RESEARCH DESIGN AND METHODS
5.3 Research Design
5.4.5 The interviews
The final method to be discussed is the use of semi-structured telephone interviews. Cohen et
al. (2011) note that there are three main types of interview that can be used as methods data
collection. These are: the structured interview, the unstructured interview and the semi- structured interview. For the purposes of this part of the study, in-depth semi-structured interviews were the most appropriate choice as it was important to go beyond the relatively simplistic, quantifiable responses permitted in the tightly structured questionnaire. The aim
103
was to complement these findings with more insightful, detailed data based on the lived experiences of those involved (Mears, 2012). The semi-structured approach to the interviews also meant that I could develop and ask a series of open-ended questions, some of which would help to elucidate some of the responses which participants had previously included in their questionnaire. The flexibility of this method also meant that issues that were important to the participant, but which had perhaps not been fully considered by the researcher, could be explored and elaborated on in more depth.
Due to participants living in various locations across England it was decided that the interviews would be conducted by telephone. Whilst often associated with short, structured interviews (Fontana and Frey, 1994), for the purposes of this research telephone interviews were more practical in terms of both time and transport costs (Thomas and Purdon, 1994). Moreover, research has indicated that in-depth telephone interviews can yield as high quality data as face-to-face interviews (Sturges and Hanrahan, 2004). In addition, they provided increased flexibility for the parents involved. If, for example, they suddenly realised they could not do the interview on the agreed day/time, it was easier to reschedule a telephone call than it would have been to rearrange travel and transport and a suitable place to meet. The telephone interview meant that participants did not have to leave their house to take part. Equally, some parents found that they could fit the interview in whilst they were at work or commuting, making it a more convenient option. This convenience was important in ensuring that parents remained engaged and willing to participate with the study.
5.4.5.1 The participants
In total, 20 interviews were carried out with parents of children attending nine different Free Schools. Parents identified themselves as being willing to participate in an interview following completion of the questionnaire described above. A total of 44 respondents to the questionnaire included their details in order to be contacted about an interview. Unfortunately, it was not possible to interview all of the parents who had volunteered. In addition, a small number who were contacted changed their mind about willingness to participate, usually due to time pressures. It was felt that 20 interviews would be sufficient for an initial study exploring the experiences and attitudes of parents who had chosen Free Schools. This number
104
would allow for some breadth across a number of different Free Schools, as well as learning about the potentially varied experiences of those applying to the same schools.
The method of recruiting participants via the questionnaire meant that they were self-selected and in no way necessarily representative of Free School parents as a whole. Their reasons for choosing to participate may have been particularly connected to their desire to want to share either positive or negative experiences of the choice process and attitudes towards their child’s school. Awareness of this was important both during the interview and analysis stages. Despite this potential for bias, the parents’ views were nevertheless valid and relevant to the objectives of this part of the study. Whatever their motivations for being involved in the project, their experiences and opinions on the role of parents in relation to Free Schools contributed to an area of study which, to date, has only received very limited attention (see for example, Higham, 2014).
Of the 20 parents who agreed to be interviewed, most (19) were mothers. Further demographic details of each participant and the dates and length of each interview are included in the appendices.
5.4.5.2 Developing the interview schedule
The interview schedule, or guide, is required to provide some structure (where required) to the interview and outline key topics that the interviewer wishes to address (Kvale and Brinkmann, 2009). The research questions linked to this phase of the study influenced the themes and questions that were included on the interview schedule. Due to the semi-structured, open nature of the interviews, it was not deemed necessary to write a heavily-structured script. Instead, the topics and questions included allowed for some spontaneity and diversion, meaning that there was more flexibility in the direction that parents wanted to take with their responses.
Four main themes were used to frame the focus for discussions about school choice and specifically the decision to choose a Free School. These were: attitudes, awareness, action and outcomes. These topics provided prompts for more detailed questioning, conceptualised in relation to the why and the how of school choice. As Kvale and Brinkmann (2009) note it is
105
important to differentiate between the research questions and the actual questions that are asked during the interview. They suggest that theoretical language should be avoided and instead interview questions should be expressed using everyday language that is familiar and accessible to all participants. This was kept in mind when creating the interview schedule and when conducting the interviews.
5.4.5.3 Coding and analysis of data
Following transcription of the interviews, a period of ‘immersive’ reading (Braun and Clarke, 2006) took place. This involved repeatedly reading the transcripts and actively searching for meaning, themes and patterns. Initial codes were then attached to the data, summarising the different (and similar) ideas and topics that were being discussed by each participant. These initial codes had often been defined a priori (Flick, 2002) as they were based on findings from previous school choice literature. The new context of the research (focused on parents choosing brand new schools for their child), however, meant that some flexibility was required and that some new codes were added where required. This method also allowed for comparisons between the current study and the findings from earlier studies. As Miles and Huberman state, “coding is analysis” (Miles and Huberman, 1994, p. 56) and this early reviewing of the data certainly began to make clear some of the overarching themes that would be relevant when moving beyond the simple identification of content. Pattern coding provided a tool for more inferential analysis, allowing for the data to be reduced in to more manageable and focused sections. The development of the pattern codes was an iterative process (Miles and Huberman, 1994), however, and as more data were collected and analysed, some codes were combined or revised and new ones introduced.