Secondary Qualitative Data Analysis

1. Secondary Qualitative Data Analysis ENRS Webinar Cheryl Tatano Beck, DNSc, CNM, FAAN

3. Secondary analysis is use of an existing qualitative data set to find answers to a research question that differs from the question asked in the original or primary study.(Hinds et al., 1997)

4. History • Glaser (1962) called man “a data gathering animal.” • “The emphasis on survey data neglects other kinds of data, particularly field data, and hence limits the potential use of secondary analysis. This research strategy can be applied to almost any qualitative data however small its amount and whatever the degree of prior analysis” (Glaser, 1963) • Thorne (1994) published the first article on qualitative secondary analysis.

5. Benefits of Secondary Qualitative Data Analysis • Extend the larger context of research results • Promote generalizability of qualitative findings • Salvage data from original study which had not been completely analyzed • Provide a means for continuance of older qualitative findings • Track progression of new trends and knowledge by reusing the older data for comparison

6. Benefits of Secondary Qualitative Data Analysis (cont’d) • Re-examine data from original research in light of different insights from others’ research • Cost effective in time when funding is difficult to secure • Reuse of data maximizes benefit of publicly funded research for good of the public • Develop qualitative theory • Contribute to health policy formation

7. Benefits of Secondary Qualitative Data Analysis (cont’d) • Reduce participant burden by not needing to recruit additional sample • Facilitate more research with marginalized, vulnerable populations difficult to recruit

8. Challenges of Secondary Data Analysis • Qualitative findings are never free from the perspective of the researcher. Since bias will exist in datasets along with interpretive methods used to analyze these data, the potential exists in secondary analysis to exaggerate the impact of those biases. • A fundamental premise of qualitative research is the critical relationship between researcher and the participants. Secondary analysts lack this crucial relationship and can be liable of making an incorrect interpretation of the primary dataset.

9. Challenges of Secondary Data Analysis (cont’d) • A concern for secondary analysts can occur when the original research was on sensitive topics. Secondary analysts did not collect the data and they might have strong emotional responses to reading the interview transcripts or listening to the tape recordings which they had not expected. • Due to the typical but appropriate small sample size in qualitative research, one issue is related to the degree that any qualitative study can represent more than the participants who comprised the sample. Secondary qualitative analysis could lead to conditions which exaggerate any representational problems of primary research.

10. Challenges of Secondary Data Analysis (cont’d) 5. Thorne (1998) warned that “because the sociopolitical consciousness of a group may undergo collective change, interpretations that would have been acceptable to members at one time may not be understood the same way at a later point.” 6. The fit between data and methods by which the primary dataset was constructed is another major challenge for secondary qualitative analysis.

11. Challenges of Secondary Data Analysis (cont’d) • Hinds et al. (1997) addressed the issue of missing data when determining the fit of the original dataset to the secondary question. These authors described the problem of missing data in qualitative studies when the primary researchers explored an issue in one interview but not in all of the interviews. • The currency of the data can be another peril with fit of the primary data set to a new research question.

12. Three Main Modes of Qualitative Secondary Analysis • Formal data sharing (archives) • Informal data sharing: The original researchers may choose to share their data with other researchers who may have had some involvement in the primary study while others may have no prior involvement. The original researchers may or may not be involved with the secondary analysis • Self-collected data: Original researchers reuse their own datasets.

13. Ethics in Secondary Qualitative Data Analysis: Informed Consent There are three layers regarding informed consent qualitative researchers need to address (Alderson, 1998). • First involves the participants’ informed consent for the original researcher to reuse the data. • Second is for different researchers to have access to the participants’ data to reanalyze it. • Third involves the participants’ consent if the data are to be archived.

14. Ethics in Secondary Qualitative Data Analysis: Informed Consent (cont’d) • Pre-planning for reuse of data with participants is best done during the initial consent process • Conversations can be had around the option of data archiving and where data will be stored and who will have access to the data • As more and more funding agencies are requiring data sharing, permission is needed to be made explicit in the initial informed consent.

15. Confidentiality • A delicate balance is needed between researchers honoring their promise to their participants of confidentiality and at the same time retaining the usefulness of the data for future reuse by other researchers. • Anonymization which involves de-identifying information of the participants from the dataset is one way to create a balance. The problem here, however, is that contextual information from specific, situation contexts that can be of value to secondary analysis is removed. If data are distorted to a degree, it lessens their potential value for reuse.

16. Concerns Associated with Qualitative Data Sharing and Archiving • Preparing a qualitative dataset for archiving is time-consuming. So much is involved in this preparation: anonymizing, transferring dataset in file formats accepted by the repository, producing metadata, and making sure of adequate data documentation • Removal of both direct and indirect identifiers may be required to ensure participants’ anonymity. If important contextual information is deleted, the primary researcher should create a log of these changes to the dataset.

