This study aims to reflect on qualitative interviewing with a particular emphasis on semi-structured interviewing (SSI), with the purpose of guiding students and young scholars of International Relations and Political Science who will use this method in their research. This study begs to differ from both radical postpositivist’s deep scepticism which makes any scientific inquiry almost impossible as well as from positivism’s unreflective, unproblematized, instrumental approach to interviewing. It proposes a reflectivist approach to qualitative interviewing that emphasizes the political nature of the interviewing process with various political, ethical and even social consequences. The reflectivist approach requires researchers to be self-critical at all times, in particular concerning their role and influence on the interview setting and the interviewee. This article proceeds as follows: It first addresses my own research on the nexus between civil society and the Kurdish question in Turkey, where SSI has been operationalized as the main research method. It then addresses the positivist and post-positivist debates on qualitative interviewing as well as the reflectivist approach that this study promotes. The article then engages in SSI in three distinct stages: pre-interview, interview and post-interview phases. Finally, the concluding part introduces some works utilising interviewing in Turkish IR and wraps up the theoretical/methodological arguments disseminated throughout the study at hand.

 1.Introduction

This study aims to reflect on qualitative interviewing as a frequently used method for data gathering in different disciplines of social and human sciences. It particularly focuses on semi-structured interviewing (SSI) with the purpose of guiding students of International Relations and Political Science who would like to use this method in their qualitative research. Interviewing with its particular forms has long been popular among social science researchers to obtain up-to-date and first-hand data. To Rapley, qualitative interviewing has even become ‘the central’ research tool for social scientists.[1] The rising interest in qualitative interviewing has triggered scholarly discussions concerning its strengths and limitations as well as concerning whether it is useful on its own or should only be employed as complementary to other methods.

I have widely benefited from SSI in my research particularly to comprehend the ways in which a protracted conflict or long-lasting public debates are understood and discursively framed by civil society groups vocal on these debates. I first employed SSI throughout my post-doctoral research (from 2006 to 2008) on securitization of different aspects of the Cyprus question by the conflicting parties. Using the multi-sectoral approach to security introduced by Buzan et. al., this research aimed to map out different discourses of (in)security articulated by different social and political actors in both parts of the island of Cyprus. Yet, as will be clear below, I have also adopted SSI in a more conscious and systematic manner in my research projects concerning civil society organisations (CSOs) in Turkey involved in public debates on the Kurdish question.

For almost two centuries positivism is known to have had great faith in empirical/observational research. To this approach, data collected through an ideal, objective interviewing process will reflect the reality and can be used to testify or falsify hypotheses and theories. Yet, this claim to objectivity is challenged by a multitude of post-positivist approaches that ‘deny science any privileged access to the objective truth about the social world’.[2] Post-structuralists argue that language and discourse and thus the conversational material collected through interviews construct rather than reveal phenomena. However, the radical post-structuralist critique, which is deeply sceptical of any conventional scientific inquiry, is often viewed as ‘categorical’ and even ‘destructive’[3] and hence, is largely overlooked by the scholarly literature on the issue.

In this study I beg to differ from both radical post-positivist’s deep scepticism which makes any scientific inquiry almost impossible as well as from positivism’s unreflective, unproblematized, instrumental approach to interviewing. Following Alvesson,[4] I propose a reflectivist approach to qualitative interviewing in general and to SSI in particular, problematizing and challenging the dominant positivist or neo-positivist assumptions, but without entirely negating them as some post-structuralists do. The reflectivist approach emphasizes the political nature of the interviewing process with various political, ethical and even social consequences. Every research may serve to the interests of distinct social groups and may serve to the reproduction or transformation of certain power relations from micro to macro levels. Hence, researchers need to reflect on potential beneficiaries of their project as well as question whether their project can ‘improve the lives of human beings in any way’.[5] Thus, in the reflectivist approach, the role and impact of the researcher goes well beyond asking questions and reporting answers. The reflectivist approach requires the researcher to be self-critical at all times, in particular concerning his/her role and influence on the interview situation and the interviewee. It suggests that both the researcher and the participant are social/political actors acting with different identities, motivations and expectations.

Against this background, this article proceeds as follows: It first addresses my research on the nexus between civil society and the Kurdish question in Turkey, in which SSI has been operationalized as the main research method. It then addresses the positivist and post-positivist debates on qualitative interviewing as well as the reflectivist approach that this study presents. The article then explores the three distinct stages of SSI: pre-interview, interview and post-interview. The discussion about each of these phases includes my personal reflections and refers to concrete examples from my previous research. The ethical issues that researchers need to tackle will be addressed as well. Finally, the concluding part introduces some works utilising interviewing in Turkish IR and wraps up the theoretical/methodological arguments disseminated throughout the study at hand.

2. My Research on Civil Society and the Kurdish Question

I first employed semi-structured interviewing as a systematic research tool while I was part of the SHUR project (‘Human Rights in Conflicts: The Role of Civil Society’) funded by the European Commission within the context of the Sixth Framework programme. When I was invited to the project it was already at its fieldwork stage and I was expected to organize the envisaged interviews including specifying interviewees, preparing the questionnaire, and doing other organizational issues. I already had some limited experience of interviewing, yet I had to learn the nitty gritty of each stage of SSI throughout the fieldwork process. One may certainly gain insightful data from methodology books and articles concerning different kinds of qualitative interviewing, yet I found, unlike with various other methods, a researcher is best trained for qualitative interviewing while practicing it. That being said, training may be most useful for the post-interview phase, in which the conversational data are transcribed, analysed and reported. The transcribed data may be analysed through either relatively conventional methods (content analysis, discourse analysis) or through more recent computerised methods of coding utilising such software as Nvivo or SPSS.[6]

There are a number of ‘universal’ rules and tricks for conducting interviews, most of which are covered in this article. However, this is also a very subjective process of learning for each researcher that may change depending on his/her expectations, ways of doing things, and the specific characteristics of the interviewees. Qualitative interviewing is unique as a research method in the sense that interaction with human beings is the main determinant at every step. Subjective experiences, perceptions, political/social motivations, formulation of questions and the narrative used by interviewees and some other incalculable factors all make interviewing a unique and context-bound experience for the researcher. Therefore, the researcher needs to be open to learning in each case of interviewing as an attentive observer. Hence, self-reflexivity at all times is a key principle for the researcher while learning and practicing this method of inquiry.

