• No results found

Data Gathering Tools: family group interviews and research journal

CHAPTER 5: RESEARCH METHODOLOGY

5.4 Data Collection

5.4.1 Data Gathering Tools: family group interviews and research journal

A major goal of this research was to gain insight into the attitudes, values and lived experiences of participating families within the case study. I wanted to elicit the families’ stories and make these discourses the object for analysis. Therefore, the data gathering tools adopted had to offer enough structure around the key themes arising from the literature review (presented in Chapter 2) whilst affording a good degree of adaptability and flexibility. After the piloting phase (see section 5.4.4) family group interviews accompanied by a research journal were selected as data gathering tools.

Family group interviews conducted in a semi-structured manner represented the most appropriate research method to achieve an in-depth understanding of participants’ experiences. Cohen, Manion and Morrison (2018) state that group interviews are a time and cost effective way of gathering data which can often generate a wider range of responses than individual interviews. For example,

having more than one person present may provide a more complete and reliable record. Group interviews provide opportunity for commentary, discourse and interaction (Kidd & Parshall, 2000). Leshem (2012) states that “the group interview is an intensive social encounter that weaves a complex web of communication styles”. Group members may correct, challenge or probe each other’s responses, and conversations between participants can result in more data being shared than would otherwise have occurred during a one-to-one interview (Jaap, 2011). In group interviews, a collective group response is

gathered - one unit of analysis and the researcher remains an interviewer asking questions compared to a focus group where s/he takes on the role of a

facilitator of discussion (Morgan, 1998; Cohen, Manion & Morrison, 2018). Similarly to studies conducted by Moskal and Sime (2016), it was important that both children and adults’ views were sought. This is in comparison with a

substantial body of previous research into family/children’s bilingualism where interviews with youngsters were either conducted separately (e.g. Hancock, 2010) or not at all (e.g. Deakin, 2016). Cohen, Manion and Morrison (2018) write that group interviews are especially useful in conducting research with children. Nevertheless, the social element of group interviews can be a shortcoming; sometimes peer pressure or group dynamics can skew responses, participants may dominate discussion or hold back due to either timidity or fear of revealing ideas in front of others (Leshem, 2012). Also ‘group think’ may occur if

participants are anxious or self-conscious to speak out in front of others (Leshem, 2012, p.3).

By conducting group interviews with family members rather than friends,

acquaintances or strangers, I believe this mitigated some of the concerns around group interviews (Smith-Christmas, 2016). There is a school of thought that children will open up more and feel more relaxed with their parents present, especially younger children however, depending on the topic and circumstances at home this is not necessarily always the case (Morrison, 2013).

I decided to carry out face-to-face interviews - versus video call or telephone - as they allow the researcher to build a rapport with interviewees so that

hopefully they would be more honest and open (Smith, Todd & Waldman, 2009; Cohen, Manion & Morrison, 2018). Smith, Todd and Waldman (ibid) also state that this method offers the scope to explain any complex questions, collect in- depth data and probe answers to delve deeper. However, Wellington (2015) advises that ‘over-probing’ can lead to bias increasing the more the researcher prompts. Denscombe highlights the skills needed to be an effective interviewer. He states that a ‘good’ researcher should be:

…attentive and sensitive to the interviewees and their feelings; able to tolerate silences; non-judgemental; effective in the use of prompts, probes and checks; and an effective facilitator in group interviews/focus groups. (Denscombe, 2014, p.192)

Similarly to Denscombe, but in contrast to other prominent case study methodologists, Merriam (1998) wrote much more explicitly about effective techniques and procedures for gathering data, for example, the importance of interaction between interviewer and interviewee, what constitutes a ‘good’ question and pitfalls to avoid.

Semi-structured interviews allow the researcher a blank page upon which

participants can write their stories whilst maintaining rigour through questioning (Bell, 2005). The interviews were built around open-ended questions that

focused on key themes but enabled me to pick up on unexpected issues and explore these further (Barbour & Schostak, 2011). Interviews explored the following domains: All about you; bilingualism and Spanish, English and other languages; biculturalism; daily life (see Appendix 10 Group Interview Schedule). It was important to have some structure and ensure that the questions asked would address the themes being researched, but it was equally important to be able to explore any topics that interviewees discussed that perhaps I would not otherwise have considered or deemed essential (Cohen, Manion & Morrison, 2018). This gave the flexibility to expand upon their answers and explore topical trajectories if it seemed appropriate or to go into more detail (Smith, Todd & Waldman, 2009). Semi-structured interviews give freedom for respondents to talk about what is of central significance to them rather than what the

Gillham (2000) reminds us that questions must be clear and not misleading nor biased so ensured that I kept this in mind when devising the interview schedule. Furthermore, having been a primary teacher for many years, I was used to

interacting with children and speaking to them in an age-appropriate manner, so I was confident that I had the necessary skills of being a reflexive, impartial listener as Denscombe (2014) recommends.

