Abstract: The aim of this research is to analyze possible effects of service quality of audit firms on customer satisfaction and behavioural intentions. The population of this research consists of Istanbul Stock Exchange (ISE) listed companies. Perceived service quality of audit firms was measured by using modified 22 items SERVQUAL scale. In order to test the research hypotheses, structural equation modelling technique was used. Empirical findings of the research show that perceived service quality of audit firms affects customer satisfaction directly and positively; while perceived service quality does not have a positive direct effect on behavioural intentions, customer satisfaction affects behavioural intentions directly and positively. Consequently, it can be said that perceived service quality leads to satisfaction and satisfaction leads to behavioural intentions.
INTRODUCTION
Service sector has produced approximately two-thirds of worldwide GNP for the twenty first century (Kara et al., 2005). In the service sector, the quality of service is a key strategic value (Lewis et al., 1994) and has become a strategic instrument for firms since 1990s (Donnelly et al., 1995). Because, service quality has close relationship with customer satisfaction, customer retention, positive word of mouth (Buttle, 1996), repurchasing behaviour and extended market share (Siu and Cheung, 2001). In this situation, firms must assess their service quality in order to achieve strategic advantages. This makes that service and service quality are very important research subjects.
Service is not a physical product, but it is a process or performance (Lovelock, 1991). Services are also intangible, heterogeneous, perishable and inseparable of production and consumption and entailing simultaneity (Jain and Gupta, 2004; Zeithaml et al., 1985; Sureshchandar et al., 2001). All these aspects differentiate both services from product and service quality from product quality. Because of elusive and abstract nature of quality and services, there is no objective measure to assess service quality. In the absence of an objective measure, the customers` perception of quality has been used as a measure to assess service quality (Aldhizer et al., 2002; Bamert and Wehrli, 2005).
Parasuraman et al. (1988) have emphasized perceived quality for defining service quality. They defined service quality as a customers` judgment about firm`s excellence or superiority. According to Llosa et al. (1998) service quality is a kind of attitude which is the result of expectations and perceived performance. Customer assesses the service quality by comparing the service that is actually experienced with the expected one (Donnelly et al., 1995). To put it in different words, customer rates the quality of services by the gap between perceived and expected service.
Some researchers (Cronin and Taylor, 1992) argue that there is one underlying dimension of service quality, while others (Parasuraman et al., 1985) argue that there are multi dimensions (Benington and Cummanel, 1998). Parasuraman et al. (1985) revealed 10 dimensions of service quality in the original model of service quality. These dimensions are reliability, responsiveness, competence, access, courtesy, communication, credibility, security, understanding and tangibles. In the later study, Parasuraman et al. (1988), these ten dimensions fall into five dimensions as reliability, assurance, tangibles, empathy and responsiveness. The dimensions of reliability, tangibles and responsiveness were not changed. However, assurance contains communication, credibility, security, competence and empathy contains understanding and access (Carman, 1990). In the model;
| • | Reliability is the ability to perform promised service dependably and accurately |
| • | Assurance refers to employees` knowledge and courtesy |
| • | Tangibles refer to personnel`s appearance, physical facilities and equipment |
| • | Empathy refers to care and individual attention given to customers |
| • | Responsiveness refers to help customers and provide service willingness |
This multi dimensional structure of service quality makes to measure service quality possible by using performance and expectations gap with SERVQUAL scale (Angur et al., 1999). By using five dimensions, Parasuraman et al. (1988) developed 22 items scale in order to measure service quality.
In the empirical literature, there are many alternative service quality models and instruments developed for measuring service quality. One of the most widely used models is SERVQUAL (Zhao et al., 2002). SERVQUAL is developed by Parasuraman et al. (1985, 1988, 1991, 1993). SERVQUAL instrument contains 22 items scale in order to measure the gap between perceived and expected service. In the model, Parasuraman et al. (1985) have defined service quality as a gap between customer perception of performed service and expected service. This model defines service quality as a function of five gaps that relates customer expectation to customer perception of actually performed service (Brown and Bond, 1995). The gaps in the model are the service quality gap (gap 5), the understanding gap (gap 1), the design gap (gap 2), the delivery gap (gap 3) and communication gap (gap 4) (Donnelly et al., 1995).
