Hence, the service providers are the Management, Faculty, Support staff, and all those who contribute to the overall development of students. The service receivers are mainly the students. Even though the engineering college has a higher goal to be accomplished which includes contributing to the national GDP by enriching the knowledge, skill and attitude of prospective engineers, in the context of this research it is limited to providing service to the future engineers. Hence, the primary source of data is the engineering students. The secondary source of data include teaching and non-teaching faculty, journals, periodicals, newspapers and the response from all the stakeholders of engineering institutes. The primary source of data is used for quantitative analysis and the secondary source for the qualitative analysis.

The method used in this research goes in line with Kerlinger’s (1977) procedure: ;

‘…the theory and method of analyzing quantitative data obtained from samples of observations in order to study and compare sources of variance of phenomena, to help make decisions to accept or reject hypothesized relations between the phenomena, and to aid in making reliable inferences from empirical observations’.

The type of research is partially quantitative and to some extent qualitative. It is quantitative in the sense of being correlational and testing of hypothesis. It attempts to determine whether there exists a significant relationship between two independent variable. This refers to the hypothesis testing (formulated in chapter 3), undertaken in this research. It is qualitative in the sense that it involves the collection of secondary data, which is qualitative in nature, in the form of suggestions and implications, to enhance performance and quality of engineering education.

The data collection is through stratified random sampling method. The data, which is basically qualitative in nature, is converted into quantitative form through Likert type 5-point scale. It is then processed using statistical packages Excel and SPSS version 10.0. The results obtained through analysis are used to test various hypotheses. The gap analysis is undertaken to compare the inter-department Service quality.

4.2 Research Framework

The entire project was carried out in the standard Research Framework shown in Figure 4.1. The following were the various phases involved in the project work.

Phase I: Problem Statement

In this phase of the project the research question was very clearly stated. The problem identified in this research was the concurrence issue of SERVQUAL and SERVPERF metrics. It was decided to seek whether both these metrics produced the same outcome when applied to a Service quality measurement situation or they differed in their perception evaluation (Chapter 1).

illustration not visible in this excerpt

Figure 4.1: The Research Framework

Phase II: Purpose and Objectives

In this phase of the project the research question was very clearly stated. The problem identified in this research was the concurrence issue of SERVQUAL and SERVPERF metrics. It was decided to seek whether both these metrics produced the same outcome when applied to a Service quality measurement situation, or they differed in their perception evaluation. Based on this understanding the objectives were defined (Chapter 1).

Phase III: Research Background

This phase basically involved the Literature Review of the Concepts and Theories underlying Service Quality and the study of Contemporary Research. Through literature review it was clear that the Service quality literature had different theories and metrics to measure the standard of Service quality. The contemporary research was focussed on mainly two instruments i.e. SERVQUAL and SERVPERF. Discussion on each of the model, criticism of the two instruments lead to the development of metric customized for Service quality measurement in Higher education (Chapter 2).

Phase IV: Research Design

In this phase of project the procedure, techniques, and tools & techniques to be used were decided. This being an empirical study it was decided to formulate hypotheses to answer the fundamental research question related to concurrence of SERVQUAL and SERVPERF. The types of data to be collected, sample size, instrument to be used, and methods of data collection was decided. The hypothesis testing was to be through the application of Statistical tools such as t-test and ANOVA. Software to be used was also decided and Microsoft Excel 2007 and SPSS Ver. 10 were chosen based on their versatility and wide application (Chapter 4).

Phase V: Execution

In this phase of the project, the tools and techniques chosen for research were deployed. The two important parts of this phase were collection of Primary Data and application of Statistical tools. An Engineering college was chosen to be the source of primary data and three prominent departments which were well established were selected to be the three models for cross comparison. The students were randomly picked during their free timings and the data was collected. The SPSS was used to test the reliability and type of distribution of data. Accordingly, t-test and ANOVA were applied to test the hypotheses.

