Exploring the Eclipse of Equality

Analyzing Discourse: A Methodological Note

In Culture, Domestic Politics, Intellectual History, Media and Politics on April 23, 2013 at 10:26 am

Eclipse of Equality

As part of the celebration of the release of my new book The Eclipse of Equality: Arguing America on Meet the Press, I thought it might be helpful to share in broad outline how I put together the analysis there, something the space requirements of the manuscript did not allow in the book itself. I won’t go into great detail, but I hope this will be helpful to any of my students who have an interest in cobbling together some similar an analysis in their own work.

Step 1: The Puzzle Trap

The first priority in a journey of this kind is to find a coherent stream of discourse that you would like to analyze—some discursive arena on which to set your attention. For me, this came easily. In all of my political work on the United States, and throughout my academic career, I have found it necessary to keep myself grounded by watching the Sunday shows: Meet the Press, Fox News Sunday, This Week, Face the Nation etc. At first, I didn’t think much about why these shows were so important in grounding my perspectives on survey work or for my teaching, but it was always true.

I remember in one of my early job interviews talking about how I anchored the codes for my dissertation work (a massive content analysis on the association of open-ended themes and Presidential party vote in the National Election Studies (NES)) in the Sunday morning debates. The literature in the top journals was essential to getting a bead on what mattered in American politics, but what that kind of immersion provided me was what I would now call a set of puzzles that were relevant from the point of view of the scholars rather than a set of problems that real people were confronting in their lives as practical politicians.

This contrast between the puzzles we spin and the problems our subjects manage is the first methodological insight. If you want actionable results in your social science, you can’t rely on the way you and your colleagues frame the problems (in terms of puzzles in the literature), but rather an immediate sense of how the participants frame them as problems they themselves confront, and they frame them in their own rhetorical forms. When you take these forms seriously, you find that the boundary between pure and applied research begins to dissolve and the relevance of your work increases immensely, this at the expense of veering close to the social science/journalism boundary.

This puzzle/problem distinction is a fine point, but it is easiest to see if you have had any experience in framing a survey question. There is excellent intellectual and empirical work on how to write survey questions, but in the end, when you do this you are imposing your own language on the interaction and you then have to become suspicious that you are putting frames in people’s heads if not words in their mouths. This is how we fall into the puzzle trap. In the best-case scenario of survey research, when puzzle work is done well, what you will discover is your best guess of what the problems are for the participants, which can by design be little better than a translation of these problems into the ways your community frames its puzzles about that population. This becomes painfully obvious when you use historically rich datasets like the NES or the General Social Survey. The questions asked rarely survey a generation of researchers.

You can avoid the puzzle trap to some extent by using more qualitative and open ended forms of data collection, like observation, focus groups and interviews, but the puzzle problem remains in that you find yourself gravitating toward questions in data collection and themes in data analysis that favor the demands of the journal reviewers you will have later to satisfy. Successful work seems to invariably favor the reviewer over the respondent in these efforts, rewarding the solution of academic puzzles over social problems. The challenge is to tie the formulation of the puzzle you would like to solve in terms that make sense to the people involved in the action you study. If you do this well, the knowledge you generate can be plowed back into that setting in a way that can be immediately useful. Policy think tanks have this down, but they fail to try to push their analyses back to the kind of generality and playfulness that marks the intellectual life. Instrumentalism traps the work of the think tanker in an even worse problem, the principal trap, in which you give your funder the answer she wants to hear.

One attitude that promises to avoid the puzzle trap, the ethnographic mood, probably gets us closer to what we ought to be doing, but it is horribly particular in most cases (used for purposes of generalization amid vociferous denials of intention to do so or warrant from analytical design) and is marked by an almost romantic direction to geographical and cultural Others that one is meant to approach from a position of ignorance and humility (orientations that need not characterize interpretative social science). There is nothing wrong with this attitude, as such, but it is probably more appropriate to the era of colonial expansion than it is to our era of global convergence. The classic ethnographer studies someone else, not himself. Today, I am lucky enough in my own work to be surrounded by the global Other who is perfectly capable of studying herself. The anxious hand-wringing and paradoxical ethics of the ethnographic project feels very different when people of every ethnicity from all over the world are empowered to aspire to link their own problems to the puzzles of the towering journals of social science.