The validity, reliability and dimensions of SERVQUAL have been discussed in the literature by Carman (1990), Cronin and Taylor (1992, 1994) and Jain and Gupta (2004). SERVQUAL increasingly used for measuring service quality (Cui et al., 2003) because of its practical implication and its diagnostic nature for improving service quality (Zhou, 2004; Sureshchandar et al., 2001). However, there are some critics for SERVQUAL. For example, five dimensions of model are criticized for not being universals and for the high degree of intercorrelations between them (Buttle, 1996). The other critic is that performance only scale (SERVPERF) is enough for measuring service quality (Cronin and Taylor, 1992). According to Cronin and Taylor (1992), service quality judgment and satisfaction stem from the evolution of actually performed service.
Most of the studies about quality in the accounting and auditing literature have focused on audit quality (DeAngelo, 1981; Frantz, 1999; Hansen et al., 2006). Since 1981, researchers have studied on definitions, dimensions and measurement of audit quality (Wooten, 2003). Most empirical studies about audit quality have focused on the relationship amongst audit quality, client risk, agency (Arrunada, 2000) and audit quality components (Barbadillo et al., 2004) that are auditor technical competence and auditor independence. Technical competence relates to expertise to discover all major errors and omissions present in financial statements (Moizer, 1991). Independence requires both independence in mind and independence in appearance (Houghton and Jubb, 2003).
Audit firm is a firm that performs auditing and non auditing services. These non auditing services are accountancy, taxation, share registration, management consultancy, financial consultancy, investigations, liquidations, receivership and trusts (Woolf, 1979). For the past decade, auditing firms` consulting service revenue has increased very sharply while audit service has increased slowly (Turner et al., 1999). This means that auditing firms need to understand the non audit service quality as well as the audit quality in order to obtain and maintain high revenue.
Unlike the audit quality, the non audit service quality has no objective instrument to be measured. In this case customer judgment about service quality has been used to measure both the audit and the non audit service quality. In this study, SERVQUAL scale was used to measure customer judgment about service quality of audit firms.
According to the marketing literature, service quality and satisfaction is related subject (Cronin and Taylor, 1992). Satisfaction is the emotional evaluation of perceived discrepancy between expectations and performance of product or service (Oliver, 1980). In the literature some of the researchers like Bolton and Drew (1991) argue that satisfaction is antecedent factor to service quality. But according to other researchers like Cronin and Taylor (1992), service quality is antecedent factor to satisfaction. According to the second relationship model, high service quality may lead to satisfaction and then satisfaction may lead to favourable customer behavioural intentions (Olorunniwo et al., 2006; Yuan and Jang, 2008). A behavioural intention is the customer decision about future activity and service provider (Hume et al., 2007). And it can be obtained by measures such as repurchase intensions, word of mouth, loyalty, complaining behaviour and price sensitivity (Olurunniwo et al., 2006).
Both service quality and satisfaction play an important role in the forming of customer behavioural intentions (Cronin et al., 2000). Not only does service quality affect satisfaction in forming customer behavioural intentions but it can also affect customer intentions directly (Oh, 1999; Cronin et al., 2000).
In accordance with the literature, the developed research model is shown in Fig. 1. In the research model, the perceived service quality of audit firms was measured by using five dimensional SERVQUAL. In addition, the perceived service quality was hypothesized to have a direct effect on behavioural intentions and an indirect effect, through satisfaction, on behavioural intentions. Satisfaction was also hypothesized to have a direct effect on behavioural intentions.
In order to test causal relationship between perceived service quality, satisfaction and behavioural intentions, four hypotheses were tested in this study. These were:
| • | Hypothesis 1: The SERVQUAL consists of the five-factor construct is suitable for measuring the perceived service quality of audit firms |
| • | Hypothesis 2: Perceived service quality of audit firms has a positive effect on customer satisfaction |
| • | Hypothesis 3: Customer satisfaction with the audit firms has a positive effect on customer behavioural intentions |
| • | Hypothesis 4: Perceived service quality of audit firms has a positive effect on customer behavioural intentions |
| Fig. 1: | Research model |
MATERIALS AND METHODS
Study Population
The population of this research consists of Istanbul Stock Exchange
(here after ISE) listed companies. These companies should have to receive
audit service from independent audit firms in Turkey. There were 326 publicly
held companies in the list published from the ISE web site for the year
2007. List of companies was taken from the ISE web site, 326 companies
were selected from the sampling frame and questionnaires were sent to
companies` accounting manager.