Phase VI: Analysis & Results

The concept of Meta-analysis was adopted for the processing of Secondary data. The procedure involved screening of the available data from Journals, Books, Encyclopaedia, Periodicals, Conference proceedings, and Dissertations and collecting the most relevant ones for this study. They key features from these literature were recorded in the chronological order and the findings were listed and contradictions were noted down. This gave insight into the fact that SERVQUAL and SERVPERF had concurrence issues and both the metrics had context based applications. So, the Primary data was to be collected and the analysis was through empirical investigation and statistical procedures. This lead to the selection, modification, validation, application and analysis of the data thus collected. The results thus obtained were analysed department-wise and on the overall basis, and inferences were drawn to accept or reject null hypotheses. Implications were drawn based on these analyses (Chapter 5).

Phase VII: Conclusion

Various conclusions were drawn with reference to Reliability of the study, Distribution pattern, Descriptive statistics, Factor analysis, Rank order correlation, Hypotheses testing, Inter variable correlation, Gap analysis, Implications and suggestions (Chapter 6).

4.3 Sample Selection and its Rationale

The main purpose of this research is to compare the two Service quality metrics (SERVQUAL and SERVPERF) make suggestions for the enhancement of service quality in engineering education. The rationale for selecting the engineering institutes which is an affiliated college of a university is that it typically has most of the characteristics and features of a premier private college ranked among the top ten.

The Sampling Design: Random sampling constitutes the sampling design in the data collection. Convenience sampling is adopted as the method owing to the fact that data from students is freely available. A random sample of final and pre-final year students constituted the strata.

Sample Size (N): The universe of the study is finite with 360 students in three Engineering disciplines viz., Mechanical, Electrical Electronics and Computer Science. The sampling unit is a private engineering college which is a constituent of a Deemed University.

The approach of specifying the precision of estimation desired first, and then determining the sample size necessary to ensure it (Kothari, 2000) is adopted, according to which, the sample size necessary is 124 (eqn. 1).

illustration not visible in this excerpt

Again, the optimum size of the sample in management/social research is based on the nature of the empirical study, time and resources available, and various other considerations such as size of questionnaire, size of universe, nature of classes proposed etc. In practice, the complexity of the competing factors of resources and accuracy means that the decision regarding a sample size tends to be based on experience and good judgment, rather than relying on a strict mathematical formula (Hoinville et. al. 1978). Also the use of surveys in social research does not necessarily have to involve samples of 1000 or 2000 people or events. Instead, research involving a number between 30 and 250 cases is adequate (Denscombe, 1999). In this research, the sample size selected is 184 for the combined sample based on the above formula. In comparison to similar research the number is adequate (Credit card customers n = 187, Long distance telephone customers n = 184, Bank customers n = 177, Zeithmal et al., 1990).

The Instrument: The instrument used in this research are self-administered questionnaire. The reason for choosing this instrument is that it is a relatively systematic and standardized method of collecting data, which lays emphasis on measurement and conversion of data from qualitative to quantitative form. Moreover, the objective of the research was to compare the results of SERVQUAL and SERVPERF to check whether they concur and hence the metric in the form of questionnaire has been used for data collection. Finally, questionnaire method is considered to be economical and convenient for this kind of research.

The questionnaires have been designed to study the perceptions of the respondents on Service quality. They obtain the answers to the research questions and provide the necessary data to test various hypotheses.

The research has made use of the standard SERVQUAL/SERVPERF questionnaire. It used a five-point Likert-type scale, measuring the degree to which the respondents believed the statements in the questionnaire to be true, the highest being ‘True to a very great extent’ and the least being ‘True to almost no extent’. The delivery was on a personal mode to the students, and on a collective mode to the students in the randomly chosen classrooms in the three major disciples of engineering viz., Mechanical, Electronics and Computer Science Engineering.