17. Thorne’s (2013) Typology of Secondary Qualitative Analysis • Analytic expansion, in which the researcher makes further use of his or her own original database to answer questions at the next level of analysis or to ask new questions as the available theory base expands • Retrospective interpretation, in which the database is used to consider new questions that were raised, but not thoroughly examined, in the context of an original study.

18. Thorne’s (2013) Typology of Secondary Qualitative Analysis (cont’d) 3. Armchair induction, in which those whose talents lie in theory development, rather than engagement with the phenomenon under study, can apply inductive methods of textual analysis, such as hermeneutical inquiry, to existing sets. • Amplified sampling, in which wider theories can be generated through the comparison of several distinct and theoretically representative databases • Cross validation, in which existing data sets are employed to confirm or discount new findings and suggest patterns beyond the scope of the sample in which the researcher personally has been immersed.

19. Heaton’s (2004) Types of Secondary Analysis of Qualitative Data Supra analysis Transcends the focus of the primary study from which the data were derived, examining new empirical, theoretical or methodological questions. Supplementary analysis A more in-depth investigation of an emergent issue or aspect of the data which was not considered or fully addressed in the primary study. Re-analysis Data are re-analyzed to verify and corroborate primary analyses of qualitative data sets. Amplified analysis Combines data from two or more primary studies for purposes of comparison or in order to enlarge a sample. Assorted analysis Combines secondary analysis of research data with primary research and/or analysis of naturalistic qualitative data.

20. Assessment of the Feasibility of a Primary Dataset Primary Research Team • Credentials of the Principal Investigator • Credentials of the entire research team • Availability for consultation Available Contextual Information • Audiotapes • Videotapes • Field notes/memos • Whole interviews • Only parts of interviews (need discursive history of interviewers’ responses)

21. Assessment of the Feasibility of a Primary Dataset (cont’d) Available Contextual Information (cont’d) • Detailed transcription of interviews (interaction between interviewer & interviewee) • How interviewers were selected • Time, place, & setting of interviews • Background characteristics of interviewers • Background characteristics of interviewees Completeness and Quality of Primary Dataset • Complete data for every participant • Richness and depth of interviews • Quality of audiotapes • Missing data • Sufficient data are available to answer secondary research questions

22. Secondary Analysts Relationship to Original Dataset Original Researcher • Original researcher(s) reanalyze one primary dataset • Original researcher(s) reanalyze multiple datasets • Original researcher(s) reanalyze one primary dataset(s) with new author(s) • Original researcher(s) reanalyze one primary dataset(s) plus collect new supplementary data • Original researcher(s) reanalyze dataset with new additional author plus collect new supplementary data • Multiple original researchers combine their datasets & reanalyze them together • Multiple original researchers combine their datasets & reanalyze with new additional authors

23. Secondary Analysts Relationship to Original Dataset (cont’d) Secondary Analyst • Secondary analyst(s) with no previous involvement collaborate with original researcher(s) • Secondary analyst(s) with no previous involvement alone with original dataset • Secondary analyst(s) with prior involvement with original dataset collaborating with original researcher(s) • Secondary analyst(s) with prior involvement with original dataset not collaborating with original researcher(s)

24. Research Process Involved in Secondaray Qualitative Analysis

25. Secondary Qualitative Analysis Review in the Discipline of Nursing

26. Results • 274 secondary qualitative analyses conducted by nurse researchers The earliest published study was in 1988

27. Nurse researchers from which countries are leading in conducting these analyses? • The top five countries are • U.S. (n = 153) • Canada (n = 61) • the U. K. (n = 16) • Australia (n = 12) • Sweden (n = 11).

28. Primary study research design • Descriptive qualitative (n=85) • Mixed Methods (n=56) • Grounded Theory (n=37) • Phenomenology (n=25) • Ethnography, RCT, Combo (n=14) • Others (n=57)

29. Secondary analysis studies: Datasets • 144 (52.6%) studies incorporated the entire primary dataset while 128 (46.7%) used a subset.

30. Secondary analysis studies: Datasets • 191 (69.7%) used one primary dataset while 74 (27%) used multiple primary datasets. • In 9 studies (3.2%) a primary dataset plus new supplementary data were used.

31. Primary Researcher(s) involvement

32. Secondary analysts • The most frequent relationship (n = 57) occurred when the secondary researcher, who had no previous involvement with the original dataset, collaborated with the primary researcher. • In 10 studies the secondary analyst did have prior involvement with the original dataset and collaborated with the primary researcher. • When the secondary analyst had no previous involvement with the original dataset and worked alone in the secondary analysis, this occurred in 7 studies.

33. Data Analysis

34. Secondary Qualitative Analysis Approach Used • Only 68 of the 274 (24.8%) studies identified the specific type of secondary qualitative analysis that was used. • Thorne’s typology was the most frequently cited one (n = 38), • Heaton’s (n = 25), and then • Hinds et al. (n = 5).

35. Metaphor Analysis of the Anniversary of Birth Trauma

36. Table 1 Comparison of the Traumatic Experiences of Shoulder Dystocia: The Obstetric Nightmare

37. Thank You