Qualitative Data Analysis

1
Qualitative Data Analysis

• Finding or creating and then analyzing texts

2
The Coding Problem

• coding of texts and finding patterns.
• Coding turns qualitative data (texts) into
  quantitative data (codes)
• codes just as arbitrary as the codes we make up
  in e.g., the construction of questionnaires.

3
Qualitative Inquiry

• Purpose
• - to produce findings. The Data Collection
  process is not an end in itself. The culminating
  activities of qualitative inquiry are analysis,
  interpretation, and presentation of findings.
• Challenge
• To make sense of massive amounts of data, reduce
  the volume of information, identify significant
  patterns and construct a framework for
  communicating the essence of what the data reveal
• Problem
• have few agreed-on canons for qualitative data
  analysis, in the sense of shared ground rules for
  drawing conclusions and verifying sturdiness
  Miles and Huberman, 1984)

4
Critical Thinking

• calls for a persistent effort to examine any
  belief or supposed form of knowledge in the light
  of the evidence that supports it and the further
  conclusions to which it tends (Glaser, 1941)
• - means weighing up the arguments and evidence
  for and against.
• Key points when thinking critically are (Glaser,
  1941)
• Persistence Considering an issue carefully and
  more than once
• Evidence Evaluating the evidence put forward in
  support of the belief or viewpoint
• Implications Considering where the belief or
  viewpoint leads what conclusions would follow
  are these suitable and rational and if not,
  should the belief or viewpoint be reconsidered

5
Analytical Thinking

• involves additional processes
• Standing back from the information given
• Examining it in detail from many angles
• Checking closely whether each statement follows
  logically from what went before
• Looking for possible flaws in the reasoning, the
  evidence, or the way that conclusions are drawn
• Comparing the same issues from the point of view
  of other writers

6

• Being able to see and explain why different
  people arrived at different conclusions
• Being able to argue why one set of opinions,
  results or conclusions is preferable to another
• Being on guard for literary or statistical
  devices that encourage the reader to take
  questionable statements at face value
• Checking for hidden assumptions
• Checking for attempts to lure the reader into
  agreements

7
The Credibility of Qualitative Analysis

• depends on three distinct but related inquiry
  elements
• Rigorous techniques and methods for gathering
  high-quality data that is carefully analysed,
  with attention to issues of validity,
  reliability, and triangulation
• The credibility of the researcher, which is
  dependent on training, experience, track record,
  status, and presentation of self
• Philosophical belief in the phenomenological
  paradigm, that is, a fundamental appreciation of
  naturalistic inquiry, qualitative methods,
  inductive analysis and holistic thinking.

8
The Product of Qualitative Data Analysis

• "Naturalistic inquiry is always a matter of
  degree"
• extent to which the researcher influences
  responses and imposes categories on the data.
• The more "pure" the naturalistic inquiry, the
  less reduction of data into categories.

9
Bogdan and Biklen

•  "working with data, organizing it, breaking it
  into manageable units, synthesizing it, searching
  for patterns, discovering what is important and
  what is to be learned, and deciding what you will
  tell others" (1982145)
• challenge
• to place the raw data into logical, meaningful
  categories
• to examine them in a holistic fashion
• to communicate this interpretation to others.

10
Common stages of analysis

• Familiarisation with the data through review,
  reading, listening etc.
• Transcription of tape recorded material.
• Organisation and indexing of data for easy
  retrieval and identification.
• Anonymising of sensitive data.
• Coding (or indexing).
• Identification of themes.
• Re-coding.
• Development of provisional categories.
• Exploration of relationships between categories.
• Refinement of themes and categories.
• Development of theory and incorporation of
  pre-existing knowledge.
• Testing of theory against the data.
• Report writing, including excerpts from original
  data if appropriate (e.g., quotes from
  interviews).

11
3 broad levels of analysis that could be pursued

• Simply count the number of times a particular
  word or concept occurs (e.g., loneliness) in a
  narrative The qualitative data can then be
  categorised quantitatively and subjected to
  statistical analysis.
• For a thematic analysis want to go deeper than
  this.
• All units of data (eg sentences or paragraphs)
  referring to loneliness could be given a
  particular code, extracted and examined in more
  detail. Do participants talk of being lonely even
  when others are present? Are there particular
  times of day or week when they experience
  loneliness? In what terms do they express
  loneliness? Do men and women talk of loneliness
  in different ways? Are those who speak of
  loneliness also those who experience depression?
  Themes could eventually be developed such as
  lonely but never alone or these four walls.
• For a theoretical analysis such as grounded
  theory you would want to go further still.