The main task of the SHUR project was to understand the role and impact of civil society organizations in ethno-political conflicts through the comprehensive and cross comparative analyses of four different cases: Israel-Palestine; Bosnia-Herzegovina; Turkey-Kurds; and the Cyprus dispute. I was part of the team working on the Kurdish question, and I co-authored a publicly accessible report and an article with Nathalie Tocci on the role of civil society in Turkey’s Kurdish question.[7] In this report and in the subsequent journal article we examined the multi-faceted roles and impacts of Turkish and Kurdish civil society organizations (CSOs) in Turkey’s Kurdish question, i.e. fuelling the conflict, easing the subjective incompatibilities, constructing dialogue.

To fully grasp the different manifestations of the impact of civil societal actors, we proceeded through a four-phased methodology. Having critically engaged in the academic and non-academic writing on the Kurdish question and Turkish civil society and having set the theoretical/conceptual framework of the research, we specified CSOs which were vocal on the Kurdish question and came from different socio-political, ideological and social backgrounds. For analytical clarity, the CSOs were also classified mainly according to the political stance and position they held throughout the debates on the Kurdish question, i.e. anti-establishment/nationalist Kurdish CSOs and establishment/Turkish nationalist organisations. As I will mention below, this necessitated in the pre-interview phase a detailed examination of the CSOs and their discursive/deliberative positions in the conflict. In the third stage our fieldwork mainly relied on semi-structured and in-depth interviews with the representatives of the CSOs and other relevant activists specified in the previous stage. The questions about the impact, strategies and deliberative positions of the CSOs were also cross-checked both with other CSO representatives (involved in similar fields) as well as with academics and officials. In the fourth and final phase, the empirical data gathered through a detailed discourse analysis of the transcriptions of these in-depth and cross-checking interviews were assessed in the light of our main research question. This question was whether particular actions and discourses of the CSOs involved in public debates on the Kurdish question contributed to securitization, de-securitization or non-securitization of the conflict.

In a more recent study elsewhere, I extensively used SSI with the aim of exploring the impact of Europe/Europeanization on the politically mobilised civil society vocal on Turkey’s Kurdish question. In this study the impact of Europe was conceptualised as a complex constellation of top-down and bottom-up processes. The EU-induced legislative reforms (EU-ization) on the development of civic/political rights and fundamental freedoms were defined as the top-down impact of Europe on CSOs. The bottom-up impact emerged through the usage of European norms, policies and institutions (Europeanization) by CSOs to frame and justify their deliberative positions and to increase their mobilising power and visibility. Thereby, Europeanization rather than being a mere process of adaptation to the EU, referred to a wider normative/political context; a context where European norms, policies and institutions are (re)-negotiated by different European societies and institutions, and have an impact on them.[8] This context exists to the extent that European norms, values and institutions are incorporated into public narratives by domestic actors, e.g., political parties, media and CSOs.[9]

Later in this study I tried to understand how the CSOs react to and make use of the European context (Europeanization) to increase their influence and to promote their political agenda. The aim was to grasp whether and/or how CSOs that are actively involved in large public debates formulated their political demands and deliberative positions by making reference to specific European norms, policies and institutions. The study also examined the views of CSOs on the potential roles and limitations of the EU in the Kurdish question and the crippled peace process which lasted between March 2013 and July 2015. It also revealed the historically changing meanings of the EU/Europe for Kurdish CSOs from the 1990s up to the present time. Hence, another empirical goal of the research was to map out the variety of discourses concerning the EU and Turkey’s integration into Europe disseminated by the politically mobilised civil societal actors in Turkey.

Those two research projects heavily drew on several interviews that I conducted in Diyarbakır, Ankara and Istanbul with the participation of CSO representatives, activists and academics. The interviews were organized as semi-structured and open-ended and they were all tape-recorded with the consent of the interviewees beforehand. Interviews for the first research on different impacts of civil society on the Kurdish question, were made in February 2008 in Istanbul and Ankara and included individuals from both Turkish and Kurdish CSOs as well as human rights organizations. The fieldwork for the second research on the Europeanization of civil society vocal on the Kurdish question mainly covered people from Kurdish civil society with diverse social and political backgrounds as well as from several human rights organizations. Some Istanbul-based Turkish CSOs and activists, though limited, were also interviewed. Interviews for this research were mainly conducted between February and June 2014 in Istanbul, Ankara and Diyarbakır. Since SSI allows maximum interaction between the researcher and the interviewee, the two research processes became very instructive for me concerning the nature of the Kurdish question as well as the civil society involvement in the issue.

 3.The Reflectivist Approach to Qualitative Interviewing

Interviewing has long been a popular method for qualitative research in social science as it provides a direct link between the researcher and ‘the very object he/she is investigating’.[10] If, as suggested by Brinkmann, ‘the human reality is a conversational reality’[11] interviewing seems a very advantageous method of data gathering for researchers. One needs to make a distinction between a more optimistic positivist and more cautious and at times even pessimistic post-positivist approach to interviewing. The positivist approach tends to see interviewing as a very useful instrument of empirical research that provides insightful and first-hand data that may not be gained otherwise. In this approach, ‘a rich set of accounts of the interviewees’ experiences, knowledge, ideas and impressions may be produced and documented’ through interviewing.[12]

By drawing on interviews, the researcher can gain insightful data into such things as people’s subjective experiences, perceptions, group attitudes, which may be less or entirely inaccessible through other methods. As observed by Alvesson, for the positivist approaches, interviewing is an effective research tool or human encounter in which a knowledge transmitting logic prevails, and language is a transparent medium for communication of insights, experiences and facts. Moreover, it assumes the interviewee is motivated by a desire to assist science and is called upon in a sufficiently well-structured or secure and personal way so that pretence and role play do not matter much, therefore true and authentic answers are provided, and the interviewee can be mobilised as an integrated source of meaning, knowledge and intentionality and so on.[13]

The researcher takes every step ‘in order that the respondent speaks openly, authentically or truthfully to produce valid reporting on some interior or exterior state of affairs’.[14] Positivists conceive language as a neutral transmitter that mirrors the reality told by the interviewee who often tells the truth in an objective manner. The researcher, who is viewed as a neutral transmitter of that truth, should do everything to minimize his/her role and other sources of bias.