In this study, data was gathered using self-administered direct mail survey. Self-administered mail surveys are generally criticized as having a low response rate and slow (Hair et al., 2000). In order to increase response rate and decrease the likelihood of non-response bias, the questionnaires were sent with separate cover letters. The respondents were informed about the purpose and importance of the study. In addition to encourage respondent`s willingness to fill out questionnaire and return it, a self-addressed stamped return envelope was included in each mailing. A second mailing and e-mailing were employed four weeks after the first mailing and it was followed up with a telephone call after two weeks when no response was returned.
The non-response bias was assessed with an analysis of variance between the early and late respondent groups (Armstrong and Overton, 1977). There were no statistical differences between the early and late respondents` survey responses, indicating lack of non-response bias. The 207 usable questionnaires were received with a response rate of 63%.
Measurement and Analysis
In order to measure perceived service quality of audit firms, modified 22
items SERVQUAL scale developed by Parasuraman et al.
(1985, 1988, 1991) was used.
Survey instrument for this research was comprised of three sections. Section
one contained items on demographics details of the respondent. Section two contained
the modified SERVQUAL items to measure perceived audit service quality (perception
and expectation items). Finally, section three contained the two items to measure
customer satisfaction were adopted form Oliver (1980)
and Taylor and Baker (1994) and two items to measure
repurchase intention were adopted from Cronin et al.
(2000). To measure audit service quality, satisfaction and behavioural intentions
we used the five point Likert response format ranging from strongly disagree
= 1 to strongly agree = 5.
In order to test internal consistency of the SERVQUAL five dimensions Cronbach`s coefficient αscores for each dimension were computed. To test the validity of the five-factor structure in service quality exploratory factor analysis was performed. The research model was tested with structural equation modelling (SEM) by using AMOS program. The maximum likelihood (ML) method of estimation and two-stage testing process were followed in the SEM process (Anderson and Gerbing, 1988). Firstly, confirmatory factor analysis (CFA) was conducted and measurement model fit was evaluated and then the structural model and research hypotheses were tested.
RESULTS
Profile of the Respondents
As the data indicates, the large group of respondents (34.8%) was
between 36 and 45 years old. Overall, the respondents were highly educated
with 69.6% having graduate education and 29.5% having master degree (Table
1). The average length of service experience with their auditor was
3 years, ranging from less than 1 year to more than 8 years. Respondents`
company industry sectors ranged from manufacturing (35.7%) to technology
(2.4%).
| Table 1: | Profile of respondents |
Reliability of the Scales and Exploratory Factor Analysis
In order to test the reliability of the SERVQUAL scale and internal
consistency of the five dimensions, Cronbach`s coefficient αscores
were computed for each dimension using data on the differences between
the perceptions and expectations scores. The reliability coefficients
for the five dimensions are shown in the last column of Table
2. The results in Table 2 show that the five dimensions
have internal consistency measures higher than 0.70. The overall reliability
of the gap scores (0.90) were quite high. The high alpha values indicate
satisfactory levels of internal consistency. Thus, the SERVQUAL scale
is reasonably satisfactory to measure perceived service quality of audit
firms.
To test the validity of the five-factor structure in service quality, exploratory factor analysis was performed on the perception minus expectation gap scores for the Audit firms` customers. A principal axis factor analysis with oblique rotation was performed on the 22 SERVQUAL items. We used the same methods as by Parasuraman et al. (1988) to find the same factor structure. The results of the exploratory factor analysis are shown in Table 2.
The statistical test results (KMO = 0.865, Bartlett Test of Sphericity = 2338.636, Significance = 0.000, factor load = >0.50) indicate that the data was appropriate for the factor analysis. Five factors with Eigen values greater than 1.0 were extracted, which accounted for 66.5% of the variance in the set of the items. The exploratory factor analysis results in Table 2 indicate that the 22 items SERVQUAL scale matched the five-factor structure as described by Parasuraman et al. (1988).
Measurement Model
As mentioned previously, confirmatory factor analysis was employed firstly
to ensure that five dimensions of the SERVQUAL are reliable and valid. Confirmatory
factor analysis statistically tests whether the sample data confirm the theoretical
model (Bollen, 1989). The SEM allows a statistical test
of the goodness of fit for the proposed confirmatory factor solution, which
is particularly useful in validating scales for the measurement of specific
constructs (Hair et al., 1998). The results of the
confirmatory factor analysis for the SERVQUAL items are shown in Table
3 and 4.
