Fun fact: no one in grad school teaches you how to write a qualitative research paper. If you’re like me, you didn’t read a ton of qualitative research during coursework, and much of what you did read was quick case illustrations tacked onto a statistical analysis. Or you read books, which are excellent for a whole host of other reasons but not really for how to condense interview or field data into 20 pages.
I recently went through the publication process for a qualitative, interpretivist-y article (my first!) at a top IR journal, and I want to share what I learned about the decisions involved in presenting such research. So I’ve written up some notes on the piece (you can read it here if terrorism and anti-racist institutional change are your things) that go section by section and talk about the choices I made in writing after the research was already done. My choices are not the only choices, and some of them may not even be “good”—that’s for readers to judge. But hopefully it’s helpful to see not only one way of doing things, but also some commentary on the logic behind doing things that way. I am indebted here to two anonymous reviewers at ISQ for giving the kind, constructive feedback that everyone deserves but is too rarely offered. I also benefited tremendously from reading qualitative work by Erica Simmons, Rachel Schwartz, Benjamin Meiches, and Lisa Stampnitzky. Though I may have borrowed techniques from them, all errors are my own.
Introduction
In social science, we talk about “framing” as a critical component of an article, but we don’t often talk about how to do it. I think about framing as answering two questions: why should anyone continue reading this article after the first page, and whose voices will be centered in this piece?
As a reader, I don’t like being dropped into the literature at the top of an article. This is a surefire way to make me switch from close reading to cavalier skimming. In my mind, research is rooted in the real world, so we should always take care to respect those roots. So I knew before I started writing that I wanted to open with Walter Lübcke’s assassination. This was also the incident many of my interlocutors used to frame their thoughts on counterterrorism in Germany (as it happened shortly before I started fieldwork), and it felt important to replicate a bit of the environment in which I did my research.
The other major choice I made up top was how to describe what I ended up calling “white supremacist violence” and “violent white supremacy.” There were a number of possible alternatives—far-right extremism, white supremacist extremism, white supremacist terrorism, etc. As a critical scholar, it was important for me not to replicate the practices of reflexively calling certain things “terrorism” or “extremism” and others not that I call out in the article. At the same time, I felt compelled to confront head-on the avoidance I often encounter in academic work of whiteness under the guise of being “apolitical” (as if that were a thing). So I chose a redundant term, violent white supremacy, to call attention both to the ideology underlying violence and the inherent violence of that ideology. Any discomfort I or white readers felt about confronting the mundanity and pervasiveness of white supremacy was much less important than calling a spade a spade.
I thought of the rest of the intro as convincing non-terrorism, non-race, and non-Germany scholars why they should read the piece. I tried to do this in as many ways as possible: discussing institutional change and stability for the institutions folx, frequently referencing international narratives for the IR folx, and talking about global patterns of violence for the policy crowd. I ended on definitions, to which I wish more people would devote time in their introductions. The language we use to describe the world cannot be an afterthought—or, worse, not a thought at all. Our terminological choices are in fact choices, and like all choices, we should tell the reader upfront when we’re making them.
How we understand government responses to terrorism (or what we might call “the lit review”)
This is my process: I write in one straight shot, in order, very quickly. I put together the first draft of this article for APSA 2019, and I wrote each section in the order it appears over the course of about three days. I work like this because training as a journalist ruined me forever for healthy approaches to deadlines, but it also gave me skills to synthesize a lot of material and find the narrative under pressure, and for that I am perpetually grateful. (tl;dr: Don’t try to replicate this method unless you already know that it works for you. You will burn out and die.)
One side effect is that caveats, anticipated questions, and “oh crap I need to mention this thing” digressions appear in my writing as they occur to my manic, caffeine-based brain. You can see one such instance at the beginning of this section, where I engage with a common critique of my work: “but why won’t you define terrorism, Anna.” At the time, I figured I would have to move this bit elsewhere in a later draft, since it’s sort of definitional and rather foundational and thus might have a home in the introduction. But I like it here, because it does what we’re told literature reviews should do—situate our work within often opposed theoretical and methodological traditions—but that they rarely do in practice. And the introduction was already too long, so there’s that.
I approached the rest of what we might call the “literature review” section as an opportunity to address what I see as a lack of existing literature on my topic and an interrogation of why that might be. Initially I failed to cite most of what I cite here, falling into the common trap in scholarly work of knowing too much about my topic and assuming everyone else a) knows an equal amount and b) agrees with my characterization of that topic area. I’m grateful to Joe Young for making me go back and explain in more detail why I think existing literature is helpful yet insufficient for the outcome I’m trying to explain, particularly when it discusses “far-right extremism” and other euphemisms I eschew. At one point, there were even more citations here; I dropped a few when I was trying to make the article shorter, because, fun fact, removing reference list entries is the fastest way to cut words (especially when those entries are chapters in edited volumes—those things are like 50 words long, each). I understand more now the reasons junior scholars, women, and POC have to fight so hard to get cited, and I wish journals would stop including reference lists in their word limits as they work to combat citation biases.