On the other hand, as Alvesson suggests, an interview is a ‘social situation and that which is said is far too context-dependent to be seen as a mirror of what goes on outside this specific situation’.[15] As post-positivists argue, the interview situation is not independent from power relations and interest calculations. Therefore, it must be viewed as a political act with non-trivial consequences. The interviewee or the participant may ‘act in his or her interests and or the interests of the social group with which he/she identifies’. Unlike positivist or neo-positivist approaches, interviewees are not seen solely as honest, moral story tellers or neutral information givers, but seen as politically conscious and ‘politically motivated actors’.[16] Interviewees may exploit interviews for self-promoting or ‘for their own political purposes, they may cheat or lie or they may very well tell the (partial) truth as they know it but in, for them, selective and favorable ways’.[17]

Post-structuralists argue that language and discourse and thus the conversational material collected through interviews construct rather than reveal phenomena. While positivists tend to see the interview as an effective research instrument, post-positivists are quite sceptical about the idea of using interview data.  I would argue that the extreme arguments of both sides are flawed, rather, following Alvesson,[18] I prefer a reflectivist approach to qualitative interviewing, one which problematizes and challenges the dominant positivist or neo-positivist assumptions, but without entirely negating them as some post-structuralists do. A reflectivist approach emphasizes the political nature of the interviewing process ‘as a socially and linguistically complex social situation’[19] with seminal political, ethical and even social consequences. By the reflectivist approach I refer to thinking carefully and persistently at all stages of interviewing, i.e. the persons and institutions to be interviewed, the questions to be posed, the conditions of the interview situation, the materials to be included and/or excluded in the analysis, and the potential beneficiaries of the research outcomes. Thus, to the reflectivist approach, the role and impact of the researcher goes well beyond asking questions and reporting answers. Rather than being an objective and neutral reporter, the researcher acts as a political subject with a capacity to intervene and configure every stage of interview situation including the post-interview process. Therefore, the researcher has to be self-critical at all times, in particular concerning his/her role and influence as a social-political actor both during and after the interviews. This reflexivity should be at its maximum when interviews are conducted with the disadvantaged sectors of society, i.e. the children, the elderly, the disabled and others who are discriminated against, silenced and marginalised in one way or another.

The reflectivist approach also suggests that interviewees or participants may not always be motivated purely by a scientific reason nor presenting always the most accurate account of what is asked to them. They are ‘frequently political, conscious actors’[20] who may tend to give misleading or partially accurate information. The participants’ social status, economic and political expectations, ‘job related tasks and duties, personal objectives and plans, worldviews and ideologies’ may all influence their accounts throughout the interviewing process.[21] These and other similar subjective factors influence in varying degrees what is or what is not told, what is highlighted or what is omitted/hidden, what is prioritised and what is trivialised by participants. It is not possible for the researcher to isolate the conversational material from those factors and to attain pure, objective data. Yet, it is possible to explore certain argumentative threads, typologies of representations and perceptions, self/other conceptions, policy formulations, deliberative positions, strategies and actions adopted by different actors, in different political-social contexts and time periods.

3.1 Discussing semi-structured interviewing

Qualitative interviews may be differentiated in terms of their ‘structure, the number of participants in each interview, different media and also different interviewer styles’.[22]  These can be formal, informal, structured, semi or unstructured, focus group and telephone interviews. The researcher may either rely on one distinct form of interviewing or in some cases may exploit different kinds of interviews simultaneously in a way as to complement each other. Each has different strengths and weaknesses. This article, rather than presenting an exhaustive discussion of these, limits itself to semi-structured interviewing as probably the most frequently used form of qualitative interviewing in social science research.[23]

Nevertheless, it is worth elaborating on the notion of structuring interviews to comprehend better the distinctive characteristics of SSI. Apart from the standardised, questionnaire/survey kind of interviews that are generally associated with quantitative interviewing, there is no completely structured qualitative interviewing. Completely unstructured interviews where, for instance, only one question is prepared in advance are also seen in very exceptional cases. Thereby, we can rather speak of a ‘continuum ranging from relatively structured to relatively unstructured formats’.[24]

SSI is structured around data that the researcher wants to obtain and may well be reconfigured in accordance with the statements of the interviewee. Far from being a passive recording of what is told or experienced, it takes place as a social interaction between the researcher and the participant. Drawing on the dialogue between these two, SSI recognizes enough space to the interviewee to emphasize the issues that she deems important. It also gives the interviewer ‘a greater chance of becoming visible as a knowledge producing participant in the process itself rather than hiding behind a preset interview guide’.[25] The researcher comes up with a list of questions. Yet, this is not an exhaustive list and depending on the course of the interview, new questions may be added, some others may be removed or the formulation of others may be modified.

As mentioned before, this study will examine SSI in three stages: these are the pre-interview or preparatory stage; the fieldwork or interview stage; and the post-interview or analysing and reporting stage. While delineating each of these phases, the emphasis will be placed on realising the full potentials of SSI and overcoming some of its limitations. Yet, it should be noted that these stages need not be conceived as necessarily sequential steps of research. On the contrary, these are often ‘overlapping and cycling’ stages,[26] as researchers may, for instance, need to conduct new interviews after having completed and analysed a previous set of interviews. The conducted and analysed interviews may necessitate new ones or the researcher might be compelled to renew his/her pre-interview preparatory work. Thus, unlike what the three-stage conceptualization implies, SSI should not be imagined as a linear process of research necessarily advancing stage by stage.

3.2.The pre-interview phase

 Before embarking on the fieldwork, meticulous research on potential interviewees is needed.  The researcher needs to attach due importance to familiarise him/herself with the relevant actors, since the selection of the correct interviewees is of significance for time saving and accessing relevant data. The relevant participants may be identified through a detailed scanning of online and printed sources on the research topic. Yet, the researcher should not hope to have an exhaustive list of interviewees at this early point. As will be discussed below, new interviewees are likely to be added to the list by recommendations throughout the interviews.

