Substantive Dimensions of the Deliberations
To participate, you may either post a contribution to an existing discussion by selecting the thread for that topic (and then click on "Post Reply") or start a new thread by clicking on "New Topic" below.
The transition to Stage 2 of the deliberations is currently underway but will take some time to complete. In the meantime, we very much welcome additional contributions to the existing threads in this forum.
For instructions on how to follow a discussion thread by email, click here.
University College London
- Posts: 7
- Joined: Sat Apr 23, 2016 4:48 pm
Harms in study of identity, and forgotten agency
The Association for the Study of Nationalities (ASN) executive passed a resolution expressing concerns about DA-RT at its April conference. The text of the resolution is here: http://nationalities.org/news/asn-issue ... ing-da-rt/
I summarized some of the concerns in my paper for the conference, selections of which are below:
This discussion argues that research on nationalism encounters many similar problems to those in violent environments and those in authoritarian or unstable political contexts. I briefly sketch some of the primary harms that DA-RT presents for research sources, for researchers, and for the researcher-source relationship. These problems cross research areas due to the fact that it is conflicts of power that underlie all of these research areas. These inherent power problems indicate that the scope of research amenable to DA-RT does not rest on the “quantitative” and “qualitative” identity groupings for research that have emerged in political science. Instead, these are logical problems inherent in the research of power-laden environments that take human agency seriously, and which exist even beyond the fields discussed in this paper. I develop these arguments further below, and suggest some potential fixes for these problems.
I. Research that Causes Harm
What do we owe to those whom we research? And why do we take up their time and invade their lives to research them? A large part of those involved in research on nationalism or the politics of identity are committed to goals that are external to promoting their own academic careers. Many of us hope to improve the state of the world, whether naïve or not. An effort to understand and explain political struggles relating to ethnicity is often related to a sense that if we could only understand the nature of such struggles, we could help ameliorate understanding between those in said groups. In addition, there are some who desire to advance understanding of groups that have been historically marginalized with a sense of improving the lot of such groups, as in the case of much research on the Roma in Europe. Researchers driven by these aims will usually have a basic goal of respectful treatment of their sources as well as an aim to protect them from harm.
There is, however, another view about why we conduct research. Much of the language on the need for transparency comes from the base assumption that researchers are focused on advancing their own careers, and that they thus face great incentives to mis-represent results in ways that will favourably aid their careers. And indeed, there have been some cases of such instances. Such cases are held up by DA-RT advocates as demonstrating a need for greater research transparency, along with the intent to provide an ability to replicate study results to confirm their validity. While an effort to address the potential problem of careerist deception is laudable, the intended DA-RT “fix” will create problems far out of proportion to the potential harm of deception. As noted by Pachirat (2015), “DA-RT risks becoming a solution that generates problems that did not exist before” (30).
There are three clear types of harms to persons that can result from the implementation of the DA-RT requirements. The first type involves harm to individual sources who provide interviews or materials to researchers, and the second the potential harm to researchers in terms of their continued ability to conduct research. There is also a third area of harm that relates to the damage that can be done to the relationship between researchers and sources. This relationship damage constitutes a personal harm as well as a violation of what we think we are doing in research. This section emphasizes the personal aspects in the three harms, and the next section addresses the broader questions for social science research.
Following the bureaucratic language of university ethics boards, the DA-RT requirements describe those whom we research as “human subjects.” This phrase assumes a type of white-coated research style imported from the hard sciences. Such language has been frequently adopted by those conducting field experiment work, in which a population is divided into groups and one receives an experimental “treatment” in a study modelled on the logic of drug trials. Was Subject 1 exposed to this leaflet of information? Did it change how he or she responded to a survey question? Experimental studies tend to constrain individual responses in the context of the research project, in a way that can make them look more like data points than like human beings. A roundtable on “Fieldwork” at the 2015 APSA meeting represented this view widely, with just one panel participant who had experience with sustained engagement and intensive interviews with human beings in the field. The field of experimental research is not a dominant form of research in the study of nationalism and identity, as much of what we study relies on understanding actual dynamics on the ground. Conducting fieldwork for any length of time means that a researcher ends up eating meals with and drinking with sources, living with them, attending weddings and funerals, and well, thinking of them as human beings rather than “human subjects.” Even those doing archival research who hope to avoid the morass of ethnography will be surrounded by human beings during the course of their work in the field. Does this mean that our research is “contaminated” by this exposure, as a fundamentalist interpretation of the scientific method might imply? Of course not. It means that we are social scientists in a social world rather than evaluating chemical responses in a lab.
