You are on page 1of 22

Analysing qualitative data

in health service issues

By: Dr.sc.hum. Budi Aji, SKM, M.Sc.


Principles of data analysis:

1. Acceptance that reality is subjective

2. Acceptance that respondents’ view may differ from


researcher’s view

3. Attention to nuances in respondents’ language

4. Importance of context

5. Theory both guides and results in qualitative research

6. Understanding emerges slowly and non-linearly


Approaches to qualitative data analysis:

a. deductive Analyse data to prove/


disprove ideas/theories

b. inductive Analyse data to


generate ideas
Qualitative data analysis methods:

a. case studies & narratives

b. content analysis
Case study:

• “systematic inquiry into an event or a set of related events


which aims to describe and explain the phenomenon of
interest” (Bromley 1990)

• the focus is on the individual case …

• … with the aim of understanding a phenomenon in relation


to the specific context within which it emerged
Case study (continued):

• it relies on variety of data collection methods

• it reports on the interaction between factors

• it is most often disseminated through a narrative

• it is often used in conjunction with other analytical


methods to illustrate the phenomenon in context
Content analysis:
• A set of procedures for organizing non-structured
information into a structured format that allows one to make
inferences about the characteristics and meaning of the
material collected

• Material assumes “structured format” as it is coded

• Coding refers to the process of attaching labels to the


material …

• … in the attempt to tease out common patterns


• Qualitative data analysis, and in particular content
analysis, implies a shift from the EMIC to the ETIC
perspective …

• … organising insider’s concepts (EMIC) in such a way


as to make them understandable also to the outsider
(ETIC) …

• … but without betraying or losing the point of view of


the insider
Emic and Etic:
The emic and etic principle was named in 1954 by the linguist Kenneth
Pike, who used the terms to refer to the fact that people distinguish
phonemes (the basic set of underlying constructs which generate the
sound of language) from their phonetic representation (what people
actually hear).
In 1956, the anthropologist Ward Goodenough insisted that this principle
could be borrowed from the field of linguistics and applied to areas of
culture other than phonology to distinguish between the insider’s, i.e.
emic, and the outsider’s, i.e. etic, point of view.
Although ontological and epistemological debate continues to surround
the exact application of emic and etic, most current anthropologists, and
to a lesser extent other social scientists, continue to use the terms to
distinguish between the insider’s, i.e. the one being studied, and the
outsider’s, i.e. the researcher, conceptualisation of a given situation or
event (Bernard 1990).
Content analysis can be applied on different
sources of information:

- Interview transcripts
- Researcher’s notes & memos
- Observation reports
- Images (photos, movies, ads., etc.)
- Documents
Approaches to content analysis:

Deductive
- generate codes a priori & apply them to the data
- commonly used in theoretical research

Inductive
- generate codes as you move through the data
- commonly used in applied research (including
public health)

In applied research, researchers often used a hybrid approach


Reading Coding

In the field

Questioning Interpreting Verifying

At the desk

Reducing Displaying
Steps in content analysis: (1) Reading

1. READ, READ, READ Begin to identify


emerging themes
1a. Read for content
and develop
1b. Read for quality tentative
explanations
1c. Read for language
Steps in content analysis: (2) Coding

…What do people think? General comment


People think it is a good thing,
because at times a disease Illness
comes to find you and you Illness coping
have nothing in your hands. So
even if you ask for a credit, if Traditional
you cannot pay back, you will Risk-sharing
not get it. But with the
insurance, they will treat you
as a gift. It works like this, if CHI benefit
you give some money today, CHI as aid
they will help you later.

EMIC ETIC
Steps in content analysis: (2) Coding
• Codes are equivalent to labels …
• … which allow researcher to organise material
• No standard rule as to how to assign codes
• Approaches to coding:
- match the ideas in the data
- match the language in the data
- match the interview guide
- borrow abstract terms from social science literature
• Always helpful to define codes & keep track of changes
• Timing of coding: continuous or after data collection?
Sequence in coding

2a. Assign labels


2b. Construct a coding sort
a collection of similarly coded blocks

Leads into following step: displaying


Steps in content analysis: (3) Displaying

• Similar to process of coding, but focus now is on sub-


portions of material (similarly coded blocks)
• Displaying entails:
- taking an inventory of all is known on a topic
- capturing variation in each topic
- separating qualitative & quantitative aspects
- noting differences between individuals/sub-groups
Sequence in displaying

3a. Identify principal theme is each block of material


3b. Identify sub-themes in each block of material
3c. Look for evidence that supports or denies each theme
and sub-theme
3d. Look at what language is used to express ideas
3e. Record emergent ideas as memos
3f. Look for link between themes
Sequence in displaying: an illustration

3a. Main theme: Health care seeking behaviour: people’s


access to health services is poor
3b. Sub-themes: different factors hamper access to services
3c. Found evidence for each factor in text of interviews
3d. Language: expresses disempowerment
3e. Recorded emergent ideas as memos
3f. Health care seeking behaviour linked to quality of care,
poverty, harvest, cultural construction of illness, etc.
Steps in content analysis: (4) Reducing

• Refers to process of distilling the information to make


visible the most important concepts & relationships
• Allows researcher to establish the boundary of each theme
• Often useful to develop matrices & diagrams
• Not necessary for all codes
• Process is highly intertwined with data displaying
Final steps: Interpretation

• Making sense of the themes & the relationships which have


emerged from the material
• A credible interpretation is one that would make sense to
people who have been studied
• The challenge is to interpret data so that they are
meaningful to a wider audience, but do not betray the
respondents’ reality
• Interpretation is context-dependent – all elements are
considered at once
Any questions?

You might also like