The results of the confirmatory factor analysis in Table
3 indicate that the measurement model overall goodness-of-fit indicators
were X2 = 197.35, df = 179, p = 0.17, GFI = 0.97, AGFI = 0.90,
CFI = 0.99, TLI = 0.99 and RMSEA = 0.02.
| Table 2: | Exploratory factor analysis results |
| Extraction method: Principal axis factoring, Rotation method: Oblimin with kaiser normalization | |
A non-significant X2 value with 179 degrees of freedom (p>0.05) indicates the measurement model statistically fit the data (Bollen, 1989). The GFI value was higher than the cut-off value 0.90. The AGFI was found as a value of 0.90. This value is more than the suggested cut-off value of 0.80. The RMSEA value was less than the cut-off value of 0.05. The absolute fit measures indicated an acceptable level of fit for the measurement model (Hair et al., 1998).
The incremental fit measures exceed the critical value of 0.95, further supporting the measurement model. The TLI value, IFI value, CFI value, NFI and RFI value indicated excellent fit and were higher than the cut-off value of 0.95. The incremental fit measures indicated an acceptable level of fit for the measurement model (Schumacker and Lomax, 2004).
The parsimonious fit measure also indicated an acceptable level of measurement model parsimony. The Normed X2 value was 1.10 and was below the upper cut-off value of 5 (Schumacker and Lomax, 2004). All of the fit measures (absolute fit measures, incremental fit measures and parsimonious fit measures) indicated that five dimensions of the 22 items SERVQUAL gap score statistically fit the data well.
After establishing that hypothesized five dimensional SERVQUAL construct fitted data well; parameter estimates, reliability, convergent validity and discriminant validity of the measurement model were assessed.
In order to evaluate the internal consistency of the scale, composite reliability
estimates were computed with confirmatory factor analysis results (Fornell
and Larcker, 1981). As shown in Table 3, these were 0.81
for assurance, 0.84 for reliability, 0.84 for responsiveness, 0.85 for empathy
and 0.83 for tangibles. All of the composite reliability estimates exceeded
the 0.70 threshold recommended by Fornell and Larcker (1981).
| Table 3: | Confirmatory factor analysis results for the measurement model |
| Construct reliability = (Sum of standardized loadings)2/(Sum of standardized loadings) 2+ (Sum of indicator measurement error). Average variance extracted = (Sum of squared standardized loadings) / (Sum of squared standardized loadings) + (Sum of indicator measurement error) | |
Convergent validity can be assessed from the measurement model by determining whether each item`s estimated loading on its assigned construct factor is significant (Anderson and Gerbin, 1988). As shown in Table 3, all the standardized loadings for each factor were significant (p<0.001) and high, which satisfies the criteria for convergent validity. In addition, average variance extracted of all five dimensions exceeded the 0.50 threshold (Table 3) convergent validity was assured.
Discriminant validity can be examined by comparing the intercorrelations of the construct to the square root of the average variance extracted (Fornell and Larcker, 1981). Table 4 reveals that the square root of the average variance for each of the factors is greater than any of the intercorrelations of the constructs. This finding supported that the five-factors of SERVQUAL have discriminant validity.
On the basis of the confirmatory factor analysis results, it can be concluded
that the data reasonably fit to the five dimensions (tangibles, reliability,
responsiveness, assurance and empathy) of SERVQUAL, as proposed by Parasuraman
et al. (1988). The measurement model showed evidence for convergent
and discriminant validity, as well as reliability. Therefore, it can be said
that these results have supported hypothesis 1.
| Table 4: | Intercorrelations among dimensions |
| The bold diagonal elements are the square root of the variance shared between the constructs and their measures. Off-diagonal elements are the correlations between constructs. For discriminant validity, the diagonal elements should be larger than any other row or column | |
| Table 5: | Standardized parameter estimates and goodness-of-fit index for the structural model |
Structural Model
The proposed research model (Fig. 1) and hypotheses
were tested by using structural equation modelling (SEM). The structural
model-fit statistics are shown in Table 5, reveal that
the structural model reasonably fits the data well (X2 = 269.88,
df = 245, p = 0.132, GFI = 0.91, AGFI = 0.892, CFI = 0.991, TLI = 0.989
and RMSEA = 0.022).
