4. Educación Popular: una necesidad metodológica
4.11. E DUCACIÓN P OPULAR Y M ETODOLOGÍA P-COA ACEM
Qualitative data analysis is a continuous process that begins at the initial phase of data collection and aims at bringing meaning to the object under study (Rabiee, 2004). Data
121 collected using qualitative research methods is often large in quantity and hence the process of data analysis involves reducing this data into a format that is manageable, with the end product being the generation of new knowledge or hypothesis. Qualitative data analysis is mainly concerned with uncovering people‟s opinions, views and feelings about the object of study, irrespective of the validity of this views (Thorne 2000). Due to the large quantities of data collected in qualitative research, it is essential to have the data collection and analysis taking place in tandem so that subsequent interviews and discussions build onto the previous ones (Dey 1993;Krueger and Casey 2000). Analysing of qualitative data is therefore an inductive process that is highly iterative, time consuming and largely driven by the researcher (Lofland et al. 2006). The purpose and objectives of the study are important elements of qualitative data analysis, as the analytical process is geared towards answering research questions raised at the onset of the research process. In this study, a thematic framework approach (Ritchie et al. 2003a) was used to analyze qualitative data collected. The analytical process was systematic and followed the following five key steps (see Figure 16) (Pope et al. 2000):
Transcription of the tapes and field notes
Checking and validating the transcripts
Development of the thematic framework
Coding of the transcripts using the thematic framework
Charting and interpreting the data
While the five steps have been listed in a linear manner, the data analysis process was not always linear but rather it involved moving back and forth between the different steps in order to seek clarifications of the responses and understand the context within which particular responses were made (Ritchie, Spencer and O'Connor 2003a).
Figure 16: Analytical process for qualitative data
Transcription of tapes and field notes Checking and validation of transcripts Development of thematic framework Coding of transcripts Charting and interpreting the data Qualitative data analysis
122
4.9.1 Transcription of the tapes
Transcription is the process of transferring audio- or video-taped materials into verbatim text documents for convenient reference, storage and analysis (Cope 2009). Transcription of the data collected in this study was done by the research assistants under the guidance of the principal researcher. For health care providers and young people whose interviews were conducted in English, the verbatim transcriptions were done directly in English. For young people and community members whose interviews were done in Kiswahili, the verbatim transcription was done directly into English from Kiswahili. Initially the plan was to have the tapes, where the interviews were conducted in Kiswahili, first transcribed in Kiswahili and then translated into English but it became apparent that the process would take longer and the cost would be beyond our means. A decision was then made to have the tapes carefully transcribed directly into English. This was also supported by the fact that members of the research team were conversant with both English and Kiswahili languages. Moreover in most instances, respondents mixed the two languages during the discussions. The transcripts were typed out and each transcript saved as an individual word document with clear labelling showing the study site, type of interview and respondents gender.
4.9.2 Checking and validation of the transcripts
In order to ensure accuracy and consistency of the transcripts, checking and validation of all the transcripts was done by the Principal researcher, who is conversant with both English and Kiswahili languages. Transcript checking was done so as to ensure their accuracy and conformity with what was said by the respondents. This was a rigorous activity that involved listening to the tapes while, at the same time, reading the transcripts to ensure accuracy in language translation, making amendments where necessary. The process was done in batches of transcripts according to the respondents in the study. Checking and validation of the transcripts was also used as a way of familiarization with the data and initial identification common thematic areas (Robinson 1999). To ensure accuracy of the transcripts, six of the transcripts were randomly selected and reviewed by an independent researcher conversant in both Kiswahili and English, who cross-checked them for accuracy and language translation consistency (Mays and Pope 1995). Apart from the typing errors the transcripts were found to be accurate and of good quality.
4.9.3 Development of the thematic framework
Following familiarization with the data, a thematic framework for each set of transcripts, by category of respondents, was developed. This was guided by the research questions, objectives of the study and the major themes and concepts that emerged from each set of transcripts
123 (Richie and Spencer 2002) or rather the repeating ideas that came from the raw data (Auerbach and Silverstein 2003;Spencer et al. 2003). All the data was explored inductively to generate thematic categories (Pope, Ziebland and Mays 2000). This stage led to the development of a thematic framework which was subsequently used for coding or indexing. A thematic framework was developed for each category of respondents: young people, community members, health service providers, and facility managers (see Appendix I). The identification of broad categories of the thematic framework was informed by the research objectives. Initially, five transcripts of each category of respondents were read one by one while noting on the right hand margin of the transcript the themes generated from each paragraph section of the data. Once all the five transcripts had been read, the themes recorded on the margin of the transcripts were then sorted and merged together to construct the initial thematic framework. The sorting of the identified themes involved listing all the themes identified and re-grouping them into main themes and sub-themes or sub-categories (Ritchie, Spencer and O'Connor 2003a). In developing the thematic framework, the following questions were asked while reviewing the raw data and using a social constructionism orientation;
What is the meaning of what the respondent is expressing in this section of the data?
Is this expression similar to what has been said earlier?
Answering these questions helped to identify emerging themes and group similar sections of the data into similar categories. For example “SRH problems” was a thematic area that was derived from sections of the data where respondents identified the SRH problems young people experienced. The specific health problems like unsafe sex, early/unwanted pregnancy and abortion formed the sub-themes or categories (see Table 19).
Particular attention was paid to the frequency with which emerging issues were mentioned among the different respondents, differences in views across the respondents and specific examples of personal experiences given by the respondents (Krueger and Casey 2000). Identification of similar thematic areas across the study groups helped increase the trustworthiness of the data (Patton 1999). The data analysis process also involved identifying majority views, minority views and also conflicting arguments.