Incorporating hegemonic national identity (or what we might call “the theory”)
Much of the content in this section and the previous one has been the same since about October 2019. What’s changed, however, is the ordering. Paragraphs have moved between and within these sections with only slight alterations to transition sentences so that things still make sense. The lesson, to me, is that there are numerous ways to tell the same story, so if you can’t quite thread the needle with yours, you may already have all of the pieces you need; just try them in a different order. (I do this in multiple Word documents, which I keep track of using nested directories and files named only with dates. This is profoundly inefficient and bewildering to others, but it works for me.)
I also want to draw attention to what I think is a widespread but little-discussed fact in social science research: oftentimes we come across an article that just happens to “click,” and so it becomes a central part of our argument no matter how random it seems. Brudny & Finkel (2011) did that for me. I don’t remember how I discovered it initially, but I do remember reading it at The Wormhole in Chicago and knowing immediately that this concept—hegemonic national identity—was the missing link my ideas needed in order to make sense. One might not think ex ante that reading an article about post-communist identity politics in Ukraine and Russia was a good use of my time, but it’s a testament to the importance of reading widely and branching out beyond our narrow subfields. There’s tons of useful stuff out there!
Last thing: given that this is the last section before we get to “methods,” one might expect it to end with formally stated hypotheses, in the classic formula for empirical political science research. May I state an unpopular opinion? I hate formal hypotheses! I hate forcing what are fundamentally assumptions about the world into standardized language that strips away the personal and identity-based biases that shape those assumptions in the first place. And I think our single-minded focus on scientism encourages us to boil real-world processes down to simplistic statements of dual-variable cause and effect. I don’t know that what I did instead—stating my assumptions in a lot more words—gets around these problems, necessarily. I need to study more how interpretivist scholars do this, since I was never taught an alternative to finagling my ideas into a formal hypothesis structure. But I hope this is a step toward recognizing the complexity inherent in any statement, whether or causal or constitutive or descriptive, about the real world.
Methods: Elite perspectives from Germany (or what we might call “methods and data”)
A faculty member who shall remain nameless but not genderless once said that quantitative and experimental research is held to a higher standard than qualitative research. I would like to see that faculty member try to write the methods section for a qualitative article (and also, y’know, spend enough time learning about a context other than his own to understand and discuss controversial topics in his non-native language; I digress). This shit is hard.
I’ve been told before that my methods sections read like research diaries, and I fail to understand why this is a bad thing. Pretending that our methodology unfolds in a linear fashion, and that we know what we’re doing every step of the way, is not only factually incorrect, but it does a disservice to junior scholars who encounter hiccups and think they’re doing something wrong. (Hi.) There were originally a lot more personal anecdotes here, which in retrospect I regret cutting for length purposes, because those are the bits of research that are real. (They were also, inevitably, one of the first things I got comments on, despite appearing halfway through the article. Why the aversion to the personal, political science? What gives?)
Anyway. I wrote this section initially as a stream of consciousness, with choices I’d made coming back to me and getting thrown in at random places. I advocate writing this way: it’s useful to get all of the bits and pieces out of your head and onto the page where you can see them and the narrative becomes clearer. I didn’t know what wasn’t essential to describe my methods until I wrote it all in prose. Writing this all out was also useful for considering choices I’d made quickly or subconsciously (whoops) and how they’d affected my work without me realizing it in the moment.
All of the components you’re told about are here: justification for case selection, description of the “data” (here, my interlocutors), how the “data” were selected and collected, and a paragraph on researcher positionality (which is less typical but so so important). Digression: a faculty member once told me he disliked anthropology because it was “too reflexive.” While there is certainly a scenario in which one’s research becomes a navel-gazing exercise, this statement betrays a popular reluctance for considering how our identities influence both the contexts we research and how we make sense of them—a tendency that perpetuates the racist, colonialist, heteropatriarchal structures that dominate political science.
And a quick note, because I’ve gotten questions on this: I use “interlocutor” rather than “interviewee” or “research subject” on purpose. Both “interviewee” and “research subject” imply that research is a one-way street: at best, the communities we study give us something, and at worst we take it. But research actually works through the co-production of knowledge and understanding—through conversations and mutual exchange. Or it should, I think. And so I honor that ideal here.
Event selection (or, WTF is inductive research)
The other hard part of writing a qualitative article is deciding, in all of the rich information you’ve collected, what details you’re actually going to write about. For quantitative scholars, imagine not only that every cell in a spreadsheet comes with a story (and every cell does! we just don’t think about it!), but that you also have the full text of that story as told to you by one of the protagonists. Do you summarize the story? How briefly? Are there key lines that encapsulate the takeaways better than a summary ever could? Which ones do so best? (You have limited space, remember.) Will you be accused of cherry-picking no matter how thoughtful and transparent you are? (Yes.) The questions continue.