However, designation of interviewees is not only an issue of time saving or accessing relevant data. It has normative and political implications as well, since only selected interviewees or organizations will have the chance to make themselves heard and to influence the research outcomes. The selection of participants will determine ‘whose worldviews, opinions and interests will be taken into account and whose will be ignored and excluded’.[27] The feeble civil society literature in Turkey has long remained confined to the usual suspects, i.e. business associations, and mostly Istanbul-based CSOs and organizations with international experience and networks. The CSOs which have been pushed to the margins of the political system due to their social/political activism or the CSOs which operate outside the three major cities, e.g. Istanbul, Ankara or Izmir, and operate with inadequate personnel and financial facilities have often been overlooked by this literature. This has resulted in the over-representation of the usual suspects at the expense of others. To avoid this, Caputo suggests interviewers need to do everything to make sure that ‘the debate is fair’, ‘no one’s voice excluded or demeaned’, and ‘the vested interest of the powerful who usually end up having their way are restrained’.[28] Caputo certainly defines an ideal situation which probably can never be fulfilled in practice. Yet, it is still important for ethically and politically engaged researchers who wish to represent the voice of the unheard, the silenced or the marginalized rather than contribute to the reproduction of the existing power relations within the civil society sector.

Researchers who have already contacted potential interviewees due to previous research, will be in an advantageous position to access relevant organizations and persons. For instance, when I started my research on Europeanization and Kurdish civil society, I had already established an extensive web of contacts with relevant actors due to my previous research in the SHUR project. This enabled me to access relevant organisations and persons more conveniently. When potential interviewees know you as a researcher developing a real interest in their issues and motivated by purely scientific ambitions, they often agree to arrange an interview more easily. Building a relationship of trust is bound to impact the content of the interviews as well. In such cases, conversations often develop more fluently and participants tend to speak in a more sincere manner. Mutual acquaintances between the interviewer and the interviewee or approaching the latter via a reliable reference bears particular importance while conducting research in cases as delicate as the Kurdish issue in Turkey.

Meticulous research at the pre-interview stage may also provide the researcher some insightful information on the actions, discourses and identities of the relevant actors as well as on the web of relations among them. If there are activists, NGOs, journalists, political leaders among potential interviewees, it would be very useful to read and examine what they have written and what they have publically told in relevant topics for the research at hand. A careful analysis of academic and non-academic publications, news analyses and commentaries in newspapers, reports and press releases by civil society groups, research centres, as well as public declarations and official documents in this particular area would provide necessary and useful data to the researcher to determine both who is to be interviewed and how interviews can be structured. If the researcher is expected to interview a large number of interviewees and if the aim of the research is to understand how a particular issue or debate is framed by different groups of actors, a meaningful classification of these actors would be needed at this stage. This helps the research to understand common and differentiated argumentative threads among these actors. For example, in my first research, which aimed at assessing the impact of CSOs with different identities, strategies and actions in the Kurdish question, we classified CSOs mainly according to their political stance and the positions they held throughout the debates on the issue, i.e., Turkish nationalist/establishment or Kurdish anti-establishment CSOs.[29]

Yet, it should be noted that this kind of classification is in fact a subjective intervention on the part of the researcher and may always be deficient or misleading due to different reasons. First, assessing the agendas and declared or undeclared intentions of potential interviewees or organizations that would be covered in the research may not be possible. Second, different actions and discourses of the same organization or interviewee could fall into different categories or may dovetail with different group of actors. To illustrate, in the same research on civil societal involvement in the Kurdish question, we observed that while the Human Rights Association has often made a non-securitizing impact, ‘some of its actions and discourses had a securitizing impact even if the organization did not aim to do so’.[30] Therefore, instead of taking a CSO as a homogeneous actor, we rather focused on the ways in which particular actions of particular CSOs have moulded the Kurdish question.

3.2.1.Ethical issues and taking consent of participants

The researcher needs to engage in all ethical issues in the pre-interviewing stage that may arise throughout the research. It has increasingly become a common practice that institutions funding research projects demand researchers to present a report addressing potential ethical issues which in turn must be approved by an ethics committee at the local (e.g. the host university) or at the national level. As qualitative interviewing is conducted through human to human interactions, addressing ethical issues is an essential part of the process. As a first step, the researcher may send an information sheet to interviewees via email or by post covering the necessary information on the details of the research, e.g. what the aims and objectives of the investigation are, what kinds of information will be asked from them, how and in what ways the data provided will be used, protected and destroyed when the research cycle has ended. The information sheet should clearly explain to participants the measures taken to ensure the privacy and confidentiality of the data provided in their native language. Should the participants agree with these terms, they may be asked to give their oral or written consent prior to the interviews. Through this procedure, the participants will understand the issues at stake in the research project before giving their consent.

Informed consent could be either in oral (e.g. by phone) or written form. Although the latter is much easier to document for the researcher, the preference could be left to the interviewee to minimize any discomfort. In case of an oral consent, it needs to be witnessed by someone such as another colleague or participant or the research assistant of the project if there is any. The written consent may be provided either via email from the interviewee stating his/her consent to participate in the research or via a written statement by the interviewee. The participants, however, will be free to withdraw their consent for being interviewed any time without liability if they deem it necessary. They have also every right to withdraw from the interview whenever and for whatever reason they wish. The questions posed in interviews should not invoke political or legal sensitivities and the participants will be free to avoid any question they do not wish to answer.

The researcher may fail to envisage and prevent all risks the participants may face due to the content of their answers and the opinions they express. Yet, s/he can make every effort to minimise these risks such as stigmatisation or discrimination of interviewees due to their participation in the research. The most effective way of doing this to obtain their approval before using the data they have provided in the publications. Under certain circumstances it may not be possible to know how the data obtained during interviews will be used in outcomes/publications and it would not be sufficient to inform the participants at the very beginning of the research. In order to minimize this risk, the researcher needs to maintain ongoing communication (oral and written) with the participants. Particularly, in the cases where direct quotations from an interviewee will be used, the researcher has to ensure in writing that they agree with these quotations as they will appear in publications.

