The purpose of this review is to consider both current and past literature concerning the main concept areas associated with the study in question and also to provide some rationale for conducting the study. The literature review will focus on critically analysing tools which have been used in the past to measure service quality such as, SERVQUAL (Parasuraman et al 1988) and SERVPERF (Cronin and Taylor, 1992). By critically analysing these tools through the views of past research, it will provide a better understanding of which tool should be used in the current study.
A service is an act or performance that one party may offer to another that is strictly intangible and does not result in ownership of anything (Kotler, 1991). Fitness centres base their businesses primarily on the provision of services; therefore it is vital the service provided meets customer requirements. Services have four unique characteristics which distinguish them from goods. Services are intangible, perishable, variable and inseparable (Lamb, Hair and McDaniel, 2008). This view on the four characteristics has been criticised by some authors on the basis that the characteristics stated are not applicable to all service sectors (Afthinos et al, 2005). Also, focusing heavily on these characteristics can overlook the consumer role in the delivery of a service (Afthinos et al, 2005). A key feature of the services is inseparability, as it clearly highlights consumer-employee interaction as a vital part of production and consumption of a service (Chelladurai and Chang, 2000). Organisations which deal with tangible goods are able to measure quality by the number of defects produced; organisations such as fitness centres are unable to do this as they need to measure the service they provide. One way of measuring the service provided is to ask customers to give feedback through certain tools.
A variety of past studies have been conducted to assess service quality. Much of the initial work in developing a model to assess service quality came from Parasuraman et al (1985); who noticed that discrepancies existed between organisations and customer perceptions of the service quality delivered. Parasuraman et al (1985) developed the SERVQUAL scale, consisting of 22 expectation and 22 perception questions, which were rated on a seven point likert scale, ranging from strongly agree to strongly disagree. They suggested that when the perceived experience is less than the expected experience, it implies less than satisfactory service quality. After two stages of purification, the SERVQUAL scale was adapted from a model with ten dimensions to five; tangibles, reliability, responsiveness, assurance and empathy (Parasuraman et al, 1988).
Empirical evidence has shown that SERVQUAL has good trait validity and a reliability of between .80 and .93 (Young et al 1994). This shows that the SERVQUAL scale is valid and the results gained from a study are likely to be reliable, however it must be taken into account that the trait validity and the reliability will vary depending on the type of industry that is being evaluated. It was hypothesised by Parasuraman et al (1988) that when there is a negative discrepancy, the performance gap will cause dissatisfaction, whilst a positive discrepancy will cause consumer delight. The positive and negative discrepancies are calculated through subtracting customer perceptions from expectations.
The problem facing a service provider who does not directly deliver a service to the customer, is how to analyse the service quality gap. A solution for this problem was to use the SERVQUAL scale and identify the areas where expected customer service level is not achieved in each of the five dimensions (Gagliano and Hathcote, 1994). The results from the SERVQUAL scale could then indicate where a gap analysis is required. During the late eighties and very early nineties the SERVQUAL scale was advantageous to managers of organisations as it was easy to use, as well as being the first qualitative method to measure service quality. The SERVQUAL scale invented by Parasuraman et al (1988) has been supported by various researchers (Mckay, 1989; Brady and Robertson, 2001).
The scale has also been tested by McKay (1989) in Canadian municipal parks, and he was able to extract the same five dimensions as Parasuraman et al (1988), confirming the reliability of the scale. SERVQUAL has also been tested by Brady and Robertson (2001) who employed the scale to test the quality of service received from travel agencies. They found that it is still a valid and reliable model with which to evaluate service quality provided by travel agencies (Brady and Robertson, 2001).
The SERVQUAL scale focuses on the fifth service quality gap, which is the discrepancy between customer expectations and their perceptions of the service delivered; when customer expectations are greater than perceptions. Parasuraman et al (1988) believed that by using their performance minus expectations theory, they would be able to analyse and close this gap. Managers and employees would know what the customer requirements were and therefore they could establish a plan to close the gap. Once the plan of action had been in place for a certain amount of time they would conduct the study again to see if they were any closer to closing the service quality gap (Parasuraman et al 1988). This was a good plan as it allows an organisation to see how far away they are from closing the service quality gap. If the research is conducted again and the results are more positive then the organisation knows they are heading in the right direction. However, if the results are negative then the organisation knows that the plan they have implemented needs to be altered.
Chellaudri et al (1987) were one of the first researchers to evaluate service quality in fitness centres as they developed The Scale of Attributes of Fitness Services (SAFS). The scale consisted of five dimensions and was similar to the SERVQUAL model. Parassuraman et al (1988) criticised the SAFS scale by suggesting that it had a lack of structure, and it was confusing as to what in particular was being measured. Also, the wordings of the dimensions were not clear, which in turn questions the scales validity and reliability. If the wording is not clear to the customers then the investigator may think he is measuring one thing when in fact the customer has interpreted the question in another way. Parasuraman et al (1988) suggested the new and improved SERVQUAL scale was the way forward as it was the superior scale to use when measuring service quality across numerous industries.
