• Tidak ada hasil yang ditemukan

Note

Chapter 4 Methodology

10.1 Data collection techniques

There were two qualitative data collection processes used in this research study viz.

interviewing and focus groups. The combination of individual and group data collection processes provided slightly different perspectives on the research issue. The data collection process in any research study has implications for the validity and reliability of the research.

Conventionally, validity is a concern with the degree to which a method investigates what it intends to. Hammersley (1990, p. 57) argues that it is the “extent to which an account accurately represents the social phenomena to which it refers”. A concern with validity in research is a concern with ‘truth’. However, in qualitative research, the notion of an objective, universal ‘true’ account of a phenomenon is critiqued (Kvale, 1996; Silverman, 2005). Kvale (1996) argues that knowledge is a social construction of reality and truth is constituted through dialogue. Silverman (1997) highlights the constitutive and interactional character of the interviews, and this could also be applied to focus groups. He argues that the ‘story’ told in an interview is only the one that is most tell-able under the circumstances, or the ‘social milieu’ of the communication process. Participants are thus not passive providers of

information, but actively construct versions of reality in the social interaction of the interview and focus group process. This makes the interview or focus group site a site for the

production of meaning, and the interviewing or focus group process an interactional project.

Wardekker (2000, p. 269) argues that in the CHAT approach, as the historical and actual dynamics of practical activities organised in activity systems “are the result of meaning- making discourses, research should be sensitive to the socially constructed character of

activity structures”. In the interview (or focus group process), “subjects are fleshed out, rationally and emotionally in relation to the give and take of the interview process, the interview’s research purposes and its surrounding social contexts” (Gubrium & Holstein, 2001, p. 15).

Within this social interaction of the interview or the focus group, the researcher also has a particular subjective position. The interviewee positions the interviewer or facilitator and responds to her in a particular way. Topics raised in the interview or focus group may also

‘incite’ respondents to voice subjectivities never contemplated before. This affects how one considers the validity of qualitative research data. The value of interview and focus group data thus lies both in their meanings and in how meanings are constructed. Kvale (1996) argues that validity in this framework should therefore focus on defensible knowledge claims. The defensibility of knowledge claims is also related to how the data is analysed, and this will be discussed in the section on data analysis below.

Reliability pertains to the consistency of the research findings and is related partly to the way in which the research was conducted. Kvale (1996) argues that the reliability or consistency of the research findings can be checked throughout the research process. For example, the reliability of interviews and focus group discussions can be checked through examining how the questions are asked, and the process of translating and transcribing the data can also be checked. A thorough documenting of the research process is a critical aspect of ensuring reliability, and this is provided in the section below.

10.1.1 Interviews

Interviews provide a means to access the perspective of the research participants on the phenomenon. In this study interviews were conducted with participants in order to obtain accounts of their experiences of sexual activity in relation to HIV and AIDS.

In this research study, 27 interviews were conducted. I conducted 15 of these interviews.

Three of these 15 interviews were conducted with couples, and 12 were conducted with individuals. A further 12 individual interviews were conducted by two CADRE researchers.

All of the interviews were conducted in the participants’ homes.

Interviews which I conducted

Interviews conducted by the CADRE

researchers

TOTAL number of interviews

conducted Couple interviews 3

Individual interviews

12 12 27

Table 7. Interviews conducted

The interviews were conducted using a semi-structured interview guide devised in relation to the research question (see Appendix 1 for the interview schedule used for couples and individuals). The focus in the data collection process was on the sexual ‘biographies’ of the participants. Key areas of focus were:

· Introduction to sex or sexual ‘socialisation’ (how one learnt about sex, who introduced you to sex)

· The nature of early sexual relationships, particularly first sexual experience or ‘debut’

sex (age, feelings about this experience, social norms and expectations; contraceptive use; parents’ awareness of the activity)

· The dynamics of sexual relationships (their experience and feelings, age, gender, initiating relationships, social norms and expectations; contraceptive use; multiple partners)

· Cultural practices in relation to sex education and sexual interaction

· Perceptions of HIV (including knowledge of HIV and HIV transmission; exposure of HIV)

· Perceptions of vulnerability to HIV risk

As is typical of a qualitative research process, each interviewer addressed the central issues, but followed the interview questions in slightly different ways.

Some of the interview questions addressed personal and sensitive information. There are obviously constraints on this kind of data. Information about sexual activity is not readily volunteered. The accounts tended to be normative and did not allow for an investigation of the discrepancies that usually exist between precept and practice (Delius & Glaser, 2002).

However, most of the participants responded to the interview questions and were quite forthcoming about their experiences and feelings. I think this was largely due to the rapport established in the interview setting but also a consequence of the my long standing connection with the research context.

10.1.2 Focus groups

Data were also collected through focus group discussions, a method particularly suited to exploratory descriptive studies (Berg, 1998; Stewart & Shamdasani, 1998). The focus group method prioritises the perspectives and experiences of the research participants and has the advantage of de-emphasising the prominence of the interviewer (Krueger & Casey, 2000).

