DATA ANALYSIS IN QUALITATIVE RESEARCH In qualitative research

DATA ANALYSIS IN QUALITATIVE RESEARCH

� In qualitative research data collection and data analysis are not as strictly separated as they are in quantitative research � If we discuss data analysis we must discuss data collection as well

QUANTITATIVE RESEARCH Hypothesis Data collection Data analysis Discussion of findings and results

QUALITATIVE RESEARCH Hypothesi s and findings Data analysis Data collection

The structure of qualitative research is not characterized by step-by-step-procedures but by interrelations between hypothesis, research methods, way of data collections, and findings Qualitative research is a permanent learning process – for the researcher and the people he/she is studying

Hypothesis Data collection and analysis

COLLECTING INTERVIEW DATA

�I some aspects, an interview is the most natural thing in the world � A fair proportion of what appears in media consists of interviews � These interviews don’t have a predetermined pattern but are free-flowing and open-ended � We all identify qualitative research with these open-ended interviews � Some experience in qualitative research however demonstrates that things are not so straightforward as me might think

� Defining a research problem � Deciding how many interviews are needed � Designing an interview protocol � Recording and transcribing interviews � Using multiple methods

DEFINING A RESEARCH PROBLEM

� The crucial thing is that qualitative research is based on models how people perceive or construct reality � By using qualitative interviews we want “to look into the head of people” to find out how they experiences or their emotions are � A positivist approach is focusing on experiences, a emotionalist on feelings of the interviewees, on how the interviewees are producing replies to questions.

EXAMPLES

How residents in a community of elderly people feel about their lives? � How do drug users present themselves to keep self respect during the interview? � Factors and contexts of gender discrimination, and their influence of life and employment trajectories. � How do different actors experience the implementation of labor standards? � How do families construct stories about grief and recovery after the dead of relatives? � Responses of patients who consulted unqualified Western practitioners. �

�A political science student was interested in what seemed to be an important social issue: � How water and sanitation policy had been influenced by the decentralisation of Tanzanian local government since 2000. � He was planning to find answers by interviewing key stakeholders. � What is the problem with his approach?

� It is difficult to find out what happened in the past by interviewing people today.

� Reformulate the research problem as a study of contemporary stakeholder views. � Use different data, for example contemporary data of 2000 and today

DECIDING HOW MANY INTERVIEWS ARE NEEDED

FOLLOW THE EXAMPLE OF QUANTITATIVE RESEARCH � In case of small numbers of cases try to select all (10 of 14 practitioners in a district for example). � If a quantitative study has been done already, randomly select a small sample (15 of 160 school teachers)

THE MOST COMMON TECHNIQUE IS PURPOSIVE SAMPLING 16 Ph. D candidates from two different universities and from different departments (psychology and social studies) � Studying gender inequalities, women selected to include various stages of the life cycle (after high school diploma, vocational or university education, pregnancy, children’s education etc. ) � Care providers of children whose mothers work oversea (20 in total, three groups: grandparents, aunts, and others). � Snowball sampling: using the social network of the initial informants. �

DESIGNING AN INTERVIEW PROTOCOL

Usually, prepared questions only serve as guides in qualitative research. � But some researchers prefer sets of more structured questions. � If so, the basic research questions should never be asked directly, as the respondents may be aware of the research problem which can effect the responses. � It can also lead to lazy research in which carefully data analysis is replaced by reporting back what people tell. � It is anyway crucial to pilot an interview schedule. �

RECORDING AND TRANSCRIBING INTERVIEWS

� Simplified Transcription Symbols (PDF)

USING MULTIPLE METHODS

� How can I combine data from interviews with other observations and information? � Non-interview data as guide to structure the interviews. � After completing the analysis of the interviews, use multiple methods to put your findings in a broader context.

COLLECTING ETHNOGRAPHIC DATA

� Access � Narrowing down � Number of cases � Recording observations

ACCESS

� Getting access is the central problem when collecting ethnographic data. Usually access is limited or restricted � In closed or private settings (organizations, deviant groups) access is controlled by gatekeepers � In open or public settings access is theoretically available but often with severe difficulties, either practical or ethical

� Two kinds of research access: � Covert access without subjects knowledge � Overt access based on informing subjects, agreement often through gatekeepers

� Access to collect ethnographic data is mainly based on trust. To increase its level it is recommended to use the following ways: � Impression management (adequate dress, behavior, compliance) � Top-down access in formal organizations � Being non-judgemental � Offering feedback

NARROWING DOWN

� In settings where an abundance of information is available research questions and strategies must become more specific

NUMBER OF CASES

� Consider limiting your study � Consider that even if you are only studying a single case you have already rigorously limited your research � Find the “Black Swan”. As Popper has reminded a single observation of a black swan already falsifies the hypothesis that all swans are white. Falsification is a central element of science and of qualitative research in particular

RECORDING OBSERVATIONS

� Record what you see, not what you think! � Record how you are being treated!

DEVELOPING DATA ANALYSIS

� For data analysis in qualitative research the following steps of are common �

Analyse of data already in the public sphere Analyse your data when gathering them Published Key questions statistical material, Permanent published reports, evaluation of What, which, papers etc. hypothesis, how? methods and questions

ANALYSE YOUR DATA AS YOU GATHER THEM

� Ask critical questions: � Do I feel comfortable with my preferred method? � Is my data analysis suggesting interesting questions for further research and discussion? � Can I generate interesting generalizations? � Do previous findings apply to my data? If not, why not?

ASKING KEY QUESTIONS ABOUT THE DATA OBTAINED

� What are the main units in my data and how are they interrelated with each other? � Examples: Children, villages, school classes, doctors, administrations, farmers associations etc.

� Which categories are actually used by the people we are studying? Qualitative research does not impose abstract scientific language on people. � How does a farmer speak about problems, which categories and words is he using? How do children describe their learning experiences?

� In which context do our subject use which categories and language? � How does a farmer or student speak at home and at school?

� How can I use difficulties in my research (access, wrong questions etc. ) to improve my study and to which potential these difficulties do have for further research? � Why nobody is talking with me? Why I don’t get access?

ANALYSIS OF CASE STUDIES

� Start with analyzing data which are of high quality and easiest to collect. � Look at one process within these data. � Narrow down to one part of that process. � Compare with different subsamples.

ANALYSIS OF INTERVIEWS

� Start with analyzing findings which are of high quality and easiest to understand. � Treat the different voices of your interviewees and differences in their responses rather as a finding than a problem. � Describe the range of voices and responses of your interviewees. � Examine how these voices are articulated in interrelation to each other and describe the interactional contexts in which this happens.

ANALYSIS OF FIELDNOTES

GENERAL IDEAS � Identify and follow processes in witnessed events � Understand how members themselves characterize and describe particular activities, events etc. � Convey interpretations for when, why or how things happen � Identify practical concerns, conditions and constraints of peoples everyday life

IDEAS FOR FIELDNOTES � What people, events or situations were involved? � What were the main themes or issues in the contact? � Which research questions did the contact bear most centrally on? � Are there new ideas, hypothesis etc. after the contact? � What should the fieldworker do next?