Unlike the research of chemicals in a lab, the research of human beings can harm them in ways that have moral implications. This first potential harm implicit in the DA-RT requirements is the most obvious. A best practice in our field is to render sources anonymous (Hall 2015). There are common sense reasons for this that made people do so even before anonymity was formalized by university ethics boards. When I discuss controversial political ideas with someone, whether around a dinner table or in a formal interview, the things that person tells me can harm them if they become fully “transparent.” Ethnic majorities disagree with ethnic minorities, and vice versa. However, they often work in diverse environments and are friends with individuals outside of their own group. Their views on ethnic politics, if revealed, could harm those interactions and perhaps even their careers.
Taking potential harms seriously means that formally marking down someone’s identity as linked to controversial comments is dangerous (Fujii 2016, Lynch 2016). Let’s imagine that a PhD student following the DA-RT requirements on supervisor advice wishes to name sources and dutifully collects several interview consent forms during interviews on tense ethnic politics. Let’s imagine that said student has also dutifully collected voice recordings for the interviews to ensure a full set of research materials to be ready upon request. Let’s assume that said student is then stopped at the border and her suitcase is searched. In the process of the search, police take the consent forms and the recordings, placing her sources in clear danger. It is here that a Catch-22 with university ethics boards becomes clear: consent forms themselves can endanger sources. For this reason, anonymity is often the best way to protect sources within tense political contexts.
The second harm to researchers can follow from the above example. Let’s imagine instead that said student decided to preserve full anonymity for sources, conducting research according to the dangers of the research context rather than according to the DA-RT requirements. This would involve careful note-taking, and no recording devices (persons can be voice-identified, Hall 2016: 29) or consent forms. She then produces extensive handwritten notes and transcripts of interviews, with code names for sources. In essence, a good rule to follow for protecting sources is this: could they figure out who these people are without me? If not, then proceed. Let’s imagine that the student plans to publish the material in the form of articles. What happens next will depend greatly on the journal to which she submits. If the journal editor has signed DA-RT and have decided to proceed with implementation, the editor has the discretion to request all interview transcripts, code names, handwritten notes, and any other materials.
In addition, if those notes or transcripts are not in English, some fear that editors might request that these also be translated (Lynch 2016). The burden that DA-RT places on researchers using large databases is relatively straightforward, in terms that it is the database itself that is requested by researchers. For our student and her interview notes, however, the parameters of what editors can request are less clear, and she could receive a desk reject with a note that the submission can only be taken for review with the provision of a list of materials that is simply unworkable for someone with limited time or no budget for translation or assistance. The disadvantage placed on researchers conducting interviews and fieldwork is disproportionate to the perceived benefit of transparency (Parkinson and Wood 2015). Moreover, the stated requirements that such materials are to be placed in an online repository that can be publically accessed returns to the potential harm to sources. Interview transcripts might identify (or-misidentify) individuals. Faced with these problems, our student withdraws her piece and wonders what her next steps should be. As several have questioned with regard to the DA-RT requirements, what precisely is to be gained from this type of heavy-handed reporting on the research we do (Hall 2015, Lynch 2016)? Is the perceived gain of transparency proportionate to the burden of the requirements on researchers? And are journal editors the appropriate gatekeepers on such matters (Yashar 2016)?
