Structure Matters

Early in my essay, I distinguished the theory-driven, explanatory structure of analytic films from the kind of character-driven, dramatic storytelling that is used in many documentary films. I made this distinction because I think method and structure should be taken just as seriously in audiovisual scholarship as it is in written scholarship. We have no problem distinguishing between causal explanation and dramatic storytelling in written work. The same should hold true with respect to audiovisual work. Sunita Parikh disagrees. She says that my distinction between theory-driven analytic films and character-driven documentaries “creates a needless and potentially problematic schism.” She cites the documentary series Eyes on the Prize and the oeuvre of Frederick Wiseman as examples of films that are presumably “similar in their construction to . . . analytic filmmaking” due to the fact that they “take historical and social scientific events and processes as their subjects.”

I agree with Parikh that Eyes on the Prize and the works of Frederick Wiseman are important and insightful works about politics and society, but I disagree that this necessarily qualifies them as good models for social scientific filmmaking. In fact, I believe these two examples help to highlight rather than undermine the importance of the distinction I was trying to make. For instance, Henry Hampton, the executive producer of the Eyes on the Prize series, is not known for structuring his films around causal explanations, as I recommended in my target essay, but for his use of dramatic storytelling techniques and character-driven narratives. Sam Pollard, who co-directed two episodes of Eyes on the Prize, reflected on Hampton’s style and influence in an interview with Sheila Curran Bernard, another Eyes on the Prize alum, in Bernard’s book Documentary Storytelling:

Similarly, in a 1998 interview, Frederick Wiseman discussed the role of dramatic storytelling in his work: “The first thought: I’m trying to make a movie. A movie has to have dramatic sequence and structure. I don’t have a very precise definition about what constitutes drama but I’m gambling that I’m going to get dramatic episodes.” ^{Footnote 2} Elsewhere, Wiseman has stated that he rejects “thesis-oriented films” and approaches filmmaking “just like someone writing a novel.” ^{Footnote 3}

The point is this: Social scientists do not structure their written publications so that they “escalate dramatically and emotionally.” They do not gamble that their field notes will be full of “dramatic episodes.” Why should social scientific filmmaking be any different? From the social scientist’s perspective, the problem is not necessarily that a researcher-filmmaker would choose to explore a dramatic topic or that audiences would feel emotion when watching a work of audiovisual nonfiction, but that structuring one’s film with the goal of telling a story that escalates “dramatically and emotionally” creates incentives to present information in ways that prioritize the arousal of emotions over the presentation of the most accurate, objective account. For example, looking back on one of the episodes he directed for Eyes on the Prize, Pollard says that although he presented Muhammad Ali’s “real story,” he also “feels a sense of artifice”—that the story “feels a little jerry-rigged.” ^{Footnote 4} These are precisely the kinds of reactions that good social science seeks to avoid. Furthermore, theoretical explanation will take a backseat when one’s primary goal is telling a dramatic story with protagonists and antagonists and a beginning, middle, and end. Three-act narratives are simply not how social scientists structure their publications. For this and other reasons, my objective was to propose an alternative approach to nonfiction filmmaking that uses causal theories, not plot, as its backbone. To be clear, I am not arguing that creative nonfiction is incapable, whether through text or film, of producing and communicating knowledge in ways that are insightful and important. Rather, I am arguing that creative nonfiction is not a good model for social scientific filmmaking.

Although Dvora Yanow has her doubts, the submission guidelines and decisions of the principal organizations that fund, exhibit, broadcast, and distribute serious social and political documentaries make clear the documentary filmmaking profession’s emphasis on character-driven, dramatic storytelling. ^{Footnote 5} At the same time, Yanow is not incorrect when she says that “there is more to documentaries than ‘eccentric,’ ‘outlier’ storytelling films and to documentary filmmakers as those identifying as storytellers or focusing primarily on ‘engaging characters.’” For example, some documentaries, like Alex Gibney’s Taxi to the Dark Side ^{Footnote 6} and Charles Ferguson’s Inside Job, ^{Footnote 7} are more concerned with advancing causal explanations than telling character-driven stories. These thesis-oriented documentaries, however, are relatively rare, vary widely with regard to scientific rigor, and are often more interested in making normative rather than positive arguments. Furthermore, dozens of important ethnographic films have been produced over the past century. Many social scientists, however, are not ethnographers and will be uncomfortable with the kinds of “reflexive, collaborative or participatory methods” that tend to characterize ethnographic filmmaking. ^{Footnote 8} Therefore, if social scientists are to use film to disseminate their findings, they may seek an approach that more closely aligns with the qualitative approaches they and their colleagues are already familiar with. My objective was to advance a filmmaking approach that is more consistent with my training and the training of many other political scientists.

Jeffrey Gould suggests that “Germano’s efforts might be more fruitful if he blurred the lines somewhat between the strictly expository quality of analytic film and the individual and group stories of transformation that characterize documentary film.” Similarly, Henry Farrell suggests that “the analytic narratives employed by Robert Bates, Avner Greif, Margaret Levi, Jean-Laurent Rosenthal and Barry Weingast present a different, and perhaps more palatable model for analytic filmmaking.” Farrell goes on to say that “the tension between the abstract and the particular, the model and the story it is supposed to illuminate, is a useful one.” I agree with Gould and Farrell that hybrid structures—a sort of mix between analytic and ethnographic filmmaking—could be useful. Scholars using a hybrid structure would establish theoretical pillars to develop their films’ expository structure. Then, to illustrate theories and processes, they would embed ethnographic or historical vignettes into the larger theoretical structure. I experimented with this approach somewhat in The Other Side of Immigration. The story of Carolina’s family [32:35–41:47] illustrates dynamics that are discussed in more general terms by various interviewees. Scholars who use ethnographic or historical vignettes in their analytic films should make an effort to present a wide range of stories selected on the basis of some set of theoretically relevant criteria or to present stories that represent the average tendency. There must also be an effort to connect individuals’ stories to larger trends in order to demonstrate how consistent or inconsistent these stories are with the average tendency.