However numerous researches and testing have not been supportive of the SERVQUAL scale and it has been heavily criticised. Carman (1990) admitted that the SERVQUAL model has good stability; however the five dimensions stated are not always generic, as the dimensions will alter depending on the organisation being surveyed. The method of collecting customer’s expectations after the service has been delivered has also been questioned; instead it should be collected before the service (Carman 1990). Even when the expectations are taken before a service has been delivered there is still no relationship between one another. One of the major objections against the SERVQUAL scale is in relation to the performance minus expectation gap scores. Many authors have struggled to find a positive fit between the quality measured through the SERVQUAL scale and overall quality measured directly through a single-item scale (Babakus and Mangold 1989, Fin and Lamb 1991). Although the use of the performance minus expectations gap score is naturally appealing and theoretically sensible, the ability of these scores to provide extra information beyond that which is present in the perception component of service quality scale is under doubt (Carmon, 1990).
Churchill, Brown and Peter (1993) further criticise the SERVQUAL scale suggesting that problems of reliability, discriminant validity and variance restrictions are present. These problems exist due to the scores on the SERVQUAL scale being “difference scores”, as the perception scores minus the expectation scores. Churchill, Brown and Peter’s (1993) results displayed that even though the SERVQUAL model had high reliability, a non-difference score rated higher in this aspect. They also found that the distribution of SERVQUAL scores was not normal, and the scale also failed to achieve discriminant validity from its components.
Due to many authors, notably Carman (1990), criticising the SERVQUAL model for not being industry specific, Crompton and McKay (1990) developed the Recreation Quality (REQUAL) scale. The REQUAL scale used the SERVQUAL scale as its basis and it was developed for the evaluation of recreation and leisure services. The scale consisted of the same five dimensions as SERVQUAL. Backman and Veldkamp (1995) tested the REQUAL scale and found that the scale was a valid tool of measurement in the recreational field. They recommended that the scale can serve as a template for other researchers to use in their investigation of recreational service quality. Apart from this study, there has been a real lack of confidence in this scale and it has not been used anywhere near as much as the SERVQUAL scale. It must also be noted that this scale still complies by the rules of the SERVQUAL scale. So the majority of the criticism that SERVQUAL (apart from not being industry specific) receives, is also directed at the REQUAL scale.
Wright et al (1992) also attempted to modify the SERVQUAL model, developing 30 items based on the disconfirmation paradigm between customer’s expectations and perceptions. They found that 28 of the 30 items scored negatively, therefore the perceptions did not meet the customer expectations. They concluded that their modified SERVQUAL instrument could be useful in the public leisure context. Although they state the scale may be useful, it will still cause the fitness centre problems as customer expectations always seem to be higher than their perception. Consumers always want the best; therefore the managers of fitness centres will have difficulty in improving these scores and providing the perfect service for each customer. This is one area where the SERVQUAL scale has received heavy criticism (Cronin and Taylor, 1992; Carmon, 1990)
Many authors have attempted to replicate the SEVQUAL scale and failed. The most intense criticism of the scale has come from Cronin and Taylor (1992), suggesting that the SERVQUAL model is based on very little theoretical and empirical evidence. Cronin and Taylor (1992) developed a performance based model to measure service quality, known as SERVPERF. The SERVPERF scale was developed from earlier work which was conducted by Bolton and Drew (1991), who suggested that a consumer’s perception of service quality is largely based on his/her preconceived attitude regarding the service. It was also noted that a customer’s present attitude is based on their previous experience of service quality and their satisfaction or dissatisfaction of the service (Bolton and Drew 1991). After consumers have experienced a service, their attitudes about the service quality may alter over a certain period of time and can change their future attitudes. For example, just after a person has experienced a holiday they may think it was average or below average but as the person looks back over a period of time then they may feel that it was an enjoyable experience. So the perception of services can interestingly change over a period of time.
Cronin and Taylor (1992) suggest that their SERVPERF scale is a better means of measuring an organisation’s service quality than the SERVQUAL scale. Depending on the type of industry surveyed SERPERF’s reliability ranges between .884 and .964 whilst also exhibiting convergent validity (Cronin and Taylor, 1992). Empirical evidence has been shown across a variety of industries to support their performance only scale over the SERVQUAL instrument. According to several authors, the SERVPERF scale is a clear improvement on the SERQUAL instrument (Boulding et al 1993; Carman 1990). Researchers have also found that the single – item performance based scale is able to measure greater variance in the overall service quality than the SERVQUAL scale (Grewal and Brown, 1994; Hartline and Ferrell, 1996). Many authors prefer the SERVPERF scale as it is more efficient than the SEVQUAL scale. This will allow greater feedback as more customers will be willing to answer the questions as there is a reduction in time taken to complete the questionnaires. Another benefit is that the results gained from measuring customer perceptions only are more relevant and meaningful as it is discovered straight away what the customers thought of the service. However, several authors will argue the fact that measuring customers expectations are still important (Parasuraman et al 1988; Gronroos 1984; Bopp 1990)
Teas (1994) has questioned SERVQUAL on a conceptual basis and suggested that it was very confusing in relation to service satisfaction. They agreed with Carman (1990) who suggested that the expectations element of SERVQUAL should be eliminated and instead the performance element alone should be used. They proposed that Cronin and Taylor’s (1992) SERVPERF scale is of better quality and should be used in its place. Although the SERVPERF scale is still trailing the SERVQUAL scale in terms of the amount of researchers who have used the instrument, the number of researchers who are making use of the SERVPERF scale are increasing remarkably. When the two scales have been applied in conjunction with one another, the SERVPERF scale has been shown to outperform the SERVQUAL scale (Cronin and Brand, 2002; Dabholkar et al, 1996).
