Lecture 8 Qualitative Data Analysis DHX_MET1 Methodology 1 Stanislav Ježek Faculty of Social Studies MU OVERVIEW 1.Remembering the RQ 2.Reflecting on what we have learned just by doing the interviews (observations). 3.Reading & re-reading transcriptions 1.Transcribing 4.Coding – themes and concepts 1.The more descriptive part of analysis 5.Theory building 6.Critical reflection 7.Writing 1. -1. REMEMBERING THE RESEARCH QUESTION & RESEARCH PURPOSE/OBJECTIVES •Open, wide RQ in QUAL research. •Descriptive questions •What are relevant factors in peoples‘ experience of… •What is the range of ways of understanding/experiencing/doing…. •Explanatory/Understanding questions •How people make sense of… •How do social/psychological phenomena come to be, what are their antecedents, consequences, necessary conditions… • •In qualitative research the RQ can change (not necessarily). •The richness of data and wide-open possibilities can result in a feeling the original RQ does not serve the research purpose • •Specification •Conceptual/focus shifts 0. Data production (because the analysis is not a separated step after data production) •Interviews – the major source of data • •There are two main reasons for qualitative interviewing to be un/semi-structured: •To provide freedom and space for self-expression •To be able to clarify, extend our understanding of what the person is saying (based on the theory that continually forms in our mind) • •Other sources of data – observations, texts • •We select what needs to be observed (units, behaviors) based on our developing theory Example QUAL interview: https://www.youtube.com/watch?v=9t-_hYjAKww 1. REFLECTIONS FROM INTERVIEWS (RESEARCHER IS AN INFORMANT, TOO) •Preliminary understanding of the informants‘ experience •Initial concepts and their properties •Extending theoretical sensitivity •Field notes on what parts of interviews were difficult, emotional (informant, researcher) and how the difficulties resolved • •Own data collection provides advantages; secondary data analysis still possible • 1½ INTERVIEW TRANSCRIPTION •Interviews MUST be transcribed •Time consuming, expensive process. •https://ocean.sagepub.com/blog/whos-disrupting-transcription-in-academia •Speech recognition is promising but… •still pricey as a professional service, e.g. www.happyscribe.co €12 per hour •without special notation for paralinguistic features •Software assisting with the process, e.g. https://www.audiotranskription.de • •Notation – we want more than just words •pauses, intonation, stresses, interjections (uh-huh) , and more •the level of detail depends on the goals and topic • • •https://www.audiotranskription.de/download/manual_on_transcription.pdf, p. 25 2. READING AND RE-READING •Repeated reading provides further understanding •Especially important when analyzing texts produced by someone else • •Proto-theory •At least by the time of re-reading we should have an idea of the general „shape“, or structure of our theory •What general elements form a theory – what will I look for? • 3. CODING – THE CORE PROCESS •Identification of meaning units in the text •Theoretical sensitivity, previous readings & research questions indicate what a meaning unit is •Meaning unit usually describes(mentions) one incidence of a relevant phenomenon •Assigning labels to meaning units – CODES •Codes are descriptive labels – handles - for phenomena that allow us to quickly refer to them •The words for codes come from theory, by association, from informants (=in vivo codes) •Codes are initially quite specific (i.e. not general) – represent a narrowly defined phenomena •With each next meaning unit I ask whether it is another instance of previously used code-phenomenon, or whether a new code is in order •Coding table – meaning unit + code + explanations/notes from Saldaña (2011) Goal: Career development – how people at the end of UNI think about it.. in vivo codes from Saldaña (2011) themes from Saldaña (2011) 3¼. CODING - Software •Paper&Pen&Highlighter •Office software – Word+Excel •https://www.youtube.com/watch?v=o4qa7Zb8twM •Specialised software •Atlas.ti, NVivo … •free: google free QDA software • • https://www.predictiveanalyticstoday.com/top-qualitative-data-analysis-software/ 3½. CODING – EMERGING DEFINITIONS AND HIERARCHIES •The common features of meaning units with the same code indicate the definition behind the code – CONCEPT •With each coding we decide whether a meaning unit fits the partially implicit/explicit definition. •With each coding the definition may slightly change to accommodate meaning unit •This is how concepts gradually EMERGE (as opposed to having them defined from theory before coding) •Close CODES/CONCEPTS may inspire to define a superordinate category/code/concept – a HIERARCHY emerges •More general (higher) categories are sometimes called THEMES •Sekaran & Bougie call this categorization • • „I HOPE“ in the examples earlier could lead us to a concept of reliance on good fortune. This could lead to a theme of motivation, or positive thinking or some other depending on other parts of the interview. • Počítačem generovaný alternativní text: DEVELOPING THEMES AND CONCEPTS 1. Read and reread your data. 2. Keep track of hunches, interpretations, and ideas. 3. Look for themes that occur frequently. 4. Construct typologies. 5. Develop concepts and theoretical propositions. 6. Read the literature. 7. Develop charts, diagrams, and figures. 