Finally, on the topic of structure, I think Parikh raises an important point when she argues that “if analytic filmmaking is to fulfill the promise Germano expects, then the theoretical framework and the connection between theory and data need to be presented much more explicitly so that the viewer can be guided by it as the evidence is set forth.” The question is how exactly how to do this. One approach, which Parikh recommends, is voice-over narration. Ethnographic filmmaker Robert Gardner is known for making arguments and insights explicit through narration, as are thesis-oriented documentarians like Gibney and Ferguson. Another approach is to break one’s film into chapters and use on-screen text to help explicate theoretical arguments. A third approach would be to publish short essays that explain how one’s analytic film makes its theoretical argument, similar to the “Communicating Arguments and Evidence” subsection of my target essay. These essays might appear online, published alongside the video. Scholars may also write separate articles that make explicit, reference, test, or expand on theoretical insights that are otherwise implicit to the structures of their analytic films. Finally, scholars may create an audio commentary that can be played over the regular audio track (like a director’s commentary on DVDs) or a video appendix at the end of the film. These approaches allow the researcher to explain certain methodological decisions and make theoretical concepts explicit without interrupting the flow of the film. One can imagine a film that uses all or some of the above approaches in combination.

Analytic Film and Scientific Inquiry

Farrell argues that “because films are dense, rich narratives, they can convey an intimacy of detail that standard kinds of social science knowledge cannot. Yet it is exactly the most valuable details that are least likely to fit with the promised ‘hypothesized logics.’” I disagree that there is necessarily a trade-off here. Video simultaneously transmits many different kinds of knowledge, meaning that it has the potential to convey both intimate and general information at once. For example, many people who have seen The Other Side of Immigration have mentioned to me that they found the people I interviewed to be “articulate” and “insightful.” That the hundreds of smallholder farmers and return migrants I encountered while doing fieldwork in marginalized Mexican communities tended to be articulate, insightful people is not a detail I thought to include in my written work. This detail, however, comes through naturally in an audiovisual publication and may be valuable to some viewers in the context of more general causal explanations regarding the political economy of emigration. In response to Davide Panagia, it was the ability to present such different kinds of information simultaneously that I found particularly fascinating and revelatory about using video to complement written and quantitative analyses.

Panagia raises a point I hear often when he describes video editing as “a notoriously slippery, elusive, and evasive art.” Yanow adds that “all film-makers, [Germano] included, shape the stories they tell,” and that cameras are “set up to frame some things in and others out.” All true statements, but aren’t all forms of knowledge production potentially deceptive exercises in editing, framing, and shaping? Quantitative research is full of subjective choices, among them, how to measure concepts and collect data, which variables to include in one’s model and which to exclude, how to deal with violated assumptions, and which models and results to report to the public. Writing involves countless decisions about how to interpret material and what to include or exclude in final drafts. Just as a video editor can use images to misrepresent a situation, quantitative analysts and writers can easily fabricate data and present data in ways that distort meaning. Does the fact that human beings are involved in the research process mean that we should make no effort to reduce the impact of personal biases or that we should not develop rules, norms, and review procedures that discourage deceptive uses of data? Of course not. Indeed, just as we do in written and statistical work, we need rules for audiovisual scholarship that help us approximate objectivity, make our methods and level of participation transparent, and open the door to scrutiny and replication by other scholars.

Panagia raises an important question when he asks, “What kind of evidentiary support, if any, [can we] expect from a moving image?” To begin to answer this question, I think it is important to first ask what kind of evidentiary support we can expect from any work of social science or any kind of data or evidence used in the social sciences. Most social scientists accept that we are in the business of persuasion and are rarely, if ever, able to offer definitive proof for our propositions. We attempt to persuade by advancing theoretical explanations, often in tandem with data that offers only a modicum of support for our arguments. We reduce uncertainty and improve our understanding of social and political phenomena by marshaling as many kinds of data and evidence as possible, whether in the same study or as a community of scholars. Indeed, the act of accumulating many different kinds of quantitative and qualitative evidence to develop a clearer understanding of big questions is what I meant when I wrote that analytic films help us develop a “more accurate and complete body of scholarly knowledge.” Analytic films alone do not make our understanding fully accurate or fully complete, but they may complement our text-based and mathematical methods, revealing new pieces to the puzzles that occupy us. Combining different kinds of data can, as Gould argued in his essay, lead to new insights that take our research in new directions (see also note 36 of my target essay for an example).

Finally, I agree with Yanow that it is important to consider ethical questions when conducting audiovisual research. In my audiovisual work, I obtain a signed release from anyone appearing on camera, offer to protect identities when appropriate, and honor requests not to be filmed. In the editing phase, I give careful consideration to people’s reputations and future personal safety when deciding which clips to include in the final cut of the film. While I believe that high ethical standards are important, I also think it is important not to overstate the risks of conducting audiovisual research. Broadcast journalists, documentary filmmakers, and reality TV shows have filmed people around the clock for many decades. Journalists often do not obtain consent or alert those they are filming. At the university level, journalists and documentary filmmakers are typically not subject to the same institutional review board (IRB) requirements as social scientists. The risks of audiovisual research may in fact be lower than other kinds of research because just about everyone knows what a video camera is and what they are getting themselves into when they go on camera. The same is not often true when a researcher knocks on someone’s door asking for responses to a questionnaire, as I discovered while collecting survey data and video data in parallel.