Kitzinger (1994, cited in O’Kane, 2000) describes focus groups as encouraging communication especially around difficult issues such as sexual activity. Stewart and

Shamdasani (1990) argue that focus groups are appropriate to explore issues in a context, and to observe and understand attitudes and behaviour. For example, Kitzinger (1994, cited in O’Kane, 2000) used focus groups to understand how people think about AIDS.

In this study five focus groups were conducted. I conducted three of these focus groups: one with a group of women between the ages of 30 and 60; one with female youth between the ages of 16 and 25, and one with male youth between the ages of 16 and 25. Two focus groups with the youngest research participants (between the ages of 10 and 15), were conducted by the two CADRE researchers.

Focus groups which I conducted

Focus groups conducted by the CADRE

researchers

Women 30-60 years of age FG 1

Female youth 16-25 years of age FG 2 Male youth 16-25 years of age FG 3

Girls 10-15 years of age FG 4

Boys 10-15 years of age FG 5

Table 8. Focus groups conducted

Research with child participants presents unique challenges to the conceptualisation and implementation of the research process (O’Kane, 2000; Jones, 2004). Individual interview with the younger participants might have been very intimidating. The group format of focus groups enabled the children to converse with other participants, potentially mediating some of the more extreme power differentials between adult interviewers and child participants (Van der Riet, Hough & Killian, 2005).

Focus groups were conducted with the older youth participants because they provided an additional source of data to the interview data. The focus group discussions also provided an opportunity to observe some of the group dynamics amongst these participants. An

assumption within the HIV and AIDS literature is that peer group dynamics play an important role in an individual’s decision to engage in safe sex practices (Dolcini et al., 2004). To a certain extent the group nature of the data collection process replicates or mimics the ‘group’

in which the participant exists (Stewart & Shamdasani, 1990). The group influences the nature of what the individual says thus potentially drawing out particular dynamics which may be useful for the research question, for example, grouping adolescent boys together to discuss the way in which they relate to girls. Focus groups might therefore invoke aspects of peer culture, potentially illustrating the dynamics in the social world of the participant.

In this study the focus group discussions differed from the interviews in that the participants were asked to discuss the broader community of the Amathole Basin and how people

responded to HIV and AIDS. The focus was thus not on their own personal experience within an intimate relationship but more broadly on what people in the research context do in

response to HIV and AIDS. Four of the focus groups followed a set of questions on relations between boys and girls, and also on the subject of sex and AIDS (see Appendix 2 for the focus group interview schedules used for the five different focus groups). The fifth focus group (with the group of women between 30 and 60 years of age) was a more spontaneous and less-structured discussion. The focus group discussions took place in the pre-school building in one of the villages.

In this study the data from the focus groups tended to be a bit superficial, with many of the statements made by participants not explored in detail. This is partly a function of the way in which the groups were facilitated and also the age of the participants. Obtaining detailed responses from the youngest participants was not possible.

10.1.3 Data collection process: translation and transcription 10.1.3.1 Translation

All of the interviews and focus groups were conducted in isiXhosa, the mother-tongue of the participants. In the interviews and focus group discussion which I conducted, I posed the questions in English, and an isiXhosa-speaking translator then posed the question in isiXhosa.

Participants responded in isiXhosa, and the response was then translated into English for me by the translator. The translator was a research assistant from outside of the research context because I was doubtful that participants would feel free to discuss the rather personal content matter of the interviews and focus groups with someone whom they knew, and would have to see again after the research process. The interviews and focus groups conducted by the CADRE researchers were conducted directly in isiXhosa and then translated into English during the transcription process.

The interviews which were conducted in isiXhosa without a translator flow a bit more freely and are longer than the other interviews. It would have been preferable to engage in the research process in the language of the participants. However, given my language abilities, and those of the research participants, this was not possible. Conducting research in a context in which people were familiar with me I think, in part, helped to alleviate some of the

constraints on communicating in different languages.

Translation is a difficult process because it relies on the skill and fluency of the translator in two languages, and it also adds a layer of interpretation to the data. Translation thus affects the validity and reliability of the data. There were no checks done on the translation process within the interviews and focus groups, however, once the data was transcribed some terms in the transcriptions remained in isiXhosa, without English explanations. I sought assistance from three independent isiXhosa speakers to ascertain the meaning of some of these words.

Of particular concern were some of the cultural and historical practices related to sexual activity. Interestingly, neither of the two isiXhosa-speaking CADRE researchers, nor the two independent isiXhosa speakers whom I approached, could interpret some of the words used by the participants. This might have been because of their age, and their lack of familiarity with traditional practices. Some of these words were corroborated by translations used in the anthropological and historical literature, for example, imitshotsho/emtshotshweni, intlombe, amagubu, and umguyo.

