8.0 Qualitative - data analysis Flashcards
Qualitative data analysis = making sense of the data.
Looking for patterns of ideas or themes
Deviant cases -
Examples of talk or events that run opposite to the emerging hypothesis or propositions about the research and can be used to refine them.
Interpretation process begins during data collection - can be inductive
Inductive – “analysing data with little or no predetermined theory, structure or framework and uses the actual data itself to derive the structure of analysis” (Burnard et al, 2008, p. 429)
Interpretation process begins during data collection - can be deductive
Deductive - predetermined framework is in place to analyse the data.
Coding
Applying codes to qualitative data as part of analysis. Defining what the data are about.
Data examples – interview transcripts, observation notes, emails, photographs
Coding is labour intensive and requires many readings of the data to ensure depth of analysis
Coding is a 4 stage process. What are the stages?
Descriptive or open coding - Initial coding to sort the data in preparation for further analysis
Focused coding - Begin working with the codes to start making sense of the data
Axial or interpretive coding - Reassembling and reorganising the codes for greater abstraction from the data
Selective coding- Identification of central or core themes or concepts based on previous stages of coding
Thematic analysis
Identifying recurring themes through reading and rereading data (Liamputtong and Serry, 2013).
Inductive approach, building up concepts and theories.
E.g. Research topic: Exploring barriers to collaboration between primary health care clinics.
After interviewing several staff in the clinics, you find that themes such as lack of resources or a lack of professional recognition emerge as important obstacles to collaboration between the clinics.
Steps involved in thematic analysis
Make yourself familiar with the data
Transcribe the data yourself.
Read and reread the data and write down your initial impressions and ideas.
Start to generate initial codes
Look for themes by collating codes into tentative themes
Gather all data that is related to each potential theme
Revise the themes you have developed.
* Check if the themes work in relation to the codes you have developed and the entire data set.
* You may also find it useful to develop a thematic map of the analysis
Define and name your themes.
* It is also important to carry out ongoing analysis to refine the themes so that clear definitions and names for each theme can be generated
Content analysis
Deductive approach – codes are identified before they are searched for in the data.
Categories are then explored statistically.
Used for exploring large texts unobtrusively to determine trends and patterns of words, their frequency and structures of communication.
Cannot answer the how? why? or to what effect questions – therefore best used in conjunction with other methods
Framework analysis
Similar to thematic analysis. Not as common as thematic analysis.
Read and reread, ideas/themes, transcribinb.
It’s tansparent, shows how anaylsis has happened.
Involves interconnected stages of data analysis: familiarisation; identifying a thematic framework; indexing and charting; mapping; and interpretation (Pope, Ziebland, & Mays, 2000).
Similar to thematic analysis, however emphasis of transparency in data analysis through systematic approach.
Being increasingly used in health care settings.
Less concerned with theory generation – more concerned with addressing specific questions and useful for informing policy and practice (Ward et al, 2013).
Data saturation
No new information. Sample size is big enough.
Framework analysis stages (5)
Stage 1 - Familiarisation - recurring themes identified
Stage 2 - Thematic Framework - developing a theoretical framework by identifying recurrent and important themes
Stage 3 - Indexing and Charting - themes and sub-themes refined, combined and
developed
Stage 4 - Summarising data in analytical framework -
Stage 5 - Synthesing data by mapping and interpreting
Errors in qualitative analysis
Making conclusions too early in the study
Over coding the data too early
Over identifying with the participants of the study
The halo effect.
What is the halo effect?
refers to the tendency to unconsciously give more attention to views expressed by high status informants.
For example, imagine you were conducting a study on the quality of well water in a village. In this circumstance you might be tempted to over-value the view of a regional community nurse over that of woman from the village. A similar situation can arise when dealing with articulate or well-educated respondents
The most common strategy for analysing qualitative data is A, but there are many other techniques from which to choose.
* B
* C
* D
A - constant-comparison
B -Direct quotes
C - Thematic organisation (systematic identification of recurring themes)
D- Use of descriptive vocabulary
