4.5.1 Focus Group Interviews
The literature review presented in Chapter 2 identified the primary dimensions and the sub- dimensions of service quality that might have an impact on the perceptions of Chinese university students. The interrelationships among service quality, value, image, satisfaction and favourable behavioural intentions were also discussed in Chapter 2. All of these
constructs are critical for understanding students‟ university experience. However, to gain additional insight into each of the constructs from students‟ perspectives, as well as developing a questionnaire specific for Chinese students, focus group interviews were conducted.
Zikmund, Ward, Lowe, and Winzar (2007) suggest that a typical focus group should consist of one interviewer and six to ten participants. Following this recommendation, the researcher conducted two focus group interviews. The first consisted of eight Chinese students studying in their second or third year at Lincoln University, chosen because they were very similar to the research population. This group was used to develop the English language questionnaire. The second group consisted of eight students studying in their second, third, or fourth year at He Bei Normal University. This group was used to identify service attributes specific to He Bei Normal University. The two groups of students were considered to have sufficient university experience to provide the necessary background information, to investigate the range of attributes influencing students‟ university experience.
At the beginning of the focus group interviews, the researcher provided an overview of this study. The researcher also explained the domain of each of the constructs in this study so that
participants were aware of what should be included and excluded in defining the constructs, following a procedure recommended by Churchill (1979). Then, participants were asked to define factors shaping their university experience based on the three domains of staff-student interactions, physical facilities, and learning outcomes. Participants were also asked to indicate the factors they considered to be more important in determining perceptions of each of the three domains. During the focus group interviews, the researcher encouraged
interaction between participants in order to detect any ignored or unidentified factors. The interviews were recorded by hand and interpreted. The final questionnaire consisted of 89 items which were used to measure the constructs identified from the literature and focus group interviews (see Figure 3.1).
4.5.2 Questionnaire Design and Layout
The questionnaire consisted of five sections. Section A was composed of 22 Interaction Quality items. Section B presented 30 Physical Environment Quality items. Section C consisted of 18 Outcome Quality items. Section D contained items used to measure the higher-order constructs: Service Quality, Satisfaction, Image, Value and Favourable Behavioural Intentions. Section E measured standard demographic items.
As the performance based SERVPERF provides a superior measurement over the difference scores based SERVQUAL instrument (Cronin & Taylor, 1992), this study used a performance based scale to measure students‟ perceptions of educational service quality.Moreover,
Malhotra (2006) suggests that, when using Likert scales, researchers should strive to establish a consistent scoring procedure in which the respondent‟s favorable attitude towards a subject is consistently reflected by high (or low) scores. Therefore, all of the items in the
questionnaire were positively worded except for one item (No. 20 in Section B), so that agreement with each of the statements represented a favourable response.
The Likert scale, which is typically treated as an interval scale, is a commonly used scale in marketing (Malhotra, 2006). The scale is easy for researchers to construct and administer, and for respondents to understand (Malhotra, 2006). Schall (2003) notes that the seven-point Likert scale is the optimum form of the scale that produces the best response distribution when compared to the five or ten point scale. Therefore, a seven-point Likert scale was used to measure all of the performance-only items included in the questionnaire. The bipolar end- points were labeled with “Strongly Disagree (1)” and “Strongly Agree (7)”. No verbal labels accompanied the points from two to six. Respondents were asked to indicate their degree of agreement with each statement by selecting one of the seven response categories.
Hair et al. (2010) recommend that, as a rule of thumb, a construct should be reflected by at least three items. Compared to a single- item measurement, the superiority of a multi- item measurement includes diminishing the uniqueness of each individual item, providing better distinctions among respondents, increasing the reliability and decreasing measurement error (Churchill, 1979). Therefore, each of the constructs involved in this study was measured by at least three items.
The English version of the questionnaire, which was developed based on the literature and discussions from the first focus group, was first translated into Chinese by the researcher. The translated version was then thoroughly examined by a marketing scholar, who was fluent in Chinese and English. The second focus group interview revealed that the service attributes identified by Chinese students studying at Lincoln University were appropriate for use at He
Bei Normal University. Two marketing academics reviewed the English version to ensure its content validity. Finally, two Chinese marketing scholars reviewed the Chinese questionnaire to ensure its content validity.
4.5.3 Pre-test
Malhotra (2006) suggests that all aspects of a questionnaire, including the question content, sequence and instruction, should be examined through a pre-test procedure. He also
recommends that a pre-test should be conducted on a small sample of respondents, ranging from 15 to 30 people. Respondents selected for the pre-test must be similar to the research population. Further, Malhotra advises that the pre-test is best done by personal interviews, because researchers can observe attitudes and reactions of respondents. A convenience sample of thirty students studying at He Bei Normal University was conducted. These students were asked to read the questions and give comments on ambiguous statements. Responses from the pre-test were used to make minor modifications in the questionnaire. The cover letter and the final version of the questionnaire are presented as Appendix 1 and
Appendix 2 respectively.