12
1. Analysis Considerations

• Words
• Context (tone and inflection)
• Internal consistency (opinion shifts during
  groups)
• Frequency and intensity of comments (counting,
  content analysis)
• Specificity
• Trends/themes
• Iteration (data collection and analysis is an
  iterative process moving back and forth)

13
Grounded theory constant comparative method

• open coding (initial familiarisation with the
  data)
• delineation of emergent concepts
• conceptual coding (using emergent concepts)
• refinement of conceptual coding schemes
• clustering of concepts to form analytical
  categories
• searching for core categories
• core categories lead to identification of core
  theory

14
Analysis begins

• identification of the themes emerging from the
  raw data, "open coding" (Strauss Corbin 1990)
• identify and tentatively name the conceptual
  categories into which the phenomena observed will
  be grouped.
• goal - to create descriptive, multi-dimensional
  categories which form a preliminary framework for
  analysis.
• Raw data are broken down into manageable chunks,
  and researcher devises an "audit trail".

15
Next stage of analysis

• Re-examination of the categories identified to
  determine how they are linked "axial coding.
• Discrete categories identified in open coding are
  compared and combined in new ways as the
  researcher begins to assemble the "big picture."
• Purpose of coding not only to describe but to
  acquire new understanding of a phenomenon of
  interest.
• During axial coding the researcher is responsible
  for building a conceptual model and for
  determining whether sufficient data exists to
  support that interpretation.

16
Finally

• Researcher translates the conceptual model into
  the story line that will be read by others.
• Research report should be a rich, tightly woven
  account that "closely approximates the reality it
  represents".
• Stages of analysis not necessarily linear, in
  practice occur simultaneously and repeatedly.

17
RULES OF DATA ANALYSIS

• 1 Timing of Analysis
• a) in relation to data collection
• following data collection linear
• continuing, interactive (e.g., constant
  comparative analysis) in a matrix
• b) in relation to phases of study
• cyclical approach to data collection and
  analysis specified in some designs - (e.g.,
  action research, case study, co-operative
  inquiry). Interim analysis.

18
2. Separability of Data

• a) abstraction of ideas/concepts from 'raw data'
  during analysis
• b) interaction between different datasets, e.g.,
  'melting pot' of all data vs. each tranche
  analysed separately
• c) combination - when and how datasets may (or
  must) be combined or separated

19
3. Admissibility of Data

• a) relative value or worth of different kinds
  of data and how it is assessed
• b) validation required (and how) or not, e.g., by
  members, research participants, other
  researchers, etc.

20
Analytic Principles

• Coding data
• Mark, corral, and reduce data.
• Start with codes a priori or allow to develop.
• Codes evolve with time and experience.

21
Analyzing data and codes

• Mimic quantitative by counting, correlating.
• Reduce data and focus analysis.
• Proliferate codes to see layers of meaning.

22
Computer Assistance

• Does not alter analysis process.
• Usually not a shortcut or timesaver.
• Programs fit different data needs.
• Computer Software
• Atlas-ti large datasets, unstructured coding,
  mimic paper code sort.
• NUDIST large datasets, structured coding, mimic
  quant analysis.
• NVivo less data, unstructured coding, find
  patterns/relationships in codes.
• Folio Views huge datasets, focused coding,
  search sort.

23
6 types

• Word processors
• Word retrievers
• Textbase managers
• Code--retrieve programs
• Code-based theory builders
• Conceptual-network builders

24
Practical Advice

• Start the analysis right away and keep a running
  account of it in your notes
• Involve more than one person
• Leave enough time and money for analysis and
  writing
• Be selective when using computer software
  packages in qualitative analysis

25
The Qualitative Analytical Process(Adapted from
descriptions of Strauss and Corbin, 1990, Spiggle
1994, Miles and Huberman, 1994)
Components
Procedures
Outcomes
Data Reductions
Description
Coding Categorisation Abstraction Comparison Dimen
sionalisation Integration Interpretation
Data Display
Conclusions Verification
Explanation/ Interpretation