In cases where the participants require anonymity, it is again the responsibility of the researcher to make sure that no link can be established between the identity of the participant and the data obtained. Anonymity is the key principle in addressing the privacy issues. The researcher guarantees that s/he will not register elements that could permit the identification of the individuals’ identity by a third party external to the relationship between researcher and interviewees when anonymity is requested. Any potential risks should be minimized by treating the data collected through interviews as confidential and, if required by the participant, as anonymous. It is the researcher’s responsibility and duty to safeguard the privacy and confidentiality of the data.

Moreover, the ways in which interview questions are formulated could be helpful to deal with potential ethical issues in a pre-emptive manner. For instance, throughout my research drawing on interviews I have:

• focused on the CSOs as entities rather than on the personal data or information of the individual research participants;
• examined the activities, public statements and publications of the CSOs rather than trying to understand the intentions, individual convictions or initiatives of their individual members;
• avoided contentious questions during interviews which may lead the participants to reveal their personal religious or political convictions;
• avoided questions that may reveal some personal information about participants, e.g. political conviction or political party affiliation, ethnicity, religion.

When interviews are conducted with such vulnerable groups as children, elderly, disabled or irregular migrants, additional measures may be required not to give any damage to interviewees. Under no circumstances should the interviewees be forced to give any confidential data regarding their institutions or themselves.

Interviews may be recorded or the researcher may take notes during the interview depending on the preference of the interviewee. In any case, the data obtained throughout interviews should only be accessed by the researcher and should be stored on a password protected hard disk to avoid the risk of access by third parties. The data obtained throughout the research should not be used in any context other than the stated purposes of the project. The collected data should only be accessed until the end of the research project and should be destroyed irreversibly by the researcher. Particular emphasis needs to be placed on any personal data and information which will exclusively be available to the researcher. Any printed or electronic material including the personal data and information should be destroyed by the researcher following the end of the project.

3.3. The interview phase

The fieldwork constitutes the core of the semi-structured interviewing and it is this stage that the role of the researcher as a subject becomes more evident. The interview situation should be viewed as a social interaction between the researcher and the participant rather than a mere case of asking questions and recording answers. As an intersubjective process, qualitative interviewing is deeply influenced by personal characteristics, i.e. internalised norms, cultural scripts, biases or worldviews of both the interviewer and the interviewee. Thereby, unlike what positivists often suggest, there is ‘no such thing like a neutral, non-intervening and non-existent’ or objective researcher/interviewer.[31] Face-to-face interviewing enables and even is more extensively shaped by the interpersonal contact between the researcher and the participant. Even if this is largely accepted as an asset of SSI, this could also be its Achilles heel as it highly increases the likelihood of the researcher and the participant influencing each other.

The researcher is always an active part of qualitative interviewing, and his/her role goes well beyond asking questions and noting the answers. The ways in which s/he formulates the content, language and even the order of questions have incalculable effects on the reactions of the interviewee. The interviewer often makes conscious or unconscious interventions so that the interviewee ‘makes statements he or she would not make otherwise’.[32] In 2007 I conducted research focusing on the discourses of (in)security devised by different socio-political actors who are vocal in the Cyprus conflict. Drawing on the Copenhagen School’s multi-sectoral approach to international security,[33] I aimed to explore which issues are securitized in what ways and by whom in both parts of the island of Cyprus. When I asked such questions to my interviewees as ‘do you think threats to security are limited to military issues on the island? Which issues are of urgency for the security of your community and for the whole island? Do environmental, economic or identity related problems constitute threats to security?’ I noticed that interviewees frequently tended to formulate their answers in security terms. They were more inclined to inject a security quality to issues political, economic or cultural in nature. As observed by many, interviewees tend to say what interviewers want to hear.

However, when I formulated my questions in more neutral terms, the participants became less inclined to securitize issues. They expressed their views in much more differentiated ways, which eventually convinced me to alter the language of my questions. To elucidate, I avoided using such terms as security, securitization, threat, danger/dangerous in my questions which were reframed as follows: “Do you think which issues or problems should be on the priority list of political parties and governments on both sides of the island? How do you define impacts of environmental, economic or identity related issues on the Cyprus conflict? What kinds of instruments should be used to minimise negative impacts of these issues while enhancing their potential to make positive contributions to any prospect of settlement?” This reframing allowed interviewees to mention the security-ness of an issue (if they think so) or to discuss it in entirely different ways free from any influence of the interviewer.

To Brinkmann, research drawing on qualitative interviewing is more suitable for ‘how’ questions rather than for ‘why’, ‘how much,’ or comparison questions necessitating large number of interviews for any statistically meaningful result.

A question like “How do people cope with the loss of a loved one?” is in general better for qualitative projects than questions that seek to find causal effects, such as “Does psychotherapy reduce the risk of depression of the loss?” The latter question is interesting and relevant. But, it is also extremely difficult to answer with qualitative interviewing. Instead, one would need to enlist a large number of research participants, administrate standardised tests and compare the effects statistically in order to assess whether the findings are statistically significant, i.e., not just a chance result.[34]

SSI is very useful to explore how particular issues, concepts, political developments or cultural phenomena are experienced, perceived and discursively framed by actors in the public or private spheres. It may also be relevant to know what kinds of actions and strategies these actors develop in relation to particular issues and problems. Even if SSI does not allow for statistical work, it allows researcher to ‘construct a typology of experiencing, reasoning or acting’ among the actors covered in the research.[35] For instance, when we searched for how the civil society in Turkey impacts on the Kurdish question, it became very useful to categorize CSOs in a way as to understand if there exist similarities and meaningful differences regarding the impacts, actions and strategies of different CSOs in particular stages of the conflict.