8. Write analytical memos. from Taylor (2016) Analytic Memos Topics for Reflection during coding •how you personally relate to the participants and/or the phenomenon •relevance to your study’s research questions •your code choices and their operational definitions •the emergent patterns, categories, themes, and concepts •the possible networks (links, connections, overlaps, flows) among the codes, patterns, categories, themes, and concepts •an emergent or related existent theory •any problems with the study •any personal or ethical dilemmas with the study •future directions for the study (Saldaña, 2009 , p. 40) 3¾. CODING – DESCRIPTION OF DATA •The above coding procedure is considered DESCRIPTIVE •despite a lot of subjectivity and considerable theory used •The informants‘ experiences with respect to the RQ may be efficiently summarized by describing the emerged concepts (their definitions) • •Open coding (within Grounded Theory) •Thematic Analysis, Content Analysis • •Often the analysis ends here •Categories/Concepts/Themes reported in tables •…narratives •…graphical schemes • Content analysis may represent different techniques! Quantitative CA = categories heavily dependent on theory, focus on frequencies (=should be reflected by sampling strategy) Qualitative CA = emergent categories, focus on their qualities, counting does not carry much weight, often missing 4. THEORY BUILDING •Theory of the phenomenon of interest •Depends on the researcher‘s theoretical background and existing theory of the phenomenon •Grounded Theory (Glaser, Strauss, Charmaz) •Narrative Analysis (Bruner, Polkinghorne) •Discourse Analysis (Potter, Edwards, Wetherell) • 4¼. THEORY BUILDING IN GT •Decide about the Central Category (phenomenon, concept) •The phenomenon of which the theory would be •Sometimes trivial decision sometimes not •What composes GT theory? •Properties, dimensions of the central (and other) category •Antecendents of the central category, possible causes •Consequences of central category •Context, intervening conditions • •Analytic Induction a.k.a. the Constant-Comparison method • + •Theoretical Sampling - the deductive component. Počítačem generovaný alternativní text: 3. 6. 7. ANALYTIC INDUCTION 1. Develop a rough definłtion of the phenomenon. 2. Formulate a hypothesis to explain the phenomenon. Study one case to see the fit between that case and the phenomenon. 4. If the hypothesis does not explain the case, reformulate the hypothesis or redefine the phenomenon. 5. Search for negative cases to disprove the hypothesis. When negative cases are encountered, reformulate the hypothesis or redefine the phenomenon. Proceed until the hypothesis has been tested by examining a broad range of cases. Figure 6.2 Steps in analytic induction. from Taylor (2016) Macrì, D., Tagliaventi, M. and Bertolotti, F. (2002). A grounded theory for resistance to change in a small organization, Journal of Organizational Change Management, 15 (3), 292-310. • Lillemor R-M. Hallberg & Margaretha K. Strandmark (2006) Health consequences of workplace bullying: experiences from the perspective of employees in the public service sector, International Journal of Qualitative Studies on Health and Well-being, 1:2, 109-119, DOI: 10.1080/17482620600555664 5. THINKING CRITICALLY ABOUT THE FINDINGS – VALIDITY&RELIABILITY •Validity & reliability are seldom used as terms •Instead: Authenticity, Dependability, Consistency, Applicability… • •Data should be persuasively described (represented) by presented categories, concepts, themes … theory •The reader wants to see how the concepts/theory emerged from data and what is derived form theory •The reader wants to see efforts to •become aware of the role/influence of the researcher •identify weaknesses in the support of concepts/theories • •All is reported in a study à length of QUAL studies • TECHNIQUES FOR INCREASING VALIDITY •Transparency in all parts of analysis •Overt discounting, critical evaluation, explicit falsification attempts •Thick description, verbatim quotes •Triangulation – methods, sources, analytics •Member checking •… TECHNIQUES FOR INCREASING VALIDITY •Transparency in all parts of analysis •Overt discounting, critical evaluation, explicit falsification attempts •Thick description, verbatim quotes •Triangulation – methods, sources, analytics •Member checking •… Počítačem generovaný alternativní text: DISCOUNTING • Solicited or unsolicited statements? • What was your role in the setting? Who was there? • Direct or indirect data? • Who said and did what? • Did you conduct member checks? What was your perspective going into the study? How has it changed? REPORTING •CONCEPTS/THEMES/CATEGORIES •Their definitions, examples of data from which they emerged •STRUCTURE - Hierarchy or theory •Narratively described, explained •Depicted •VALIDITY SUPPORTING INFORMATION •Explicit steps taken •Limitations SUMMARY •QUAL analysis revolves around identification of meaningful elements of data(text) and inductive generalizations of these elements •Substantial general knowledge of the researcher is necessary •The process is FLEXIBLE à must be well described to be trusted •The process is REPETITIVE, ITERATIVE à meaning REFERENCES •Taylor, Bogdan, DeVault, Introduction to qualitative research methods: a guidebook and resource, 4th ed. Wiley, 2016. •Saldaña, J. Fundamentals of qualitative research: Understanding qualitative research. OUP, 2011.