authorities.
In the three universities in Yogyakarta, student data is kept centrally in the administration office. Nonetheless, we initially found the student list of UIN Sunan Kalijaga to be inaccurate; because IAIN (Institute Agama Islam Negeri, State Islamic Institute) was in the process of becoming UIN (Universitas Islam Negeri, Islamic State University), data centralization was still ongoing. So, replacements were made because some randomly selected students had already dropped out or graduated. Thus, we used data that was more accurate from each faculty rather than the data from the central administration office. At UGM, the student list was obtained after about a month of bureaucratic procedures. Meanwhile, apart from at UNPATTI, where student data had been centralized, the registration list of two other universities in Ambon (IAIN and UKIM) was generated at faculty level. The sample for these universities was then drawn randomly after combining all faculties into one list.
Of the three universities in Yogyakarta, it was only at UKDW that the campus authority formally supported the research by providing a student list, meeting room and invitation for students to fill out the questionnaire. At UIN, the survey was assisted by the student union. At UGM, the organization of the survey was helped by the Centre for Population and Policy Studies (CPPS), part of the university. Data collection in Yogyakarta was first carried out by gathering the respondents in the auditorium, and then contacting them in their faculty when the first effort was unsuccessful; getting students together to fill out questionnaires was not easy. Fewer than 50% of the expected 50 invited were present. In Ambon, the data collection technique was to visit the selected respondents directly, and ask them to fill out the questionnaire personally.
Respondents varied across the different universities in Yogyakarta in their response to the invitation to fill out the questionnaire together. At UKDW, only 23 respondents participated; to meet the quota of 50, the rest were reached through their faculties. Replacements were made due to the absence of some faculty students for more than a week. In UIN Sunan Kalijaga, the student union assisted the organisation of data gathering, and was very helpful in encouraging respondents to attend the data gathering session; 32 respondents were present. At UGM, only 21 students attended. The absentees were tracked down in their departments, faculties or boarding houses to fill in the questionnaire.
The result of the pilot survey was then assessed to examine the validity and reliability of the research instruments. The correlations between items/variables were scrutinized; if they were very low or extremely high, the items were reformulated or simply removed from the survey. To some extent if, according to respondents, the wording of questions and translations was unclear, it was either replaced or revised with a more straightforward or simpler one. Likewise, some directions as to how
to answer questions were improved to prevent misunderstandings about what the questions meant. Thus, the pilot helped to create better research measurements.
3.2.2. Survey
Given what had been learnt from the pilot, the actual survey began with advance preparation of updated student lists from the selected universities, invitations to respondents from the sample lists and the organization of venues at which the respondents would gather to fill out the questionnaire. The overall goal was to collect a random sample, from each university, of students in the second or third year of their BA, so that generalizations could be made. Institutional permission to obtain the latest student lists was again required from the office of academic administration; this updated data was needed because many students had graduated, and some had dropped out. It is general policy in Indonesian universities for studies to last for a maximum of 7 years; students who take longer must leave the campus. At UGM and UIN Sunan Kalijaga, the student list was issued by the office soon after student registration for the new semester had ended. The data of the survey population in Ambon was obtained at faculty level in all three universities, i.e. UKIM, UNPATTI and IAIN. Due to the eruption of mass violence in the city where the campus was affected, the sampling frame used the data from the 2010 population obtained in the pilot. As explained in the previous section, the research sample was drawn from higher-level students and did not include the new 2011 academic year intake.
The sample was taken by random procedure. It began with calculating the total population divided by 250 as the sample quota for each university. To ensure the randomness of the procedure, the beginning number was selected randomly from 1 to 10, and was then multiplied by the interval. While selecting the sample, a replacement was also taken from the upper number of the sample selection. Thus, 500 respondents were obtained, consisting of 250 as the respondent sample and 250 as the replacement sample. This procedure was applied to the populations of all three universities in Yogyakarta. Meanwhile in Ambon, the sample was drawn at faculty level, so that the total number of respondents was proportionally equivalent to the number of students in each faculty of the respective universities.
Again, data collection was carried out via two procedures, as practised in the pilot. The first was to gather respondents in an auditorium to fill out the questionnaire together, and the second was to hunt down the remaining respondents and ask them to do it personally. The University of Gadjah Mada was targeted first, because the experience of the pilot indicated that it had given the lowest response to the survey invitation. As the capacity of the auditorium was only 125, the survey was held twice, with the invitation given to two groups. The first session was on 27 September 2011 and the second on 28 September 2011, the expectation being that 250 students would attend. Unfortunately, a total of only 52 students came. The