Charmaz (2006) suggests that only a few broad, open-ended questions are required to encourage participants to articulate their ideas and experiences. I used open questions but had prompts ready in case participants did not give detailed answers. Furthermore, I originally devised two separate interview schedules - one for children and one for parents - which incorporated age

appropriate language and vocabulary (see Appendix 8 ‘Child Interview Schedule’) and Appendix 9 ‘Parent Interview Schedule’) but as modifications to the

research design occurred these were amalgamated (Appendix 10 Group Interview Schedule) (see Pilot Study 5.4.4). The sequence of questions was flexible to allow free-flowing conversation, and the schedule acted as an aide-memoire providing structure and reasonable consistency while allowing a variety of responses (Verma & Mallick, 1999).

Critiques of semi-structured interviews say that there can be inconsistency between interviews given that the questions and line of questioning are organic (Edwards & Holland, 2013). The researcher may start with key questions around specific themes but as semi-structured interviews encourage two-way

communication, conversations may digress and answers may vary widely. Therefore, data analysis and interpretation may be time consuming and it may be more difficult to find patterns (Denzin & Lincoln, 2005; 2011).

I had considered these potential challenges and explored other data gathering tools but I felt that none would offer me the same insight into participants’ lives. For example, questionnaires would have been easy to administer and collate (Bell, 2005) but they would not have allowed flexibility of language or

taken into account the literacy skills, especially the children (Morrison, 2013). Moreover, questionnaires are arguably more effective at describing than explaining particular phenomena, answers may be constrained by the researcher’s questioning skills and response rates can be low (Brinkmann & Kvale, 2009).

Earlier in the process, I had considered conducting a focus group composed of Club Estrella parents. This might have been a good source of dialogue and discussion but I realised that personal or family accounts could be compromised or lost through group dynamics given the lesser relationship with other parents compared to their own family members (Denscombe, 2014). Furthermore, a focus group would involve switching from question-based to topic-based research and taking a more passive role as a facilitator would lessen my control over the direction of discussion (Morgan, 1998; Kidd & Pashall, 2000).

Overall, the semi-structured family interviews I conducted fulfilled my goal of obtaining a balance of perspectives from the 14 families and also personal, unique insights into their lives as bilinguals in the West of Scotland, although not to the detriment of rigour.

5.4.2 Validity, Reliability and Triangulation

In any project, it is essential that any data collected must be critically examined to inspect that it is both reliable and valid. Bell defines reliability as “the extent to which a test or procedure produces similar results under constant conditions on all occasions” (2005, p.117). However, O’Leary (2010) argues that reliable results can be difficult to collect when researching social studies as they may not always be possible to replicate due to the nature of the project and so she advocates ‘dependability’ - ensuring that methods are well designed and carefully thought through to reduce flaws. This case study involved 36 participants who were unique and diverse, results were not replicable and generalisation was never my aim, but it was able to capture a deep insight into the complexities of 14 bilingual families as recommended by Wei and Moyer

(2008).

Stake (2005) argues that difference and uniqueness are what makes case studies such an interesting lens and, although they are bound to some extent to the specific context and the individuals concerned, this is what makes the data rich. Farquhar and Michels (2016) state that in interpretivist research involving case studies, divergent findings from data analysis reflects the multiple realities of participants. Furthermore, Dubois and Gibbert (2010) and Yin (2009) assert that the greater the number of sources or units/sub-units of data, the greater the degree of validity by triangulating the data sources that have been assembled to create the case study. For example, in this case, the group interviews around the themes of language and culture allowed me to triangulate the experiences of schooling by 14 different families in Scotland, from multiple viewpoints related by the policy framework of the time. According to Stake (2005),

triangulation helps identify diversity of perception and different realities in case study research. This aligns with Cohen, Manion and Morrison (2018, p.254) who state that triangulation is “an attempt to map out, or explain more fully, the richness and complexity of human behaviour by studying it from more than one standpoint”.

Meanwhile, Wilson (2009) explains that validity determines whether the research truly measures what it was intended to measure; due to the reflexive and open approach that I took in this embedded case study, I am confident it does. Additionally, interviews are widely employed as a method of investigation of migrant communities in Scotland and have produced reliable and valid results previously (Moskal & Sime, 2016; Kay & Trevena, 2018; Ivashinenko, 2018).