The CSOs demanding the recognition of a separate Kurdish identity and collective cultural rights, and denouncing the Turkish state’s violations of human rights were categorized as the ‘anti-establishment’ CSOs. The CSOs which, to a greater or lesser degree, adopted the official state line and blamed anti-establishment CSOs of being pawns in the hands of the PKK [The Kurdistan Workers’ Party] were defined as ‘establishment’ organizations. In between these two extremes, we placed the CSOs whose identities and actions did not entirely distance them from the Turkish state, but nonetheless challenged the hegemony of the state’s secularist and ethno-nationalist policies’.[36]

This categorization helped us to understand if there exist common patterns of actions, discourses and impacts among the CSOs belonging to the same or different categories.

Unlike quantitative or questionnaire type interviewing where questions are structured in a specific order, in qualitative interviews there is more space and freedom for the researcher to conduct fully or semi or even unstructured interviews. The interviewer may prefer to remain loyal to the already prepared list of questions in a specific order. Others may prefer less structured interviewing where they can intervene in the language and order of questions. Throughout this semi-structured interviewing, new questions can be added while others may be dropped. What I prefer is to first prepare a list of questions and organize them as sets or groups of questions. Not specific questions per se, but these groupings largely guide my interviews in a way as to ensure that all research themes are covered. To illustrate, in my research on Europeanization and Kurdish civil society I classified questions as follows:

1) questions on the history, current structure, aims and activities of the CSO which the interviewee is affiliated with,

2) questions on the nature of the Kurdish issue,

3) questions on the peace process,

4) questions concerning potentials and limitations of the European and particularly the European Union’s involvement in the peace process,

5) questions about the interviewed CSO’s approach to European and other sources of funding.

I specified the questions in each of these groupings constructed in accordance with the research’s themes, aims and objectives. Even if there are some questions which are impossible to skip, it is much more important to ensure that each research theme or group of questions have been asked. Hence, rather than asking each and every question with the same wording, language and order it was much more important for me to gain enough data to address each theme of my research. It is always possible and is often needed to remove some questions while adding others throughout interviews. The researcher is also likely to discover new perspectives of the issue under research or may be given information by the interviewee entirely unknown to her. Subsequently, s/he may want to use or cross-check this information in the forthcoming interviews.

As discussed above, SSI enables a wide range of actors including those who have a limited access to the public to express their views freely and in a detailed manner. Yet, cross-checking the data provided by any interviewee with the data provided by others substantially increases the reliability and quality of data gathered through SSI. It may also be very useful to interview more than one person from the same institution, as in some cases personal views of the interviewees overshadow the institutional stance on a particular issue. Even if there is not an ultimate solution to this puzzle, cross-checking via multiple interviews from the same organization can remedy it to a certain degree.

 3.4. The post-interview stage

This stage comprises transcription, analysis and reporting of the interview material. Judging from my experience, I may suggest that only a very small part of what has been said by interviewees can be included in publications and even in the analysis. ‘To accomplish a text that gives a good account in the sense of mirroring a reality, represented in all this empirical material is very difficult, if not impossible’.[37] The researcher needs to make arduous decisions about which data should be used and which will be omitted. Researchers (from then on authors) often find it more conducive to include in the final production the statements by interviewees that support their arguments. Alternatively, authors may give priority to some data over others that they classify more important. These are very subjective decisions with various implications on the research outcomes.

For the analysis of transcriptions of interview materials, different ways of coding, i.e. concept or theme-driven coding and/or of content analysis may be used.[38] Whatever the method, the process of transcription is an integral part of the analysis since at this stage the researcher makes decisions concerning which material s/he would consider examining and integrating into the analysis. Transcribing the whole material is an option. Yet, if the researcher is not involved in a kind of critical discourse analysis, this could be time and energy wasting. An efficient way could be first, to clarify what s/he is particularly looking for in accordance with the aims and objectives of the research; second, to transcribe solely this material; and third, to mark overlaps and differences among the interviewees on particular issues of concern. While listening to hundreds of hours of recorded conversations, it is advisable to know what is more and what is less important or what is ignorable during the transcription process.

As Brinkmann points out, ‘there is no golden standard of transcription, everything depends on the purpose of one’s investigation and on what is possible in practice’.[39] In some cases, the researcher may notice that a part of conversation that is not transcribed may turn out to be an important text as it marks a point or points of convergence or divergence among interviewees. Or it might draw attention to an unexpectedly significant or unknown perspective of the research problem. Hence, it is prudent not to be so conservative in selecting the material to be transcribed. This is also related to the anticipated outcome of the research. If the outcome will be in the form of a short report or a journal article or a book will certainly affect the size of transcribed conversational material. As Brinkmann observes for qualitative research, writing does not solely mean to document what you have heard. It means also ‘to experiment with analysis, compare different perspectives on the empirical material and try out a number of alternative ways of presenting readings of the material. Writing should therefore be treated as an intrinsic part of the methodology of interview research and not as a final post-script added on at the end’.[40]

As mentioned before, SSI is very useful to explore how particular issues, concepts, political developments or conflicts are experienced, perceived and discursively framed by actors in the public or private realms. In such a case, an efficient way of analysing huge conversational data coming out of interviews could be to focus on convergences and divergences among interviewees in seminal issues or points for the research, instead of comprising all their utterances. These crucially important issues or points may be considered as separate paragraphs or even distinct subheadings. For each point of reference, similar and diverging argumentative threads could be specified and classified. Each group of arguments need to be represented in final texts (e.g. reports, journal articles or book chapters) by making references to their advocates. Yet, the researcher does not have to give the name of the interviewee or his/her institution directly in the text. Each interviewee can be enumerated and the in-text references could be given to these numbers. As such, it will be possible to reveal which group of arguments is articulated by whom or by which group of actors. This method also demonstrates the changing popularity of different arguments among the actors interviewed. To illustrate, some sentences are quoted from my article on de-Europeanization of Kurdish civil society:

The bulk of the interviewed CSOs converge on the notion that EU-required legal and constitutional reforms have widened the political space in favour of the civil society in Turkey (Interviewees 7, 15, 19 and 22)’. ‘Mainly two views stand out: Kurdish question as a ‘fundamentally human rights question’ (Interviewees 24, 4, 13, 5, 7, 29) and as a ‘national/political issue’ (Interviewees 17, 18, 23)’.