At this juncture, the question for me was how to look at the myriad examples my interlocutors described to me and decide which ones I would analyze. I found myself spending much more time justifying my decisions here than elsewhere in the methods section—see the aforementioned cherry-picking critique. But also, and I think this is important, I made these decisions after collecting the data. I was home; my head was clearer; my environment was familiar. I had the luxury of time and space to think through things, which is not always available in the moment in the field. The only limitation was the word limit.
And…well. The word limit. I get it, and also: what a barrier to both equitable citation and rich qualitative work. This article is 11 double-columned pages long, with no tables or figures, and it is one word under the word limit. I bring this up here because it strongly influenced how much evidence I was able to present (and I think a number of episodes I chose not to analyze would aid understanding, such as events around the 2006 World Cup), but it also matters for citations. I’ve learned recently that it is not true across disciplines that journals count citations in word limits. Especially as we move to more and more online publishing, there is really no excuse. If we can’t be allowed to ramble along for as long as we want (and that you all are spared me is a good thing, let’s be clear), at least let us attribute the ideas that influence what rambling fits within the word limit.
Analysis (or what we might call “findings”)
If you thought writing a qualitative methods section was hard, try writing a good case study. I found it challenging to estimate how much my audience would already know about the instances I was describing, and thus how much space I needed to devote to background information.
Ultimately, I decided on a couple paragraphs at the start of each case study to sum up what had happened and explain what this case would tell us about responses to white supremacist violence in Germany. I also decided to keep background info to one short paragraph—and in the case of reunification, to punt on explaining most of the politics. I don’t know that this was the right choice, at least if the goal was to make sure readers were informed enough to evaluate the story I was telling, but I didn’t get any feedback from reviewers asking for more information, so I took that as a good sign.
I also dealt with how much to rely on direct quotes from my interviews or statistics from government reports I’d read vs. my impressions and takeaways, which political science struggles to consider as evidence. As any field researcher will tell you, insights tend to come from exposure and piecing together impressions over time—not the sort of thing you can take space to describe in an article. And so I tried to translate what felt like an organic (and in some cases, years-long) experience into a line of logic, with outside sources standing in for “after absorbing much German media, I had an epiphany in the shower.” This meant going out and finding corroboration for personal experiences, which is an odd thing.
The most useful piece of feedback I got from the reviewers came in this portion of the article: specifically, they pushed me to consider the role of international factors in shaping German domestic security policy. Anyone unfortunate enough to be familiar with my dissertation knows that I think this directionality is fundamentally backwards: domestic biases, embedded in global systems, shape our construction of international “patterns” and priorities. But it took someone opposing that viewpoint to help me find the best way to articulate it. In this entire article, I am most proud of the paragraph in the RAF section about the Years of Lead and other periods of prolonged right-wing violence. The instances I mention here are criminally understudied—I had to go to French-language sources to find an academic discussion of 1970s-era violence against the Algerian diaspora—and I’m excited to explore them more thoroughly in future work.
Conclusion
Fact: I adore introductions, and I abhor conclusions. I write introductions first, always, and belabor them far more than any other parts of my writing, because if I don’t figure out firmly where I’m going, I never get anywhere. Conclusions, to me, feel unnecessary, and I often don’t read them in academic work. (Sorry.) This may be because they often function as summaries, which I don’t need because I just read the damn article, and this contributed to my initial confusion with how to write one here.
The first draft spent much more time on the Lübcke anecdote, aiming toward nicely bookending the article. This was derided by everyone who read it—and, in retrospect, rightly so. Conclusions are most useful, I’ve learned, when they talk about not directions for future research (this is frequently speculative, self-indulgent, and boring) but rather possibilities for taking a theoretical framework and spinning out its logic to shed light on other cases. I suppose that’s a different way of saying future research, but I tend to think of the goal of research as being broader societal understanding, not a series of lines on a CV, and so I don’t like to frame “broader implications” as the latter.
Last thing: I end the piece with a quote from Sara Ahmed’s Living a Feminist Life, a masterwork that I turn to whenever I need shoring up or the validation of having my experiences believed. This quote initially appeared in the introduction. Though no one ever said outright that this quote was the cause, I received frequent comments on earlier drafts that my introduction was self-righteous, bordering on polemic. I never understood this feedback, and no one was ever willing to point out exactly what was so objectionable. Yet the minute I moved Ahmed’s quote on changing white supremacist institutions—a quote that, in context, refers specifically to academic institutions—to the conclusion, these comments stopped.
I wonder if this is evidence that, like me, many academics do not read conclusions, or that white readers need 10 pages to acclimate to discussion of white supremacy. I also wonder if it is a general aversion to thinking about what research findings mean for our own conduct as we go about our day-to-day lives. But it gives me hope, as a scholar-activist, that all I had to do was move the quote, not excise it. It gives me hope that, with a lot of work (and rambling) first, we can get to a place that sees racial justice as central to our research, our writing, and our academic institutions.
