Interviewing Methodology of CDS Fall 2020 Doc. Marek Rybář, PhD. Some of the topics reviewed in the lecture •Types of interviews in political science research •What are interviews good for? •When and whom to interview •How to select respondents •What are interview protocols •Capturing information and conducting interviews •Missing data •Gaining access to respondents •Reflections on the role of interviewers Types of Interviews •the most common method of data collection for political scientists who engage in fieldwork •positivist and interpretive traditions •three types of interviewing: •in-depth interviews with individuals, •oral histories, and •focus group interviews In-depth interviews Oral history interviews Focus group interview No. of participants 1-2 1 6-10 Role of researcher provocateur active listener facilitator Degree of structure from loose to highly structured low; researcher identifies starting point and participant guides narrative flow moderate; researcher puts questions order shaped by participants Question type and ordering open ended; variable order open-ended; chronological narrative a list of open-ended question; order shaped by group Advantages unique perspective; sensitive information unique historical perspective and causal process reveal tensions and disagreements Disadvantages specificity of viewpoints; inefficiency particularity; represented as unscientific group think What are interviews good for? Formulating research questions & case selection •useful for doublechecking that the ways we frame our research questions are correct, •and for evaluating whether one's project is original, compelling and possible to execute •many researchers carry out informational individual or even focus group interviews •interviews can help scholars to select cases for study that will allow testing the most promising hypotheses • What are interviews good for? Conceptualization, measurement & generating hypotheses •they can help to identify contested nature of important political concepts: respondents may (implicitly or explicitly) suggest a range of understanding connected with a particular term •respondents can illuminate important causal factors, aiding scholars in developing their arguments •we can ask the respondents why a phenomenon occurred as it did to get plausible hypotheses: some responses can be directly converted into an argument When and whom to interview •it is good to postpone interviews with high-profile political actors until we have solid baseline knowledge of the topic through interviewing academics, journalists, and others who study the phenomenon •such softer interviews can help to identify which actors to interview, assess available data, refine interview questions, and construct better protocols and interviews guide •whom to interview depends on our interviewing objectives, practical considerations, and the field context itself • How to select respondents •quota sampling: we set proportions so a sample includes certain segments of the population •purposive sampling: selection based on characteristics of respondents that are relevant to the analysis •snowball sampling: an initial set of respondents suggest further respondents •convenience sampling: interviewing those who are available and agree to be interviewed • Interview protocols •we need to draft the protocols well in advance, to gain intense familiarity with the question guide •1. questions should be theoretically motivated but expressed in colloquial language free of jargon •2. it is good to identify colloquial language that is appropriate in the given context •3. pre-testing is also desirable: ideally, pre-tests are done on a sample of the target population •4. most researchers' protocols change over the course of the fieldwork; we need to explain and document consequences of changes to our protocols • Conducting interviews and capturing information •interviewer effects, rapport, and the ethics of interviewing •recording interviews vs. taking notes •transcribing is time-consuming but usually indispensable for subsequent analysis •it is useful to prepare a coding scheme even before the first interview; •coding schemes are modified as the interviews become more numerous but it is useful to have in mind some theoretical construct of what we expect to gain from the interviews • Types of questions (Beth L. Leech) •grand tour questions: ask respondents to give a verbal tour of something they know well; it gets them talking but in a focused way (Could you describe a typical day in your organization?) •example questions: similar to grand tour but more specific ("Can you give me an example of a time that you used grassroots lobbying?") •structural questions: ask respondents to verbally structure their world by listing all the different types of something and how they relate to each other (e.g. "I am interested in getting a list of all the different types of advocacy activities your organization has undertaken") •prompts: keep people talking, and they help when responses are unsatisfactory Validity and reliability of interview data •it is not the obligation of respondents to be objective and to tell us the truth, we may try to •1. use multiple sources (extremely time-consuming) •2. move the respondents away from their case to the politics of the situation, or using a third party [a newspaper], as a way of taking the respondents away from their own perspective •if our respondents are "obviously" biased, we may move to questions where we can turn their bias into something beneficial to our research • Missing Data •respondents who declined to be interviewed and the consequences for resulting data bias •the least worrying in elite interviews - we have ways of assessing how rejection of a particular (group of) respondents affect our data, because we know a great deal about their attitudes and behavior from other sources •missing interview data leading to biased results must be openly acknowledged and built into the conclusions • Gaining access to respondents •respondents are sometimes eager to speak with researchers, yet most scholars experience challenges in accessing the people they want to interview •we may be unable to locate those whom we wish to interview •respondents may feel uncomfortable being interviewed •reaching interviews may be difficult in authoritarian or highly controlled or dangerous environments •not being able to reach desired respondents can have analytical implications • Gaining access to respondents •if we seek to develop claims that are generalizable to a larger group on the basis of interviewing a sample of that group •always assess precisely how necessary it is to speak with the individuals we have difficulty to interview: •How will not reaching them affect the quality and quantity of our evidence? •if they are essential to your research, identify a connection to them: an associate, relative, friend, colleague, employee • Gaining access to respondents •attempt to enter the network in which targeted respondents operate (types of events, mailing lists etc.) •interviewing "ordinary people" may seem easier than interviewing political elites but it is not necessarily the case •elites are used to talking to people, why ordinary people may have little motivation to convey personal information • Conducting Interviews •deciding upon location •whether to send interview questions in advance •evaluating the need to offer gifts to respondents •introducing oneself and one's work •probing •asking follow up questions •wrapping up interviews or reacting to its early termination by a respondent •requesting additional contacts •gaining rapport How to approach the role of the interviewer •e.g. several scholars remark how being married, or having children, can elevate or diminish their status and facilitate or complicate mutual respect in their interviews •positivists and interpretivists differ in their views on how such effects shape the interviews exchange and influence the data derived from it, and what should or could be done about it •interpretivists explicitly acknowledge the impact identity has on the research experience, positivists actively seek to identify and estimate how their identity affects data collection •