4.4 Development of SERVQUAL/SERVPERF Metric

Service quality models SERVQUAL and SERVPERF basically differ in their philosophy of defining Service quality. According to SERVPERF, Service quality is based on the Perceived service quality which is defined to be the difference between ‘Perceived service’ and ‘Expected service’ (Parasuraman et al., 1991) and in SERVPERF model Service quality is simply based on the Perceived service quality (Cronin and Taylor, 1992). Both make use of five dimensions of service quality namely, Reliability, Assurance, Tangibility, Empathy, and Responsiveness (Appendix I). The original metrics have been modified to suit the requirements of higher educational service providing requirements (Appendix II). The descriptive information of the dimensions and the sample items is given in table 4.1.

illustration not visible in this excerpt

Table 4.1: Descriptive Information of SERVQUAL/SERVPERF Dimensions

4.5 Organizational Profile and Demographics

The engineering institute selected for the primary data collection of this research is one of the top notch institutes among the private ownership at the national level. The institute is a self-financed college which has completed five decades of fruitful contribution in producing engineers spread across the world and is affiliated to a UGC approved university. It has sixteen under graduate and twenty post graduate courses and Ph.D. in all the departments. The institute has over 450 teaching and 350 non-teaching faculty and about 5000 students. The institute is very well connected through railway, airport and bus facility. Separate hostels are available for boys and girls and a number of messes and food courts are available to meet the food preference of students. Mutli-gym, playgrounds, swimming pools, indoor and outdoor game facilities and other related amenities are also provided to cater to the overall development of students. ; The institute has a very well equipped library in the country operating in over 70,000 sq. ft. area, with 75,000 books, 9,000 bound volumes, 800 videos related to engineering and technology. It has alumni of over 15,000 graduates and 1000 postgraduates spread all over the world.

The focus of this research is mainly to compare the outcomes of the two metrics of Service quality, and hence, the choice of organization for this research is not of prime importance, as both the metrics are served to the same respondents and the same set of service quality dimensions are compared for correlation and significance in difference of perceptions. Even though dimensional preference will be identified, detailed suggestions for quality enhancement in relation to the specific details is beyond the scope of this project, and hence, the work is not conducted as a case study in a specific organization.

4.6 Reliability, Validity and Practicality

Sound measurement must meet the tests of Reliability, Validity and Practicality. These are the three major considerations used in a research, which involves data collection through instruments such as questionnaires (Kothari, 2000).

‘Reliability’ has to do with the accuracy and precision of measurement procedure (Litwin, 1995). A reliable instrument should give identical responses if the questionnaire is served two or more times.

‘Validity’ refers to the extent to which a test measures what we wish to measure. Even though validity to a great extent depends upon the judgement of the researcher three types of validity: content, criterion and construct validity are strongly recommended.

;‘Practicality’ of a measuring instrument is judged in terms of economy, convenience and interpretability. Economy consideration of practicality suggests that some trade-off is needed between the ideal research project and that which the budget can afford. The length of the questionnaire is an important area where economic pressure is felt. More items in a questionnaire will give greater reliability (Kothari, 2000) but this is time consuming and tedious.

4.6.1 Reliability of the Instrument

The ‘stability’ aspect of reliability is concerned with securing consistent results with repeated measurements of the same person with the same questionnaire. But for a sample size of 124, as in the present case, with three disciplines in an engineering college it is not very practicable, and hence, the method of determination of the degree of stability by comparing the results of repeated measurements has been adopted. The most common approach of estimating the reliability of an instrument that is presented to respondents only once is ‘split-half reliability’. In this approach the test is split into two equivalent halves and the scores for respondents on one half are correlated with those scores on the second half of the test. The difficulty in this approach is determining whether the two halves are equivalent. Chronbach proposed the coefficient ‘alpha’ (called Chronbach’s Alpha), which may be thought of as the mean of all possible split-half coefficients. A test with ‘robust’ reliability would be expected to display a Chronbach’s Alpha in excess of 0.9. However, values above 0.7 are usually acceptable indicators of internal consistency as suggested in the literature (Santos, 1999; SPSS, 2000). The reliabilities of all the three instruments have been tested on this basis.