Finally, there is a third harm – that to the relationship between the researcher and her sources. Let’s imagine that after immense professional pressure, our student has decided to proceed with the posting of her interview notes online. On the positive side, she receives an article acceptance from Mega Journal that will rapidly advance her career prospects. On the negative side, the internet can be reached by everyone. Our researcher receives an email from one of her former sources, stating that he is fearful of his own job now that he has seen that the interview notes are available online. While anonymous, the notes contain details that the source considers to reveal his identity. The note states that he wishes that he had never met our researcher, and will never speak to her again.
What response should we have to that outcome? If we view human sources as mere “data points,” then the response might be an “oh, well,” about the sacrifices needed to be an Important Political Scientist. As outlined by Parkinson and Wood, the DA-RT requirements set up a conflict of interest between researchers and sources, pitting “the researcher’s incentives and norms at odds with those of her interlocutors” (2015, 26). Much of the information a researcher can gain in sensitive political environments follows the establishment of a relationship of some trust between the researcher and her sources. The conversations between researcher and source constitute privileged information rather than discussions intended to end up on the internet for general consumption (Cramer 2015). While some degree of technology could be applied to limiting the audience for those transcripts, the principle of the problem remains: the release of the material for other eyes is a violation of the trust built into the relationship that produced the conversation in the first place.
Over time, this violation of trust creates a practical problem, with the result of “self-censorship” on the part of researchers and the researched (Parkinson and Wood 2015: 24). If individuals fear that there is any danger that their identities will not be protected through confidentiality, this policy can make it even more difficult for researchers to find sources for researching sensitive topics in sensitive environments. The notion of how sources might think about researchers is often quite under-estimated in terms of its implications for obtaining information. In my first fieldwork experience on the Shoshone-Bannock Reservation in 1990, the first question I was asked by those I wished to interview was often this: “Are you an anthro?” Weary of being prodded by anthropologists, reservation residents had decided to try to discourage further anthros from writing about them yet again. The notion that we interact with individuals who have agency in the course of our research, rather than with data points that we can manipulate and use to our will, is the notion of reflexivity in social science (Bourdieu and Wacquant 1992). Among DA-RT proponents, questions of reflexivity have been largely dismissed as problems that emerge only in ethnographic or constructivist research. However, the reality is that this interactive relationship exists even for those who think that they are working with “human subjects” rather than human beings.
II. Power in Politics, and Implications for Research
One striking aspect of the DA-RT discussions that produced the policy is that the absence of those with experience of conducting research on the developing world (Lynch 2016: 36). Perhaps for this reason, the ongoing practices of researchers working in fraught fieldwork environments and authoritarian regimes were largely ignored in the process of DA-RT construction (Lynch 2016: 37). Those conducting research in the context of authoritarian regime environments such as China must take particular care with regard to potential harms to sources and to themselves, as these can involve physical risks as well as the harms mentioned above. A set of pre-DA-RT adaptive techniques have emerged that allow researchers to continue to conduct work (Shih 2015). As Shih outlines, collaboration with academics in China can be dangerous to those academics, and thus should in fact be avoided for topics considered particularly sensitive by the regime (Shih 2015: 21). The wording of such topics is also kept implicit throughout research projects, both in the conduct of research such as surveys and in the presentation of the final research product. Such considerations also require that fieldnotes should not be shared, and laptops have been known to have spyware installed by state authorities such that password-protection for notes is not sufficient (Shih 2015: 21-22). In short, many of the requirements set by DA-RT for transparency go against standard practice for research on authoritarian regimes. The notion that a journal editor might grant an “exemption” to DA-RT requirements for such research requires that editors should be aware of all of these potential considerations in terms of each potential research setting, as gatekeepers for what should be granted an exemption (Yashar 2016).