The debate between the two scales has been continuous. In 1994 Parasuraman et al defended their SERVQUAL model against the numerous amounts of criticism it received. They stated that although Cronin and Taylor (1992) suggest that there is little empirical evidence to support the theory of perceptions minus expectations gap, many researchers have supported this theory (Crompton and Mckay 1987; Gronroos, 1982). Parasuraman et al (1994) also argue that the SERVQUAL model attempts to measure consumers attitude level, whereas SERVPERF tends to focus on the formation of attitudes, therefore according to Parasuraman et al (1994), the two scales are not measuring the same variables. By suggesting that the two models are measuring different variables they are implying that the criticism of the SERVQUAL scale and the preference of the SERVPERF model is irrelevant, however Cronin and Taylor (1994) disagree. It is also noted that research has shown that the SERVQUAL scale has greater convergent and discriminant validity than the SERVPERF model (Parasuraman et al 1994).
The authors do however admit that their perception minus expectations method is less predictive than the SERVPERF scale, but they believe the SERVQUAL scale has better diagnostic value. Cronin and Taylor (1994) responded to the arguments made by Parasuraman et al (1994) and suggest that they are not the only researchers to challenge the SERVQUAL model and many others have found faults in the model (Grewal and Brown, 1994; Hartline and Ferrell, 1996). They go on to say that based on the literature they reviewed, SERVQUAL does not exhibit construct validity. Finally they stress that SERVQUAL has little research that conceptually supports the scale, therefore the question needs to be asked if the SERVPERF scale can validly and reliably measure service quality. Cronin and Taylor (1994) strongly believe that their scale is valid, reliable and very useful in measuring service quality and consumer attitudes. One of the most significant views on the SERVPERF versus SERVQUAL debate comes from one of the founders of the SERVQUAL scale. Zeithamal states that “our results are incompatible with both the one dimensional view of expectations and the gap formation for service quality. Instead, we find that service quality is directly influenced only by perceptions” (Zeithamal, 1993). The findings support the SERVPERF scale as they are in favour of measuring service quality through consumer’s perceptions only. This statement by Zeithamal shows the superiority of the SERVPERF scale, as he was one of the founders of the SERVQUAL scale and is now admitting that his opposing model is superior to his own. Researchers started using the SERVQUAL and the SERVPERF scales as platforms to develop new models with which to measure service quality.
Kim and Kim (1995) developed the Quality of Excellence in Sports Centres (QUESC) instrument, which consisted of 11 dimensions with only sevencorresponding with the SERVQUAL model. The instrument was tested in Korean sports centres and is similar to the SERVQUAL model as it focuses on gaining results from the difference between customers’ expectations and perceptions. After reviewing the literature on service quality, Kim and Kim had the option of creating an instrument to measure customers expectations and perceptions (similar to SERVQUAL) or customers perceptions only (similar to SERVPERF). They decided on the former due to the explosive growth in sports centres over the past few years in the Republic of Korea, which coincided with the emergence of “Sport for All” (Kim and Kim 1995). The fitness centres in Korea at this time were gradually increasing, as demand was increasing. Consequently, the managers of fitness centres wanted to know what customers in this new field required so they could implement it in their fitness centre and gain a competitive advantage over existing and new competitors.
Summary of literature review
A variety of different models and tools have been assessed in the literature review. Although the Parasurman et al’s (1988) SERVQUAL scale has been present for numerous years, praised by several authors and used in many industries, it has been continuously criticised by researchers for a number of reasons. Cronin and Taylor’s (1992) SERVPERF scale has been proven to be more popular than the SERVQUAL scale as the majority of the literature reviewed has been in favour of this model and it has shown greater reliability and validity. The SERVPERF scale has been used more and more as years have passed and it is becoming a strong tool of measurement for service quality. Although other models have been proposed which directly measure service quality in fitness centres they have not received the same support that the SERVPERF scale has. The results gained from this study will be interesting as there has been a lack of use of the SERVPERF scale in the fitness industry.
Hypotheses 1 – There will be no significant difference in service quality between fitness centre A and fitness centre B.
Hypotheses 2 – The empathy dimension will be the highest rated dimension.