Data collection, analysis and ethical issues
Three data collection tools were used in this study: observations, interviews, and questionnaires. Non-participant observations were made of all sessions, and field notes consisting of comments about everything that was believed to be important were made (McKay 2006). These field notes informed the interview questions where interviewees were asked to elaborate on their particular classroom behaviours. Questionnaires were used at the end of each intervention to collect data about students’ evaluation of the lesson and suggestions for the design of the next intervention. The questionnaire was open ended to enable participants to make a broad range of comments (McNiff, Lomax, and Whitehead 1996), which made it possible to collect rich data about the interventions.
Semi-structured interviews were the main data collection tool and were designed to gain insight into the research problem (Kvale 1996) through encouraging the interviewees to elaborate their responses (Borg 2006). Questions were modified according to interviewee responses and this allowed the interview to become more natural, which is essential for creating collaborative conversations.
In the analysis of interviews, case analysis that focused on the responses of each person (Patton 1980) was used. The objective was not to reach a common truth among interviewees but to understand how their individual opinions differed during the term. To analyse the data emerging from open-ended questionnaires, content analysis was used because it was important to understand the predominant opinions of all of the participants, which would determine the design of the next intervention.
A number of ethical issues were addressed in this study. Students were assured that they could withdraw from the study at any time they wanted and there was an alternative class that they could attend. In order to make sure that the students felt safe and comfortable enough to share their opinions freely, the questionnaire data were collected anonymously and the interviewees were assured that pseudonyms would be used to report findings.