This method, however, will not prevent the researcher from including distinctive views by a specific interviewee. The following is a pertinent example from the same research as above: ‘The interviewee from Keskesor LGBT opines that the settlement process has already served to open up more space for civil society (Interviewee 15)’.

For the readers’ convenience, the list of interviewees should be provided in the appendix right after the bibliography. This list should comprise the position of the interviewee in a stated institution as well as the date and the place of the interview.

Interviewee 4: Director, Human Rights Association (IHD), Ankara, 14 February 2014.

Interviewee 12: Coordinator, Southeastern Industrialists and Businessmen Association (GUNSIAD), Diyarbakir, 19 February 2014.

If the required consent has been taken and if the interviewee is talking on behalf of herself rather than representing any institution, her name can be revealed:

Interviewee 3: Büşra Ersanlı, academic and activist, Istanbul, 13 February 2014.

4. Conclusion

In this paper, benefiting from my experience, I examined qualitative research interviewing with a particular emphasis on the potentials and limitations of semi-structured interviews. In writing this article an important motivation has been to come up with a guide for students of International Relations and Political Science who would like to use this method in their qualitative research. To further achieve this aim, it would be very useful to mention some works utilising qualitative interview as a method of inquiry in Turkish IR scholarship. It is fair to suggest that qualitative interviewing has not been among the most frequently used methods of inquiry in the scholarship on Turkish IR. Nevertheless, there are a few, notable examples that may inspire any researcher who would like to utilise this method in their research. While some of these scholarly works draw on interviews as the principle source,[41] some others benefit from interviewing as complementary to other methods of data gathering.[42] It may be suggested that there is a welcome upsurge of interest in interviewing among scholars of Turkish politics in recent years, yet the scholarly works employing this method in a systematic manner and detailing their ways of choosing interviewees, framing questions and transcribing and analysing the interview data are still rare and direly needed.

Following Alvesson, this article recommends a reflectivist approach to qualitative interviewing, problematizing and challenging the dominant positivist or neo-positivist assumptions, but without entirely negating them. This approach also disagrees with the deep mistrust and scepticism of some post-structuralists concerning interviewing as a credible method for qualitative research. The reflectivist approach is aware of the fact that the interview situation is political in nature and is not immune from power relations and interest calculations.

As post-positivists suggest, the researcher is far from being a neutral transmitter of what is said by the interviewee. The researcher’s role and impact go well beyond asking questions and reporting answers in an objective manner. Thus, the reflectivist approach requires the researcher to be self-critical throughout all stages of the research in particular concerning his/her role and influence during and after interviews. It is often overlooked that it is in the post-interview phase where the researcher makes arduous decisions concerning the data included and excluded in the analysis that the political and subjective role of the researcher becomes most visible.

This article examined the qualitative interviewing process over three stages: pre-interview or preparatory; fieldwork or interview; and the post-interview or analysing and reporting stage. However, rather than being a necessarily linear process of research, SSI needs to be considered as constituted by overlapping and cycling stages. SSI is flexible compared to more structured or survey type interviewing in the sense that it allows more space to interviewees to express themselves freely and in a detailed manner. It also gives more opportunity to researchers to participate in the interviewing process as active subjects that may guide the interview towards new directions when they think that this serves best to their research interests. SSI is particularly useful for identifying common and divergent argumentative threads of societal and political actors involved in large public debates. It is a complex methodological tool which can yield reliable and comparable insights when used in conjunction with other qualitative methods, i.e. discourse analysis and content analysis. Even if it is argued that SSI may be employed both at the early and later stages of the research, it best serves those researchers who have already developed a keen and historical understanding of the relevant topic of interest.

To materialize its full potential, SSI needs to be preceded by detailed research on the topic of interest, on the actions, discourses and identities of the relevant actors as well as on the web of relations among them. SSI enables a wide range of actors including those who have a limited access to the public to express their views in a free and elaborate manner. Yet, cross-checking the data provided by any interviewee with the data provided by others would substantially increase the reliability and quality of data gathered through interviews. SSI develops as a social interaction between the interviewer and the interviewee rather than as a mere case of data gathering between the subject and the object. As an intersubjective process, qualitative interviewing is deeply shaped by personal characteristics of both the researcher and the participant, i.e. internalised norms, cultural scripts, biases or worldviews. The ways in which the interviewer formulates the content, language and even the order of questions may affect the language and the attitudes of the interviewee. Likewise, the interviewee may influence the researcher so that s/he can drop some questions, while adding some others. It is this dialogical potential of the SSI that renders it a unique and invaluable method of knowledge production in qualitative research.

[1] Timothy John Rapley, “The Art(fulness) of Open–ended Interviewing: Some Considerations on Analysing Interviews,” Qualitative Research 1 (2001), cited in Svend Brinkmann, Qualitative Interviewing (Oxford: Oxford University Press, 2013), 1.

[2] Mats Alvesson, “Beyond Neopositivists, Romantics, and Localists: A Reflexive Approach to Interviews in Organizational Research,” Academy of Management Review 28, no. 1 (2003): 13.

[3] Alvesson, “Beyond Neopositivists,” 13.

[4] Alvesson, “Beyond Neopositivists”.

[5] Brinkmann, Qualitative Interviewing, 51.

[6] For different methods of coding in qualitative research, see Johnny Saldana, The Coding Manual for Qualitative Researchers, 3^rd ed. (London: Sage, 2016).

[7] Nathalie Tocci and Alper Kaliber, “Conflict Society and the Transformation of Turkey’s Kurdish Question” (SHUR Working Paper Series, 2008), accessed August 7, 2017, http://www.luiss.it/shur/wp–content/uploads/2008/06/shurwp01–08.pdf; Alper Kaliber and Nathalie Tocci, “Conflict Society and the Transformation of Turkey’s Kurdish Question,” Security Dialogue 41, no. 2 (2010): 191–215.

[8] Alper Kaliber, “De-Europeanisation of Civil Society and Public Debates in Turkey: The Kurdish Question Revisited,” in special issue (co–editor with Senem Aydın–Düzgit) “Is Turkey De–Europeanising? Encounters with Europe in a Candidate Country,” South European Society and Politics 21, no. 1 (2016): 59–74.

[9] Kaliber, “De-Europeanisation”.

