The methodology used in gathering information and evaluating the research has been outlined in this introduction chapter. The contents discussed in detail in this study are precisely as follows; the variables used in the study, their measurements, the research design employed, the sampling design chosen to collect the data and the statistical techniques used.
3.1 Research Design
Exploratory study seems to be the most appropriate research design to be used for this study. This is due to the absence of information and also very few previous researches on the service quality of non-audit service. More information or knowledge in the field of interest is expected to be gathered by using the exploratory studies. This indeed will be helpful in developing a theoretical framework and hypothesis. In general, the purpose of this study is to analyze and assess the quality of non- audit services. It will be based on the confirmation/ disconfirmation paradigm. Besides this, we also can determine whether the quality of non- audit has convincing and positive effect on the level of fees paid to comply with the clients' satisfaction.
The perception of small and medium enterprises (SMEs) on the quality of non-audit services by their service provider is the main concern of this study. As a result, SMEs is the unit of analysis in this study. Since it only focuses on SME organizations as the end users of the non- audit services, this study is cross- sectional in nature.
3.2 Research Variables
Independent variable which refers to service quality of non audit services, mediating variable which is client satisfaction and dependent variables which focuses on level of fees, are the three types of variables used in this study. Firstly, independent variable measures the service quality through five dimensions which are tangible, assurance, responsiveness, reliability and empathy. The service quality variables in this research are similar to SERVQUAL model used in Ismail, Haron, Ibrahim and Isa (2006). Secondly, the mediating variables focuses on client satisfaction about the non -audit service provided by small and medium practitioners (SMP) and finally the dependent variable is the level of fees paid for the non audit services.
3.3 Population and Samples
Both the audit service and the non- audit service are provided by some of the small and medium practitioners (SMPs). Most of their clients are from SMEs industries. For this reason, SMEs are used as respondents in this research.
The SME website which is www.SMEinfo.com.my has provided the data on the number of SMEs population in Malaysia. This is due to its availability as the most recent and up-to-date information. There are 16,398 SME companies registered in the web until 11 October 2009. They are divided into seven different business sectors as shown in table 3-1.
The SMEs Business Directory is used to select samples randomly and determine the sample size for this study. The questionnaires will be sent to the respondents through mail, using self-addressed envelopes.
To guarantee that the analysis is acceptable and can be tested, the sample size has to be ranged between 30 to 500 respondents. This is according to Sekaran (2003). As a result, the actual sample size for this study is 100 SMEs. The minimum number of samples for this study will be 35 samples based on the total 7 dimensions which are mentioned in this study. The minimum number of samples are calculated by 7 dimensions times with 5 samples for each dimension (Dr. Jaya, 2009). Primary data which is obtained from the questionnaires sent to SMEs all over Malaysia is used in this study. Even though the sample size which was decided for this study is only 100, 200 questionnaires and self addressed envelopes have been distributed via postal to the respondents. This is because according to Israel (2004), a researcher should increase the number of samples to be distributed by 30% from the normal size in order to compensate with the non-response respondents. The respondents were given 3 weeks to answer and return the questionnaires. Due to very few SMEs response to the questionnaire, convenience sample also has to be used as an alternative way to collect data besides random sampling.
3.4 Primary data
This research used only 1 set of questionnaires (appendix 1) to measure all the variables. The questionnaires were divided into 5 parts:
a) Part A - General information
b) Part B - Questions on independent variables. Questions are regarding
quality service of non-audit services given by the SMPs.
c) Part C - Questions on moderating variables which refers to client satisfaction.
d) Part D - Questions on dependent variable which refers to the level of non audit fees.
e) Part E - Open end question
3.5 Measurements of Variables
3.5.1 Independent variables
Respondents were requested to rate on a scale of 1 (strongly disagree) to 5 (strongly agree) their agreement with regards to 22 statements on quality of services.
3.5.2 Mediating variables
Respondents were requested to rate on a scale of 1 (strongly disagree) to 5 (strongly agree) their agreement with regards to 4 statements on satisfaction of the non - audit services provided SMPs.
3.5.3 Dependent variables
SMEs were requested to choose the average range amount of fees that have been paid for the non-assurance services over the past 3 years to Small Medium Practitioner (SMPs).
3.6 Data Analysis
In order to achieve the research objectives, all data collected will be analyzed. The software application to be used is SPSS (Statistical Package of Social Science) version 15.0. The following section describes the technique applied to analyze data. To analyze the data collected during research, the descriptive statistics, factor analysis, reliability test, correlation analysis, multiple regressions, and frequency analysis will be used. The results obtained will be presented and discussed in Chapter 4.
3.6.1 Descriptive analysis
To study the achievement on service of non- audit service and the performance of SMEs, descriptive statistics are used. This is the fundamental information necessary for us to have a general picture of the population represented by the samples.
By using descriptive statistics, such as frequencies and percentages respondents' profile and company's background were analyzed. Means and standard deviations were used to analyze items which were measured based on 5-point Likert scale and actual figure.
3.6.2 Factor analysis
In order to reduce a vast number of variables to meaningful, interpretable and manageable set of factors, the factor analysis was used. (Sekaran, 2003). Factor analysis is used with the motive to analyze the interrelationships among a large number of variables. In addition, it is used to determine whether the information can be condensed into a smaller set of factors or components with minimum loss of information (Hair, Anderson, Tatham, & Black, 1998). The varimax rotated principal components factor analysis was carried out for quality service of non audit services and client satisfaction.
To verify the assumptions underlying the factor analysis, the Kaiser-Meyer-Olkin (KMO) which measures the sampling adequacy, Barlett's test of sphericity and anti-image correlation were used.According to Hair et al., (1998) the minimum acceptable value for KMO was set at 0.50 and Bartlett's test of sphericity was examined to be significant at the .05 level. Besides this, a diagonal entry of the anti-image correlation matrix was inspected to ensure values above .50; entries below than .50 would be deleted one by one (Hair et al., 1998).
Subsequently, principal components analysis extraction method with varimax rotation was employed. Components with eigenvalues of greater than or equal to 1.00 were selected. The resulting number of components in the component matrix or rotated component matrix indicates the number of factors in a particular variable, which was used in redefining the dimensions in the variable (Hair et al., 1998). Finally, for each item, the factor loading should be at least > 0.50 in order to be accepted.
3.6.3 Reliability test
The purpose of the reliability test to be conducted is to ensure the consistency or stability of the items (Sekaran, 2003). Cronbach's coefficients alpha was used to analyze the reliability of the instruments. However, according to George and Mallery (2003), alpha value above 0.50 is acceptable but it is considered poor. In comparison, Nunnally and Bernstein (1994) stated that the reliability acceptance level should be above .70. Alpha value above 0.50 is being used.
3.6.4 Correlation Analysis
For the purpose of hypotheses testing of the relationship between the independent variables and dependent variables, this correlation analysis is used. According to Sekaran (2003), the relationship is strong if it is close to 1. The significance of the correlations is tested at 1 percent and 5 percent level in two-tail test. Pearson correlation is used to test correlation between the independent variables.
3.6.5 Multiple Regressions
Four assumptions underlying multiple regression analysis which are normality of the error term distribution, linearity of the relationship, independence of error term, and constant variance of the error term) were tested (Hair et al., 1998), prior to the conduct of regression analysis.
This chapter explains specifically on the nature of this study, the research design, the operationalization of the variables and the research instrument used. The sample and data collection were also discussed. In addition, the techniques used to analyze the data collected from the respondents are also described in this chapter.