4.9.4 Coding of the transcripts using the thematic framework
An open coding approach, also commonly referred to as indexing, was used to code the transcripts and conceptualize the data, using inductive analysis (Richie and Spencer 2002;Strauss and Corbin 1990). Computer-aided qualitative data analysis software (CAQDAS) NVIVO8 was used for data organisation and management. The NVIVO8 programme is essentially a data management tool that systematically aids the data analysis process (Pope,
124 Ziebland and Mays 2000). It has enormous flexibility with respect to data handling and manipulation, as it significantly improves the way the data is accessed, retrieved and viewed (Blismas and Dainty 2003).
NVIVO8 was used to accelerate the coding process. This programme has a “drag and drop” feature which allows for easy coding of the text on to the different themes, sub-themes or categories which are referred to as “tree nodes.” The programme allows for multiple coding of text onto different “tree nodes” and this helps in the identification of links and associations within the data. This process is also called “cutting and pasting” where sections of the data with similar or related themes are put together (Pope, Ziebland and Mays 2000). It also allows for easier retrieval of data related to either a specific node, or a combination of nodes and searching of data associated with key words.
Table 19 shows the thematic framework that was used to code data from IDIs and FGDs with young people plus a brief description of the broad thematic areas. Broad thematic areas were further subdivided into sub-themes or categories. The build up of the sub-themes occurred as the coding progressed (Ritchie, Spencer and O'Connor 2003a). If a new theme emerged it was simply added onto the existing list. For example the broad thematic area “young people‟s experiences with available SRH services”was further subdivided into two sub-themes: “positive experiences” and “negative experiences”. During the coding process, if a data section was describing positive experiences with SRH service provision, that section of the data was coded at both the main thematic area and the sub-theme; and likewise if the description of the data was a negative experience.
To give another example sections of the data that were answering the question, [what are some of the reasons that would make young people or an adolescent not seek health care from the health facility/ youth centre?], were all dragged and dropped under the broad thematic area, “reasons young people do not seek health care”. Retrieving and reading through the coded data at this thematic area led to further categorization of the data into the categories of health provider related concerns, service delivery related and youth related concerns.
Table 19: Thematic framework for coding FGDs and IDIs with young people SRH problems: sections of the data where respondents described
the SRH problems young people faced.
Unsafe sex
Early/ unwanted pregnancy Abortion
Sexual Violence
Ignorance – inadequate information Relationship / growing up
Lack of proper parental guidance Poverty and employment Peer pressure
Where young people seek healthcare sections of the data where
young people identified the sources of health care
Parents / Friends / Relatives Health Facility
Unqualified / Herbal / Traditional Young people‟s experiences with available SRH services sections of the data where young people discussed their experiences
of the available SRH services
Positive experiences Negative experiences Other
125 Media influence
Prostitution School dropout Early Marriage
Drug and Substance abuse
Addressing the SRH problems of young people
sections of the data that described how the mentioned SRH problems could be addressed
Young people and contraception Young people and condom use
o Encourage condom use
o Condoms not used / discouraged
o Female condom concerns Parental Guidance and education Youth and RE education
Health services young people seek sections of the data
where young people identified the SRH problems that made young people seek health care
Contraception
Pregnancy related services STI
HIV related services Sexual Violence
Counselling and SRH information Drug addiction rehabilitation General health
Abortion
Reasons young people do not seek healthcare sections of
the data where young people discussed some of the reasons why they do not seek health care
Health Provider related concerns Service delivery related concerns
o Cost
o Clinic setup
o others Youth related concerns
o Fear
Youth and improving access to SRH services sections of
the data where young people gave suggestions on how services could be improved
Increase availability Improve HSP attitude
Increase awareness of availability
o IEC / Outreach / Advertisement
4.9.5 Charting and interpreting the data
In interpreting data one strives to derive meaning from verbatim texts, while being imaginative and analytical enough to identify the relationships between individual quotes and make linkages within and between the data as a whole (Rabiee 2004). After all the data had been coded, the rest of the data analysis was done manually. This involved reviewing the data in each thematic area or “tree node” and using intuition, summarising the findings in each category (Blismas and Dainty 2003) as well as examining associations within and between the transcripts (Ritche and Spencer 2002). Analytical categories were then indentified and used in writing results summaries (Ritchie, Spencer and O'Connor 2003a).
Charting of the data involved copying the data from each sub-theme or “tree node” in NVIVO8, pasting the coded data in Ms Word and constructing a three column table consisting of the following column headings: coded data, dimensions identified and analytical categories as shown in Table 19 (Ritchie, Spencer and O'Connor 2003a). Table 20 shows the analysis of data under the sub-theme “positive experiences”.
126 Table 20: Description of the data analysis process
Coded data Dimensions identified
Analytical categories “Views about available services” – positive
experiences
Adolescent FGD/01-Langata-girls
R4: They are good they help
Services are good and helpful
Good and helpful services
R1: It helps even in your house. Like that one of family planning can really help you in your house R5: You can go to another hospital and then they tell you the family planning methods are bad, they do spoil you if you are young, so there, you will be scared to use it and say it will spoil me. If you find someone else to tell you that that is a lie, if you use them you will be fine you will be happy to use
Scary advice –FP methods are bad, they “spoil” one
Would be happy to use FP after proper advice
Health providers advice
Presentation of summary findings aimed at identifying individual opinions, groups opinions, areas of disagreement and contradiction (Kitzinger 1995). Concurrent examination of findings from FGDs, IDIs and different respondents was used to identify points of convergence and departure (Lincoln and Guba 1985). This methodology of data analysis was applied to all the transcripts. With regards to data from the young people, data were analyzed to show perceptions and experiences of the available SRH service as expressed by girls, boys, younger adolescents (12-14), and the study sites (Nairobi or the districts). The identification of the original data source was maintained in the analytical process for easier reference and data retrieval.