[10] Annie Irvine, Paul Drew, and Roy Sainsbury, “‘Am I Not Answering your Questions Properly?’ Clarification, Adequacy and Responsiveness in Semi-Structured Telephone and Face-to-Face Interviews,” Qualitative Research 13, no. 1 (2012): 316.

[11] Brinkmann, Qualitative Interviewing, 42.

[12] Mats Alvesson, “Methodology for Close Up Studies– Struggling with Closeness and Closure,” Higher Education 46 (2003):168.

[13] Alvesson, “Beyond Neopositivists,” 26.

[14] Carolyn Baker, “Membership Categorizations and Interview Accounts,” in Qualitative Research, ed. David Silverman (London: Sage, 1997), 30, cited in Alvesson, “Beyond Neopositivists”, 19.

[15] Alvesson, “Methodology,” 169.

[16] Alvesson, “Beyond Neopositivists,” 22.

[17] Alvesson, “Beyond Neopositivists,” 22.

[18] Alvesson, “Beyond Neopositivists”.

[19] Alvesson, “Beyond Neopositivists,” 14.

[20] Alvesson, “Methodology,” 170.

[21] Thomas Diefenbach, “Are Case Studies More Than Sophisticated Storytelling?: Methodological Problems of Qualitative Empirical Research Mainly Based on Semi-structured Interviews,” Quality & Quantity 43 (2009): 881.

[22] Brinkmann, Qualitative Interviewing, 18.

[23] Brinkmann, Qualitative Interviewing, 21.

[24] Brinkmann, Qualitative Interviewing, 18.

[25] Brinkmann, Qualitative Interviewing, 21.

[26] Brinkmann, Qualitative Interviewing, 46.

[27] Diefenbach, “Case Studies,” 880.

[28] John D. Caputo, Radical Hermeneutics: Repetition, Deconstruction, and the Hermeneutic Project (Bloomington and Indianapolis: Indiana University Press, 1987), 60 cited in Diefenbach, “Case Studies,” 880.

[29] Kaliber and Tocci, “Conflict Society”.

[30] Kaliber and Tocci, “Conflict Society”.

[31] Diefenbach, “Case Studies,” 880.

[32] Diefenbach, “Case Studies” 880.

[33] Barry Buzan et al. Security: A New Framework for Analysis (London: Lynne Rienner, 1998).

[34] Brinkmann, Qualitative Interviewing, 49.

[35] Brinkmann, Qualitative Interviewing, 49.

[36] Kaliber and Tocci, “Conflict Society,” 192.

[37] Alvesson, “Methodology,” 173.

[38] Brinkmann, Qualitative Interviewing, 62.

[39] Brinkmann, Qualitative Interviewing, 61.

[40] Brinkmann, Qualitative Interviewing, 67.

[41] As an example to the former, Lauren M. McLaren and Meltem Müftüler Baç, “Turkish Parliamentarians' Perspectives on Turkey's Relations with the European Union,” Turkish Studies 4, no.1 (2003): 195–218 analyse the views of 61 Turkish parliamentarians on Turkey–EU Relations and includes a detailed methodological section particularly on how to choose sampling. Similarly, Özge Zihnioğlu, European Union Civil Society Policy and Turkey: A Bridge Too Far? (UK: Palgrave Macmillan, 2013); Ahmet Erdi Öztürk and Semiha Sözeri, “Diyanet As A Foreign Policy Tool: Evidence from the Netherlands and Bulgaria,” Politics and Religion, accessed March 8, 2018, doi:  10.1017/S175504831700075X; Büke Boşnak, “Europeanisation and De-Europeanisation Dynamics in Turkey: The Case of Environmental Organisations,” South European Society and Politics 21, no. 1 (2016): 75–90; Alexander Bürgin, “Why the EU Still Matters in Turkish Domestic Politics: Insights from Recent Reforms in Migration Policy,” South European Society and Politics 21, no.1 (2016): 105–18; Zeki  Sarıgil and Ömer Fazlıoğlu, “Religion and Ethno–nationalism: Turkey's Kurdish Issue,” Nations and Nationalism 19, no. 3 (2003): 551–71; Bahar Rumelili and Didem Çakmaklı, “Civic Participation and Citizenship in Turkey: A Comparative Study of Five Cities,” South European Society and Politics 22, no. 3 (2017) draw mostly on interviews conducted with political/bureaucratic elites as well as NGO representatives, community leaders  and intellectuals. Senem Aydın Düzgit, “Legitimizing Europe in Contested Settings: Europe as a Normative Power in Turkey?,” Journal of Common Market Studies, accessed March 8, 2018, doi: 10.11/jcms.12647 explores if the European Union is still perceived as a ‘normative power’ by different segments of Turkish society and draws heavily on focus group interviewing, a method rarely used by the students of Turkish politics.

[42] The works utilising interviewing as a complimentary method to other methods of inquiry are also observed in the scholarly reflection on Turkish foreign and domestic policy. Recent examples to this category of studies include Ziya Öniş and Şuhnaz Yılmaz, “Turkey and Russia in a Shifting Global Order: Cooperation, Conflict and Asymmetric Interdependence in a Turbulent Region,” Third World Quarterly 31, no. 1 (2016): 71–95; Zeki Sarıgil, Ethnic Boundaries in Turkish Politics: The Secular Kurdish Movement and Islam (New York: New York University Press, 2018); Özge Onursal-Beşgül, “Policy Transfer and Discursive De–Europeanisation: Higher Education from Bologna to Turkey,” South European Society and Politics 21, no. 1 (2016): 91–103; Ioannis N. Grigoriadis, Trials of Europeanization: Turkish Political Culture and the European Union (UK: Palgrave Macmillan, 2008); Bahar Rumelili and E. Fuat Keyman, “Enacting Multi-Layered Citizenship:  Turkey’s Armenians’ Struggle for Justice and Equality,” Citizenship Studies 20, no. 1 (2016): 67–83; E. Fuat Keyman, “Turkish Foreign Policy in the post–Arab Spring Era: From Proactive to Buffer State,” Third World Quarterly 37, no. 12 (2016): 2274–87.