10.1.3.2 Transcribing

All of the interviews and the focus groups were audio-recorded. These recordings were then transcribed. Poland (2003) comments that transcription quality is not a dominant concern in qualitative research, but that it should be. He argues for a sceptical reflexivity on the

“multiple interpretive acts that constitute the transcription process and their impact on the process of translating (re-presenting) the interview as audio re-cording and then as textual data” (ibid, p. 274). He also problematises the concept of a ‘verbatim’ transcript because he argues, it is not possible to record the many aspects of interpersonal interaction and nonverbal communication (e.g. eye gazes, facial expressions) on audio-tapes and thus this audiotape is not a verbatim record of the interview. He argues that

even when a transcriber attempts to produce a verbatim account by remaining faithful to the original language and flow of the discussion, and even when the transcriber has a suggested syntax to follow in transcription…there are a number of logistical and

interpretive challenges to the translation of audiotape conversation into textual form (ibid, pp. 269-270).

An example of these interpretive challenges is that inserting a comma or a period in the text can alter the interpretation of the text. Another problem is that of mistaking words for other similar words. In the case of ambiguity “researchers interpolate what makes sense to them as being what was likely uttered during the course of an interview” (ibid, p. 270). This is particularly problematic when the person doing the transcription was not directly involved in the research, or has not familiarity with the research issue. Poland (2003) suggests that transcription quality is maximised through paying attention to recording quality, using a notation system, reviewing the quality of the transcription, and flagging ambiguity in the interview or focus group discussion.

In this study I transcribed 3 focus groups and 15 of the interviews. This transcription was of the English translation provided by the translator in the course of the interviews and focus groups. The two CADRE researchers transcribed 2 focus groups and 12 of the individual interviews. An advantage of the transcription process in this study is that those researchers who conducted the interviews or the focus groups transcribed them, reducing the ‘levels’ of interpretation added to the data in its re-presentation in a textual form.

Conventions for transcribing, or notation systems, have been developed to faithfully represent the details of the conversation including sound, content and structure. Using these

conventions, and reporting on them in the research process, provides the reader with an opportunity to check the adequacy of the claims being made (Poland, 2003). In this study, no standard procedure or notational system for transcribing was used across the transcription by the three transcribers. This was partly pragmatic in that the CADRE researchers transcribed the interviews and focus groups for their own research purposes and I obtained the transcripts some time after this. The quality of the transcriptions could therefore not be reviewed. This might have created inconsistencies between the transcripts and within the data corpus as a whole. However, in the interviews and focus group discussions which I conducted, the conversation between interviewer and participants was recorded as accurately as possible and many of the conventions identified by Poland (2003) were followed, for example:

· indicating turns taken for speakers (interviewer, male participant, female participant, and where necessary, translator)

· pauses were indicated with a series of dots (…) where they were more than three seconds

· isiXhosa words were used when no direct English equivalent existed

· some details of the conversation other than words were included, for example, laughter

· interruptions of the interviews were recorded and indicated with square brackets [ ]

· parenthesized words using square brackets [ …] contain the author’s descriptions rather than transcriptions

· ambiguous words or phrases were bracketed and flagged with question marks (??)

Kvale (1996, p. 233) uses the term ‘analytical generalisation’ which he argues is “a reasoned judgement about the extent to which the findings from one study can be used as a guide to what might occur in another situation”. However, for this to work, sufficient evidence must be provided by the researcher for the analytic generalisations to be made. This relates to the process of data collection, and the process of data analysis. The validity of the generalisation hinges on the extent to which the attributes compared are relevant, which in turn rests upon rich, dense and thick descriptions of the case. Specifying the supporting evidence, and making the argument explicit, allows the reader to judge the soundness of the generalisation. The way in which the results are presented should illustrate whether it is possible that in other

situations the same, or at least recognizable constraints, apply. Seale (1999, cited in

Silverman, 2005) argues that reliability in qualitative research is strengthened by providing

low inference descriptors. Practically this means providing verbatim accounts of the research interaction. In reporting on the analysis of the data, presenting longer extracts of data that include the question preceding the respondent’s comments and the ‘continuers’ in the interviews, retains the link with the raw data and the origins of the interpretation. In the next two chapters, which present the outcome of the analytic process, the activity systems are described by using detailed examples from the participant’s accounts.

11 Ethical issues in the research process

Roth (2005) argues that qualitative research might have particularly complex ethical issues because there are genuine interactions between researchers and participants rather than say in survey research, or with email questionnaire surveys. Adherence to ethically sound research principles is usually regulated by ethical review boards. This study did not have formal ethical clearance because it was conducted prior to universities and research institutions in South Africa obligating researchers to submit proposed research for ethical review. However, the central principles of conducting ethical research, viz. autonomy and respect for the dignity of persons, non-maleficence, and beneficence (Beauchamp & Childress, 2001; Emanuel, Wendler & Grady, 2000), were adhered to throughout the research process. Emanuel et al.

(2000) argue for research which is socially valuable, and in which the participants are not exploited, but treated fairly and with respect.