Chapter 4: Research Methodology
4.3 Research Design
The research design constitutes the blue print for the collection, measurement and analysis of data (Cooper and Schindler, 2001; 134). A research design is the master plan that specifies the methods and procedures of collecting and analysing needed information (Zikmund, 2003; 58).
A descriptive research design was utilised for the study. Descriptive research is usually designed to provide a summary of some aspects of the environment when the hypotheses are tentative and speculative in nature (Kumar, 2002; 529). According to Cant, Nel, Nel and Kotz‟e (2003; 31), “descriptive research describes aspects such as the market potential of a product, demographics and the attitudes of consumers or behavioural information.”
65 The researcher made use of descriptive research to determine customers‟ perceptions and expectations of service quality at DTA College.
The descriptive research design was selected as it allows the findings from the study to be generalised as the opinions of the DTA College population. According to Aaker et al.
(2004; 75), “some of the benefits of a descriptive research design include the use of larger samples which is normally a good representation of target populations especially if statistical methods are utilised to select sample sizes. Generalisability of results and inferences about facts and estimates of relationships is also possible.”
A descriptive research design with the use of statistical methods allows the researcher to obtain results that can be projected to the whole population and can be of great use to DTA College for various activities that include presentations and decision making.
4.3.1 Sampling
Probability sampling was utilised for the study. Probability sampling refers to a sample in which every element of the population has a known, non-zero probability of selection and its main advantage is that it eliminates the bias inherent in non probability sampling as the process is random (McDaniels and Gates, 2001; 333).
Simple random sampling is the probability sampling technique that was utilised for the sample selection. The researcher assigned a single number to each element in the list of the sampling frame, then a table of random numbers was used to select elements for the sample (Maxfield and Babbie, 2009; 211). A computer generated table of random numbers (with all the student numbers) was used for the sample selection. A copy of the random selection table is attached in the appendices section of the research report.
Simple random sampling is simple to understand and allows the researcher to obtain unbiased estimates of the population‟s characteristics (Shiu, Hair, Bush and Ortinau, 2008;
470). Despite the limitation that probability sampling techniques take more time to design,
66 the use of simple random sampling allows generalisability of the results (McDaniels and Gates, 2001; 333). The use of a computer derived list from DTA College allowed the sampling design process to be faster for the researcher.
4.3.2 Sample size
A sample size of 58 was used for the research. A statistical sample size calculation to come up with an appropriate sample size was utilised. A population of 68 with a margin of tolerated error of 5% and a level of confidence of 95% gives an appropriate sample size of 58 (Raosoft, 2004; para5). The sample represented DTA college learners that managed to be placed during 2010.
4.3.3 Questionnaire Design
The SERVQUAL survey which was utilised for the research comprised of pairs of 22 questions in five service quality dimension categories which are briefly discussed below:
Tangibles consisted offour questions on dimensions such as appearance of facilities and personnel (questions 1 to 4)
Reliability consisted of five questions (questions 5 to 9)
Responsiveness consisted of four questions (questions 10 to 13)
Assurance consisted of four questions on dimensions such as competence, courtesy, credibility and security (questions 14 to 17)
Empathy consisted of five questions on dimensions such as ease of access or contact, communication and customer understanding (questions 18 to 22).
(Kumar et al., 2002; 690)
The questionnaire used for the study comprised of questions that covered all the five service quality dimensions. A copy of the questionnaire is attached in the appendices section of the dissertation (Appendix 2). The SERVQUAL instrument consists of structured questionnaires with closed ended questions, which require the respondent to
67 choose from a predetermined set of options. The SERVQUAL questionnaire utilises the Likert scale which ranges from strongly agree to strongly disagree (Palmer, 2001; 210).
According to Shiu et al. (2008; 330), “the Likert scale is argued to counter the concerns over the length of the questionnaire as the amount of thinking and effort is reduced on the respondents.” These authors further state that the design of the SERVQUAL instrument reduces interviewer bias as the respondent simply circles an answer in response to the question. Therefore, the use of the SERVQUAL instrument was appropriate as it provided the researcher with adequate data to answer the research questions.
4.3.4 Reliability
Reliability refers to the degree to which measures are free from random error and therefore yield consistent results whilst validity refers to the ability of a scale to measure what was intended to be measured (McDaniels and Gates, 2001; 254). The study utilised theSERVQUAL instrument which is a tried and tested instrument in terms of reliability and validity. According to Whitlach (2000; 143), “the SERVQUAL scale has been rigorously tested and has been considered high in validity and reliability.” The five SERVQUAL dimensions possess adequate reliability as measured by Cronbach‟s alpha across different service industries (Antony and Preece, 2002; 149 and Gilmore, 2003; 133).
Cronbach alpha was calculated to test reliability by the researcher to confirm the reliability of the SERVQUAL questionnaire to the particular study carried out at DTA College in Pietermaritzburg. Table 4.1 below indicates the Cronbach‟ alpha results
Table 4.1: Reliability test- Cronbach’s alpha results Service Quality Dimension Cronbach‟s alpha
Tangibility 0.805
Reliability 0.879
Responsiveness 0.850
Assurance 0.861
Empathy 0.889
68 Cronbach‟ alpha results close to 1 indicate high levels of reliability (Gilmore, 2003; 133).
The results for all the five service quality dimensions were high in reliability as the results were very close to 1 as shown on table 4.1.