QQM: topic 4: qualitative data analysis

distinct features qualitative data analysis

• focused on texts

• hermeneutic perspective

• emic focus

• iterative process of progressive focusing

emic focus

representing a setting with the participant’s terms

etic focus

representing a setting with the researcher’s terms (quantitative research)

goals qualitative research

• meaning: how people see the world

• context: the world in which people act

• process: what actions and activities people do

• reasoning: why people act and behave the way they do

good qualitative results rely on

• quality of data collector

• quality of data analyzer

• quality of presenter/writer

examples of qualitative data

• written field notes

• audio recordings of conversations

• video recordings of activities

• diary recordings of activities/thoughts

qualitative data provides depth info on

• thoughts, views, interpretations

• priorities, importance

• processes, practices

• intended effects of actions

• feelings and experiences

three steps of reading text

1. literally: focus on literal content and form

2. reflexively: researcher orientation shapes interpretation and focus

3. interpretively: researcher tries to construct own interpretation of text meaning

characteristics qualitative data analysis vs quantitative

• focus on meanings

• few rather than many cases

• in depth rather than general categories

• researcher as instrument

• sensitivity to context

• attention to impact of values on analysis

• goal of rich descriptions

steps of qualitative data analysis

1. documentation of data and data collection

2. conceptualization and coding

3. examining relationships

4. authenticating conclusions

5. reflexivity

documentation

• analysis begins with jottings, field notes, transcripts and so on

• essential for keeping track

• provides way of developing an outline for analytic process

• encourages ongoing conceptualizing

conceptualization, coding and categorizing

• may begin with simple observation

• analysic insights tested against new observations and modified

• facilitated by matrix or chart

coding

process of naming or labelling the data, categories and properties

codes

• words, phrases, meaning units, repeated in the data

• divide text into a word or phrase/label

categories

sort codes into categroies based on how different codes are related

themes

identify a central category and relate it to other categories

coding process

code => categories => themes

concept driven coding

• deductive coding

• codes are developed before the coding process begins then look for those ideas in the text/transcripts

• codes could be based on interview guide or on theory or existing literature

data-driven coding

• inductive coding

• no pre-developed system of codes

• let the words “speak for themselves” and then construct coding structure based on your review

• starting from scratch

codebook

• dictionary or guidebook of all codes in project

• provides

  • guidelines for consistent coding across coders

  • name of the code

  • definition/description of the code

  • example of text excerpt that represents the code

examining relationships and displaying data

• centerpiece of analytic process

• move from description to explanation

• can also be diagrammed in chart or matrix

• simple relationships can be extended into more complex causal model

authenticating conclusions

• how credible was informant?

  • were statements made by someone who knows the researcher or someone they just met?

  • did informant have reason to lie?

• were statements made in response to researcher’s questions or were they spontaneous?

• how does the presence of the researcher or the researcher’s informant influence the actions and statements of other group members?

reflexivity

• confidence in conclusions strengthened by account of how researcher interacted with subjects

• sensitivity to influences on interpretations of social situation or process

trustworthiness

1. credibility: confidence in truth of the findings

   • are they reasonable

   • do they reflect participants views/experience

2. transferability: findings have applicability in other contexts

   • other people, time periods,…

3. confirmability

   • neutrality of researcher

   • researcher’s reality vs participants’ realities

   • focus on participants’ realities

   • findings grounded in data

4. dependability: findings are consistent

   • study procedures documented and distinguishable

   • consistency in interview and analysis processes

threats to trustworthiness

• reactivity: potentially distorting effects of researchers’ presence on participants’ beliefs and behaviors

  • social desirability bias/response bias: tendency to survey respondents to answer questions in way that will be viewed favorably by others

• researcher’s biases:

  • choose informants who fit within their world view

  • asking questions to get desired answers

  • ignoring data that does not support conclusions

• respondent’s biases

  • withholding info or lying

  • offer answers they think researchers want to hear

  • ineffective recall

enhancing trusworthiness

• development of coding system

• triangulation

• sample size

• negative case analysis

• peer debriefing

• thick description

• member checking

• persistent observation

• external audits

• researcher’s reflecivity

• prolonged engagement

• audit trail

• peer review

alternatives in qualitative data analysis

• grounded theory

• narrative analysis

• conversation analysis

• visual sociology

grounded theory

- systematic theory developed inductively based on observations that are summarised into conceptual categories reevaluated in the research setting and gradually refined and linked to other conceptual categories

steps of grounded theory

• summarise observations into conceptual categories

• test categories directly in research setting

• refine theory again

• attempt to discover negative evidence

narrative analysis

focusses on how respondents impose order on the flow of experience in their lives and so make sense of events and actions in which they have participated

coding for narrative analysis

• typically of narratives as a whole then of different elements

• revolves around reading stories and classifying them into general patterns

conversation analysis

• method for analysing ordinary conversation

• three premises

  • interactions sequentially organised and talk can be analysed by process of social interaction

  • talk is contextually oriented

  • no interactive details are irrelevant to understand social interactions

• one limitation: can easily be misinterpreted

• the tone/impression of the statement changes depending if the participant is speaking in a sad, angry or joking manner, notes and memos are very important

visual sociology

• social world is “observed” and interpreted through photographs, films and other images

• a method for

  • learning how others “see” the social world

  • creating images of that world for further study

• video recording increasingly popular

  • video ethnography

why are mixed methods helpful?

• reinforce each other

• create greater depth of understanding

• reveal or correct errors in other methods

• fill in the steps in complex social processes

• help offset weakness of each technique

• help understand complex issues

most popular programs for computer-assisted qualitative data analysis

• hyperRESEARCH

• QSR Nvivo

• Atlas.ti

validity threats

• bias: researchers influence on study

• reactivity: researchers effect on setting or people

internal generalizability

do findings extend within the group studies?

external generalizability

do findings extend outside the group studied?

face generalizability

there is no reason to believe the results don’t generalize

ethical issues qualitative data analysis

• research integrity and quality

• ownership of data and conclusions

• use and misuse of results

denzin’s interpretative criteria for qualitative analyses

• does it illuminate the phenomenon as lived experience?

• is it based on contextualized materials?

• is it historically and relationally grounded?

• is the research processual and interactional?

• does it engulf what is known about the phenomenon?