In a discourse analysis like the one I developed in Eclipse, I tried avoid the puzzle trap through a close analysis of the political culture of the beltway influential in its natural setting: one that I understood as a deep insider myself. The interviews were not objective reflections of what the interviewees truly believed, but they were reflective of what made the material in them relevant and interesting in context. In a way this is like the impulse of the great sociology of the Lynds in their book Middletown on Muncie Indiana. Discourse analysis is self-estranging in its tendency to flirt with a kind of auto-ethnographic perspective. I ground my research questions in the flow of the debate and speculate about the great questions of the examined life throughout, but avoid the temptation to shill for the source.

Step 2: Why do I care about my data?

Once you settle on a discursive stream of discourse that seems essential to you—for some reason—you have to figure out what the heck it reveals. Why do you care about the Oprah book club? What is the family resemblance that draws you to read the American Political Science Review as if it were a coherent literary whole? How about a particular hashtag? What does the front page of this radical newspaper have to say about who we were/are? About who I am? These are the kinds of questions you have to entertain as you begin your analysis. My advice is to pick a discursive stream because you love it–you have to love your data or you are sunk in social science—but you then have to come to some perspective on why you love it (or at least cathect with it). You have to come clean with yourself about what you love about the data, what about it arrests you and what it really says about the people who produce it by living their lives. There is a potential chasm between what it means to you and what it says about us/them, but you have to find your way to the latter insight through the former.

As an aside, let me say that the great advantage on choosing to enter a particular stream of data, thereafter determining its meaning, is that it is a bounded strategy—a sampling strategy. The worst thing you can do in analyzing discursive data is to go out and start sampling the stuff you see in some random and loose way (I’ve tried this and it’s a lost cause). You become little more than a web surfer and no one will have any reason to trust that your product is worth paying any attention to. If you choose a particular data stream (like the television show Meet the Press, National Geographic Magazine, 60 minutes, etc.), the boundaries of your project are fixed by the nature of selection. I had over 3,400 episodes to choose from going back to June of 1945, but the data universe was fixed by definition (although he population grew by hundreds of episodes in the course of my writing). This gave me a bounded empirical exercise to conduct and was my Ariadne’s thread throughout a tough and long project.

So what does the record of Meet the Press say about the people who made it and watched it and about American political culture in general? Here is where you lose your scientific bearings and become something of a litterateur. You have to somehow figure out what it is you will learn when you have learned what you can from the stream(s) in question. In my case, I fell upon an inspiration that justified my dedicating four years to coming to an answer. What became clear to me was that just as a political party exists in the government (in Congress, the courts, and the executive branch, but also in the lobbies and associated organizations) but also in the electorate (the people and their associations) so too “the narrative” –the generic story about the meaning of all the blooming, buzzing confusion of political life—had a life in the electorate (measured primarily by public opinion polls), but also in government. Just as there can be a massive disconnect between the party-in-government and the party-in the electorate, so too there can be a divide between the narrative-in-the-electorate and the narrative-in government. What I had fallen upon, what ultimately fixated me about the otherwise politically narrow Sunday shows, was that they documented the narrative-in-government. If I studied the history of arguments on Meet the Press (originally I wanted to study all Sunday shows, but the data were only preserved and available for Meet the Press), I could reconstruct a genealogy of the narrative-in-government, or at least the narrative-around-government, thereby documenting the arguments that drove policy irrespective of what ideas were captured in the Gallup poll. What I had, in fact, was a regular index of the conventional wisdom captured in a systematic collection of thousands of interviews and focus groups. When I had the eureka moment, I was elated, and I directed a big chunk of my life to learning what this record of the narrative-in-government could teach me.

Step 3: Non-parametric Narrative Parsing

I was trained to be a quantitative social scientist. This means that I was taught to look for the best measures I could find through which to depict the mechanisms that drive the social process. The problem for me from the first moments I suspended my disbelief in accepting how a multiple regression model worked was that I believed that social mechanisms have to be situated or positioned in cultural space for the social process to play out. When you read Wittgenstein at too young an age, it is very hard to imagine social life in terms of even complex, non-linear and stochastic social mechanisms. I was ever alert to the game and the freedom brought by language. I was therefore naturally suspicious of the mechanistic perspective in social science that was the premise of most of the models we develop.