A non-significant X2 value with 245 degrees of freedom (p>0.05) indicates the structural model statistically fit the data (Bollen, 1989). The GFI value was higher than the cut-off value of 0.90. The AGFI was found as a value of 0.89. This value is more than the suggested cut-off value of 0.80. The RMSEA value was 0.022 less than the cut-off value of 0.05. The absolute fit measures indicated an acceptable level of fit for the structural model (Hair et al., 1998).
The incremental fit measures exceed the critical value of 0.95, further supporting the structural model. The TLI value, IFI value, CFI value, NFI and RFI value indicated excellent fit and were higher than the cut-off value of 0.95. The incremental fit measures indicated an acceptable level of fit for the structural model (Schumacker and Lomax, 2004).
The parsimonious fit measure also indicated an acceptable level of structural model parsimony. The Normed X2 value was 1.102 and was below the upper cut-off value of 5 (Schumacker and Lomax, 2004). All of the fit measures (absolute fit measures, incremental fit measures and parsimonious fit measures) indicated that structural model statistically fit the data well.
After establishing the structural research model fitted data well, standardized parameter estimates and hypotheses test results are shown in Table 5. Perceived service quality has a positive direct effect on customer satisfaction (standardized estimate = 0.727, t-value = 6.04 and p<0.001).This result has empirically supported hypothesis 1. Therefore, 53% of the variance in customer satisfaction is explained by service quality.
Perceived service quality has not a positive direct effect on behavioural intentions (standardized estimate = 0.042, t-value = 0.813 and p>0.05). This result has not empirically supported hypothesis 2. Customer satisfaction has a positive direct effect on behavioural intentions (standardized estimate = 0.834, t-value = 14.01 and p<0.001). This result has empirically supported hypothesis 3.Therefore, 71% of the variance in behavioural intentions is explained by customer satisfaction and service quality. According to these results, the predicting customer behavioural intentions (repurchase audit service and positive word of mouth about service) perceived service quality is mediated by satisfaction.
Finally, examination of standardized estimates, it can be said that assurance is the most important dimensions of audit firms service quality (standardized estimate = 0.781, t-value = 5.79 and p<0.001), followed by reliability (standardized estimate = 0.743, t-value = 5.02 and p<0.001), responsiveness (standardized estimate = 0.712, t-value = 5.02 and p<0.001), empathy (standardized estimate = 0.684, p<0.001) and tangibles (standardized estimate = 0.631, t-value = 5.62 and p<0.001).
CONCLUSION
The aim of this research is to analyze the effects of perceived service quality of audit firms on customer satisfaction and behavioural intentions by using data obtained from ISE listed companies` accounting managers in Turkey. The exploratory factor analysis result indicates that 22 items SERVQUAL scale matched the five-factor structure. In addition, confirmatory factor analysis results also show that the data fit to the five dimensions (assurance, reliability, responsiveness, empathy and tangibles) of SERVQUAL scale. The SERVQUAL scale is a reliable and valid tool for measuring service quality of audit firms. These findings indicate that the SERVQUAL scale consisting of five dimensions is reasonably satisfactory to measure perceived service quality of audit firms.
According to the analysis result, it can be said that assurance is the most important dimensions service quality of audit firms, followed by reliability, responsiveness, empathy and tangibles. These findings indicate that customers are more concerned with assurance, reliability and responsiveness dimensions to assess service quality of audit firms. Audit firm managers in Turkey should strategically focus on employees` knowledge and courtesy, ability to perform promised service dependably and accurately and ability to help customers and provide service willingness in order to improve the service quality.
The analysis results show that perceived service quality from the audit firms has a significantly positive direct effect on customer satisfaction. And also according to analysis results, customer satisfaction has a significantly positive direct effect on behavioural intention. But perceived service quality has not a significantly positive direct effect on behavioural intentions.
This study reveals that perceived service quality is the antecedent of customer satisfaction in predicting behavioural intentions. While perceived service quality from audit firms has not a direct effect on behavioural intentions, perceived service quality has an indirect effect, through satisfaction, on behavioural intentions. These finding indicates that audit firms` manager, in order to increase favourable customer behavioural intentions (repurchase and positive word of mouth), they must offer superior service quality and satisfy their customer needs and expectations.