One might argue that China is an obvious case – but how many general political science journal editors keep up to date with current regime politics in Egypt or Kazakhstan, which could involve very similar requirements from a regime standpoint? Can we expect that all university research ethics boards might be continuously following such developments, if they are used as a proxy for exemptions? Moreover, what if a place that is “safe” today becomes less safe for sources and researchers tomorrow, as in the case of the 2011 uprisings in the Arab Spring (Lynch 2016)? The study of conflict also presents problems for DA-RT requirements. A declaration that a conflict has ended does not mean that sources are suddenly “safe” (Parkinson and Wood 2015). In identity conflicts in particular, grievances will remain long after a conflict has been considered over by outsiders.
One example of these problems is the tricky case of the Boston College archive on Northern Ireland. Sources were interviewed by historians in an effort to understand the Northern Irish conflict. In that case, sources had been promised that files would only be opened upon their death (McDonald 2015). Years later, police managed to obtain these files and arrests were made as a result of historical research and the opening of an archive. The Boston College case should set aside any doubts that one might have regarding the interactive nature of research and the researched – that is, reflexivity. A legal response to these events might be that those who shoot others should be prosecuted. But as scholars of conflict, we cannot make those determinations and maintain the neutrality necessary to write about identity-based disputes. We cannot both foster understanding and become a judge at the same time, because we will become entrapped in the dynamics of the conflict that we are researching. A lack of understanding of these problems comes through strongly in the DA-RT requirements. The operative question about transparency in fraught conflict environments is always: transparency for whom? Transparency for whose advantage?
The study of nationalism and identity politics is thus fraught with many of the same problems that emerge in authoritarian and conflict settings, because all of these contexts are about a struggle to obtain or maintain power. Even if an ethnic or religious conflict is not in full-blown violence, the fact that groups have longstanding disagreements means that research in such settings is not as straightforward as it might be on, say, a population’s view of a particular presidential candidate in the United States. In a high-stakes identity conflict, sources tend to be very aware of the potential for the researcher’s work to be used to their group advantage (or disadvantage). When conducting fieldwork in mixed ethnic settings in Eastern Europe for my 2012 book, individuals repeatedly asked me to tell the story of how their ethnic group had been wronged (Stroschein 2012). Individuals were quite happy to provide me with information to advance that goal, which meant that I obtained a wealth of information from each side. It was my job to carefully sift through the interviews, newspapers, and ethnography to produce a story that could represent the political dynamics produced by these colliding views. The resulting book was a product of my agency, but also the agency of my sources as human beings with their own goals of getting their story out there.
Since I became aware of the DA-RT policies (by accident) at the conference of the American Political Science Association in September 2015, I have questioned the potential fate of the book’s material with a DA-RT compliant journal. Certainly, all of the considerations of harm mentioned above would be a serious issue, one that would have crippled my ability to do the research. But there is another disturbing aspect. At each step in the process of engaging with the research, I aimed to be as neutral as possible, as neutrality was crucial to understand the conflict issues between groups. In the course of the project I had to accept that neutrality meant that there were two (or more) versions of history that were accepted in the region. It also meant that there were two (or more) versions of what was right or just in terms of policies, as many within each group were convinced of the just nature of their cause and the lack of justice in the claims of the other group. Within this fraught environment, the result of my research would enter the discourse with potential implications for the balance of power between groups. I would be naïve to think otherwise, and most attendees of the ASN conference are acutely aware of these potential implications of their research. Those we research are human agents, who will interpret and use what we produce within the context of their own political struggles.
Once we recognize that sources or “human subjects” are human beings with agency who are acting within a fraught environment, this realization has some important implications for the DA-RT version of transparency. First, as a journal editor with a solid knowledge of my area remit, I remain also aware that I am not able to track the political orientations of each author who submits and each reviewer I select. Let’s imagine that a piece is submitted on the politics of the Turkish minority in Bulgaria. It is entirely plausible that I could select a reviewer that I consider a specialist on Bulgarian politics, but who, unbeknownst to me, has written strong articles in the Bulgarian press against the actions of the Turkish minority. Under the DA-RT requirements, the reviewer could ask to see all of the interview transcripts by the researcher, and as editor, not having reason to question these motives, I might request that the author submit them. The author, not knowing the identity of the reviewers, might comply. As journal editor, I have then unwittingly engaged in a political project to the advantage of the reviewer’s position, by providing that person with these sensitive interview transcripts and with no control over what that person might do with them next. It is naïve to think that there are protections in place to protect against such happenings. I do my best in my role as journal editor, but I cannot know the intricacies of every conflict dyad and the identity of those taking positions. Nor is it reasonable to assume that I can find reviewers without any prior positions. The best guarantee against getting embroiled in the political dynamic is to maintain current arrangements on journal procedures. This is not because of a sense of being old-fashioned, as DA-RT advocates are trying to spin their opponents (Elman and Lupia 2016: 50), but because implementing DA-RT would remove the journal from a neutral position. Of course, under current procedures a reviewer might simply reject a piece due to their political views. But due to the need for review comments, those instances are far more readily recognizable than the potential harms of sensitive interview material being released to an unfriendly party.
Second, in a fraught context in which human beings have agency, it is important that I as a researcher maintain some agency over sensitive research materials. At several talks I have given on my book, I have had to carefully navigate an audience member who strongly identified with one ethnic group in the region, and who wanted to know why that group’s story was not emphasized enough in terms of the justness of its cause. With all of my interview notes and materials on the internet, I can imagine some examples of the virtual equivalent:
• Why did I talk to a member of X organization?
• How could I not by convinced by the conspiracy theory outlined so eloquently by Y?
• Who is source Z? We find them in violation of our position on this issue and would like to question them further (from state actors).
• We are disturbed that you interviewed members of that organization. We will be terminating your visa for future travel (from state actors).
It is pure hubris to assume that only other researchers would be interested in these transcripts or materials. Making them available without discretion means that actors with strong stakes in the conflict would be more able to use our work for their own political ends. Of course, my book as the finished product is available for use in these ongoing power struggles. But this is material that I was able to carefully edit with these considerations in mind. Raw materials are not edited in this way.
It is worth mention that the book materials include databases that I used to code information into a simpler graphic form. I am more willing to distribute these materials, as they were compiled from newspapers and thus constitute public information. This willingness is not due to the fact that they contain numbers, but rather due to the public nature of the source. It is conceivable that one might code more sensitive information into numbers, and this information might be subject to the same concerns mentioned above. The reification of “quantitative” and “qualitative” identities around DA-RT is less useful than it might seem at first glance. In research on nationalism and identity, everything is political, including numbers – and thus information must be treated with care and some awareness of how it might be used within those political dynamics.
A better way to distinguish between those concerned about DA-RT and fieldwork and those dismissive of those concerns is the researcher’s understanding of human agency. “Human subjects” are not conceived of as having agency; they are acted upon. They are examined and assessed as data points that are recorded in a project like test results in a lab. In this view, it is easy to prioritize transparency, because it is difficult to envision any of the harms I have outlined here. In contrast, if one sees human beings instead of subjects in the process of conducting research, the potential harms are more immediately visible. The Kantian imperative, or the notion that one should only treat those researched as one would like to be treated, is a good general rule to follow.
These different perspectives on the agency of those we study are linked to philosophical differences. It is not methods per se that produces these differences, but philosophy of science does. An attempt to treat human beings as objects of research that exhibit the regular behaviors of natural phenomena will inevitably bump into the problem of agency (Almond and Genco 1989). Human beings in fact have agency to change their minds, read, and to have their own goals in a research encounter. It is due to agency that the social sciences exhibit inherent differences from the physical sciences. The effort to study human beings using approaches developed for non-agentic phenomena has the potential to produce useful insights. However, if researchers begin to actually treat human beings in terms of their simplifying assumptions, they will miss some of the implications of the research project itself. It appears that some of the DA-RT enthusiasm stems from confusing such assumptions with reality.
- Posts: 11
- Joined: Thu Apr 07, 2016 8:20 pm
Re: Harms in study of identity, and forgotten agency