This is not the end of the world and you can use quantitative methods in conjunction with a kind of cultural pragmatics, but there is a tension in that approach that tends to sink it. In formulating the Eclipse project, I was eager not to develop a kind of quantitative content analysis that purported to explain what was happening in my data (dump it all into your text miner and see what falls out the other end). I like these tools, but what you get in the end is little more than extra-intuitive description. So I chose instead to see what kind of story I could tell with the Meet the Press data. The story I would tell was an answer to a question: what does it say about a people that their leaders reveal the horizons of their moral imaginations in the way they do on fora like these? The data could tell a story about us, through the revealing discussions of our leaders.

The story I saw there was surprisingly exhilarating; we could learn from this television show how our ideals had transformed in the period in which the United States had its day in the sun. After all, it was after 1945 that we became the global power that I came to take for granted until the onset of this current depression and these folks who paraded before the cameras on the Meet the Press were the folks that stood out in that moment of glory. The data told our story for us, often despite our best efforts to challenge them. Most of the people I know don’t even admit to watching Meet the Press and its sister programs, but I could use the record I found there to reconstruct the larger story of conflict in America despite the unrelentingly centrist and cautious tone of the program. What a friend of my mine had once derided as Meet the Depressed, would be my guide through this period of economic depression into which I had brought my children.

The stylized narrative I came up with was the image of the Eclipse of Equality, by which I meant the gradual atrophy of a canonical category of the moral imagination. What this means is that we no longer have the chops to make powerful arguments that challenge the power of property, elucidate the horrors of class divided cultures as did Dickens and Eisenstein, and examine the actual workings of systems of opportunity and material privilege and convince large populations in ways that produce action that the (ab)use of economic power is among the most threatening forces facing a just society. We Americans just don’t buy that line anymore. That’s the main story of the book and it’s bipartisan. Neither party is much concerned about material equality unless it touches on market freedoms (the Republican answer) or cultural differences (the Democratic answer).

But how did I sift and winnow the data to come to that conclusion? The simple answer is data visualization (now valorized as non-parametric data analysis) in league with close observation of the sample (in my case the whole population of 3,500 episodes) and a habit of paying attention to the cut-points or discursive crises. I picked up the idea from Bill Gamson that we can learn the most from political talk if we pay close attention to discursive turning points and open moments at the birth of a metastory line that influence all future developments like a trauma guides a neurosis. Think of it as punctuated discursive equilibrium. To this end, I created an excel spreadsheet with the major premise for each show in the last column and then improvised thematic codes for these themes with my graduate students (especially Jana El Horr) that gave me some sense of where these cut points were. This approach also made it possible to create data subsets in which some larger conversation was hived off the larger discursive stream (in all this, the year 1978 stands supreme for reasons you can find in the book).

I found it useful to toggle between the plenary dataset and the subsets as I developed my chapter structures. In this process, the numbers of “conversations” dwindled from six to nine down to four biggies: foreign policy/war, economics/taxes, race/civil rights and class/labor. I had hoped to begin with the labor discussion because I think that the signal change in our political culture (other than the rise of the Civil Rights Movement(s)) was the falling political capital of the labor movement. The truth is that there were not as many episodes dedicated to labor after the period I describe as the Eclipse of Equality had begun, and people are just not as interested in the bitter hard history of the losers; so I moved it to the back even though this was a crucial piece of getting the story as a whole. Meet the Press began with labor depicted as almost as big a challenge to American freedom as the Soviet Union. It ended as a joke and a footnote.

Within a data subset, itself an evolving and updated category, I looked for those moments when you could easily see a before and after moment. The best example of this is in the foreign policy chapter. For most of the early foreign policy discussions lots of things are being discussed in a variety of ways, which in 1957 is immediately supplanted by discussions of every kind of missile system that you never heard of. The big event was Sputnik. After the Sputnik launch, the discussion was completely different. Remembering Gamson’s lesson, I paid close attention to these points of crisis and who the guest was in the early episode after the crisis (especially on the Sunday following the event). In this case, you learn that Meet the Press met the Sputnik crisis with the president of Harvard not with one of the leading military figures of the day. He advocated more education spread more equally and that helps to explain why our reaction was to create the system of colleges and universities we are only now abandoning as our global prestige reaches its apogee. That is a little wrinkle of the larger theme of the project and shows how little sub-narratives help to support the larger claims of the book.