4.6.2 Validity of the Instrument

The instruments used in this research have a proved ‘content and criterion related’ validity, as they are derivatives of standard instruments used before in different organizations. However, in this research, as they have been used in an educational setting, with some changes, the content validity was checked again. Even though the content validity can be primarily judgemental and intuitive, a panel of faculty members was consulted for its content validity and the following changes have been made in the instruments (Table 4.2).

The language of the questionnaire was revised wherever necessary to make the questionnaire more precise and understandable.

; Firstly, to check the ‘construct validity’, the interpretative approach by Erickson (cited in Waldrip & Fisher, 1998) was adopted. The main purpose of this was to check whether the scales were measuring what they were designed to measure i.e. they had construct validity. A semi-structured kind of discussion was adopted with the students as the mode of communication. Initially, the general aspects of Service quality were discussed to give them an idea about the nature of this research. This included promptness of service, tangibles of service, importance of reliability, individual attention etc. just to study whether these practices were streamlined in their departments. They were also asked whether the self-administered questionnaire was simple enough to understand.

illustration not visible in this excerpt

Table 4.2: The Changes made in the Standard SERVQUAL/SERVPERF Instrument

Secondly, their perception about the scales was discussed. It was also asked whether the significance of ‘Perception’ and ‘Expectation’ was clear to them and whether the questions under each dimension did justice to the main heading and whether they adequately described the main dimension with the educational set-up as the reference. The discussion indicated that most of the students were, by and large, happy with the distribution of service variables selected in the questionnaire. ;

Thirdly, their responses for different scales were discussed. The degree of variation on a 5-point Likert scale was also discussed for its adequacy. They felt it was adequate in measuring their agreement/disagreement. It was also discussed whether each dimension gave an equal opportunity to score evenly. Their reply was affirmative.

Finally, there were deliberations about their perception on the key issues focussing on the consistency of their answers. They felt that their rating would be unbiased and would remain consistent. All items were considered to be completely available for their rating.

;The questionnaires were subjected to ‘item validation’ (Pattanayak et al., 2002) through ‘Factor Analysis’ the purpose of which was to determine the internal structure of the set of given number of items. Principal Component Analysis (PCA) method with varimax rotation using Kaiser variation was used to generate factors (Chapter 5). The PCA is very appropriate when the main concern is to predict the minimum number of factors that are required to account for the maximum proportion of the variance when there is a priori set of variables (Ghauri et al., 1995). Hence, this method has been used to test the item validation of the instruments used.

4.7 Practicality of the Instrument

‘Practicality’ of a measuring instrument is judged in terms of economy, convenience and interpretability, as mentioned before. This is one of the reasons why a limited number of items with a maximum of 22 questions were used in the questionnaires of this research. However, care was taken to give a maximum coverage of the study topic.

;‘Convenience’ forms another key factor of practicality. The questionnaire was designed to be self-administrative in nature and clear guidelines were given in the instrument itself, so that the queries regarding how it has to be filled would be minimum. As Service quality is relatively known term in education system, adequate understanding of the fundamentals were evident among the respondents. However, they were not knowledgeable about the dimensions such as Responsiveness or Assurance which was not required as those terms are not used in the questionnaire. The Likert scale scoring keys were stated in the beginning and separate columns were provided for ticking under each variable. Interpretability of the items was given enough importance to see that each question gives only one meaning, free from ambiguity.

4.8 Data Collection Strategies

The following process model (Figure 4.2) was developed and deployed to assure a successful and effective survey dissemination and collection. The research questions have been consolidated into the variables of questionnaires based on the literature review and theoretical models (Chapters 2 & 3). The SERVQUAL and SERVPERF questionnaire was distributed to the students belonging to the different departments and the sequential steps are illustrated in the figure given below.

[...]