Step 4: Choose your Story-telling System

There are all sorts of causal connections that you can explore in your discursive data, but the pay-off is to draw out some larger narrative that brings the whole into a kind of focus that someone not so close to the data might care about. This is the task of developing your storytelling system. I know a good many crackerjack quantitative social scientists who view the world as a system of random variables evolving over time, but they too pick a story telling system that allows them to write their abstract. That is a precondition to most publication opportunities.

My strategy quickly settled onto a five-part structure within larger conversations. I would break the larger story into sensible sub-narratives (it is cool from the historians point of view to watch the surprising conjunctures in these unfoldings) and I would then impose a somewhat arbitrary limit on the number of crises/traumas/turning points I would recognize in the stream to give the book a consistent flavor. I found that if I picked an introductory setting, three twisty-turny moments and a move to closure, I could minimize the violence done to the actual flow of conversation while picking up key crises that pushed the larger story along. This gave me a five-part structure within the larger system of chapters with four substantive chapters. The book then reads as series of twenty historical essays couched within a larger system with an overarching theme. The strategy was then to pass four times through the data from 1945-2012. The five essays within the posited conversations reveal a kind of unity of action, but the four conversational themes have to be understood in relation to one another.

There is nothing sacred about this five-part structure and I could have played with it to try to cut nature at different joints, but these cut-points are all quite plausible considering the patterns I could discern at the level of the premise data (what the show was announced to be about). For example, I don’t consider 9/11 to be a big game changing event, rather I situate it in the context of a larger debate between Jimmy Carter and Ronald Reagan about the balance between human rights and market freedom in the global arena. That was a big choice, but it made good sense based on how I read those premise data and how the concrete conversations played out on air. To my surprise Carter is the revolutionary in this story, who brings the cultural rights agenda of Martin Luther King to the global stage. Reagan is his counterweight, revivifying the victory line of the early Cold War and waiting for George W. Bush to push the moment after 9/11 to provide a temporary push for Goldwateresque victory over global evil ambitions.

There are many literary decisions like these in the book, but that is the business. Each decision about a turning point was hard won and I agonized over the decision in every case. This was complemented by the challenge of pulling the most pithy quote for each section form the transcripts/field notes that I could use for the title and section headings.

In the end, the five-part chapter structure (opening scene, twists one through three, and final scene with climax and denouement) suited the needs of the project by providing and abstract form of presentation into which I had to fit the concrete mega-narratives of the American national conversation. Much like the Sonata form would have guided the artistic decisions of a classical composer, once I settled on this form, I found the constraints imposed by this structure liberating. The story told is not objective in some simple sense of the term, but it does reflect my own experience with those newsreels, table viewers and 10 millimeter sound recordings that provided my life its texture through the process. There are many ways to build your storytelling system, and I don’t think we lose much in placing these beneficial constraints on ourselves, in fact we gain a lot in the way of self-discipline.

Let me close with a final comment on what is left out in such a process. Just the other day, we had a great event at S-CAR in which former Senator George Mitchell came to talk about his role in promoting the Good Friday agreement in the conflict in Northern Ireland. Mitchell is a big name—hard to imagine a bigger one in conflict resolution—and as I thought about giving him a copy of the book, I gave the index a D.C. read to see how he appeared in it. He didn’t! I had watched Mitchell on Meet the Press lots of times and thought of him in the sense of someone I might meet as well. In other words, I did my best to include him, but in the end, he was crowded out by other luminaries who had more of what I might call discursive impact than he had.

To get a sense of the scale of the problem consider that Sargent Shriver barely rates in my story and President Johnson plays a bit part. The conversation of the American Century was a crowded place and it sometimes looked past the people that helped to shape it in important ways behind the scenes. Of course, what I cared about was how the drama of our lives was performed before us, accenting that over the truth of process that is merely a contributing factor to our collective memories. It is interesting to note that my literary decisions reflect the reality Senator Mitchell himself sees about himself. In his talk he demonstrated that he was wise to his workhorse vs. showhorse role in the American moral imagination when he quipped about how he was mistaken for Henry Kissinger by a woman who have traveled a long way to attend his talk. Kissinger was the kind of story breaking persona who haunts our shared memories, famous for musing on how power makes you sexy, but even he himself played a limited role in my book. This is not hard to imagine when you are in the line-up with Martin Luther King, Robert Kennedy, and Ronald Reagan among others. This is all to say that the way we imagine and remember our collected lives is quite different from how we lived them, even though the symbolic experience in as crucial to who we are and what we will become as was that lived reality.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: