Below is an interview I conducted on January 21, 2023 with my recently retired (but still research active) colleague Ben Page. Given the influence and importance of his work in areas like public opinion, elections, and democracy, I thought it would be good to get some of his intellectual trajectory down on paper. I’d add that the questions and structure of the interview are inspired by Munck and Snyder’s Passion, Craft, and Method in Comparative Politics, a book I’ve discussed earlier on this blog. Because of the length of the interview, I’ve divided it in two parts. You can find part 2 here. Enjoy!
Intellectual Formation and Training
Q: How did you first get interested in politics?
BP: I grew up during the Cold War, and my family paid a lot of attention to what was happening in the world. My mother led tours abroad and lectured on foreign affairs. There was a lot of talk about politics at home. I was worried that we were going to blow ourselves up in a nuclear war.
Q: Was your family very political?
BP: My dad, a geologist at Stanford, was not all that political. My mother was an Eisenhower Republican. When I turned Left as an adolescent and got angry about the U.S-backed coup in Iran and the overthrow of Arbenz in Guatemala and so forth, we had some lively arguments.
Q: What was your early education like?
BP: I was very lucky. The Palo Alto public schools were good, even middle school. Stanford Elementary was full of professors’ kids and some electronics wizards from the emerging Silicon Valley. David Packard the younger wired our classroom doorknob with a Ford coil and jolted our teacher with a couple thousand volts. (Not many amps, so no serious harm done.) Very stimulating.
For high school, I was especially lucky to get a scholarship to Phillips Exeter in New Hampshire. I failed to acquire the “Eastern polish” my mother hoped for, but the intellectual experience was great. I was a complete “grind” – as the social preppies used to put it – working away at Latin, German, and math, and writing weekly English essays that Darcy Curwen handed back covered with red-ink corrections. I got to meet distinguished visitors like Robert Frost and Robert Oppenheimer. Two gripping years of chem-phys, complete with foaming reactions in the lab and soap bubbles of hydrogen and oxygen exploding in the air. That set me off toward a natural science career. Fine history courses, too; I learned a little but was not very good at it.
Q: As an undergraduate you studied history at Stanford.
BP: Yes, I have to admit that I sort of drifted into history and poli sci. My plan to study the biology of the human brain quickly went aground, because at that time there was no biology of the human brain. Biologists were struggling to understand three-nerve systems in sea urchins. People had barely dreamed of AI. Also, my math classes, designed for engineers, were missing the fun theoretical stuff.
So, I started taking liberal arts courses. I loved Shakespeare, the Hebrew Bible, Rilke, and English literature. And U.S. history. But I began taking more and more old-style, historical/ descriptive political science courses about the contemporary Middle East, Africa and Latin America. Even then, it never occurred to me to become a professor, until my advisor (Latin Americanist John Johnson) made a pitch that being a professor could be a good thing.
Q: But your dad was a professor, right?
BP: Yes, he was a structural geologist at Stanford, interested in mountain-building around the Pacific rim and elsewhere. That got us great trips to Japan, Italy, and Yugoslavia. And I suppose you could say that my whole career has been an extended dialogue between my father and my mother, science and politics.
Q: You ended up going to law school. Was your plan to be a lawyer?
BP: Not exactly. It was to become an actor in politics. My family’s scheme was a little vague, but it involved getting a law degree, becoming an investment banker and making a lot of money, contributing to the Republican Party, and getting an ambassadorship. Or maybe running for the Senate or even the Presidency. Unfortunately, this plan did not take account of my actual interests or abilities. But I was happy with the idea of engaging in political action, maybe working at the State Department or a UN specialized agency.
With international action in mind, I studied international law at Harvard Law School with the great United Nations law scholar Louis Sohn and with the self-styled “international plumber” Roger Fisher, who specialized in drafting treaties that countries might actually obey.
Q: How did you end up returning to Stanford for your PhD?
BP: Well, I realized I was not at all suited for running for office. A summer at the State Department gave me a dim view of the foreign service. But it got me reading confidential files about Africa, where the CIA was currently trying to manipulate Congolese politics with South African mercenaries. Reading those files made me think about doing research instead of political action. I thought, let's take this law thing and work out a behavioral theory of international law. That might even help in getting countries to follow the law. Robert North, a wise IR scholar whom I had admired as a Stanford undergraduate, seemed like a good person to do it with, along with a couple of law professors.
Q: What was graduate study at Stanford like at the time? What sort of training did you get and who were the major figures in the department?
BP: There was one senior leader for each the three main branches of political science: Heinz Eulau, Gabriel Almond, and Robert North. And then there were these bright youngsters: Sidney Verba, Raymond Wolfinger, and Richard Brody. I started out thinking North, but very quickly I was taking courses with the younger guys.
We got along well. I published a short paper with Ray Wolfinger, and he started me off on survey research. He taught me how to analyze survey data the old way, via IBM batch processing and counter-sorter. Counter-sorters let you get to know ANES respondents personally, because each card in your hand is a person. I learned a little Algol from a computer science class but never became a good programmer. Stats training was feeble.
Dick Brody was terrific at explaining what political science was supposed to be. Law school had been very different. At Stanford, my office mate Charlie Levine had to tell me who Robert Dahl was. I'd never heard of a hypothesis. I had no idea why you might want to test one, or how you might do it. Dick had us reading philosophy of science and classic empirical studies. He nudged me towards quantitative IR. I wrote a paper on whether the outcomes of international arbitrations in Latin America might be shaped by the backgrounds (mostly European) of the judges who were imposed by the U.S.
Elections and the Median Voter
Q: Why did you move from international to American issues at this time? Did your interest in international law somehow get you you interested in American politics?
BP: Yes. The Vietnam War did it. That war was the huge international thing happening while I was a graduate student. It was absolutely horrible. Millions of people killed or injured. One of the reasons I gravitated toward Brody and Verba was they had been studying U.S. public opinion about the war and were setting up a study of the 1968 election. I figured that if anything could end this terrible, unpopular war it would be the American people, through a democratic election. So I signed on with Dick and Sidney as a research assistant. We got some NSF money and designed items to go on four ORC piggyback surveys in February, June, August and November 1968.
Q: So that took you into individual political behavior?
BP: Yes, especially issue voting. Ultimately Verba and Brody persuaded the Michigan people to put a few of their items, including their Vietnam and urban-policy seven-point scales, onto that year’s ANES survey. So far as I know, this was the origin of seven-point issue scales, which were subsequently used by the ANES for several election years. Unfortunately they had measurement problems. Including failures of unidimensionality, like people wanting to “win or get out” of Vietnam. And people’s idiosyncratic, non-comparable interpretations of the scale units.
It turns out to be very hard to study issue voting. I think Dick and Sidney were on the right track theoretically, thinking spatially about respondents’ perceived issue distances from candidates, and exploring how those perceived distances related to evaluations of the candidates. Feeling thermometers measured the evaluations fairly well. (As Stanley Kelly showed, comparative evaluations of two major-party candidates translate almost perfectly into votes.) It was right to try for interval-level measurements of respondents’ policy stands, their perceptions of candidates’ issue positions, and their evaluations of candidates. And it was right to analyze the reciprocal interplay among all three of those variables, through processes of policy-based evaluations, selective perceptions, and candidate-led persuasion.
But besides the measurement problems, statistical modelling proved to be extremely difficult. The brilliant polymath Hayward Alker helped us devise a three-equation model in which each of those three variables was a dependent variable in one equation and an independent variable in the other two. Stats whiz Gerald Kramer said, “Do you believe in your model? Well then, estimate it.” But no one could figure out how. The equations were horrendously nonlinear. And we could not come up with the instrumental variables that were needed to identify the system, variables that were believably exogenous. We essentially gave up. All that wonderful survey data came to not much. I got discouraged about the whole possibility of learning about cognitive processes with surveys.
Q: Your dissertation was not really about voting behavior, was it?
BP: No. I worked hard on the survey research side of the project. But my main contribution was different. By the time I joined up I had read Downs in Heinz Eulau’s empirical theory seminar. I said to Dick and Sidney, you're building this whole thing around issue voting. You're expecting a big referendum on Vietnam. What if the simplistic version of Downs is right – that the two parties will converge, take the same issue positions, on Vietnam and other issues? There won’t be a choice. Why should anyone vote on policy, if there is not a dime’s worth of difference between the major party candidates?
So in tandem with their project I started a dissertation built around the idea: let's see what happens with the candidates. And it was a great year to test my hypothesis, because there were tons of candidates. I can't even remember them all. But on the Republican side, besides Nixon there were George Romney and Nelson Rockefeller (Vietnam liberals) and Ronald Reagan (a Vietnam hawk). On the Democratic side, Robert Kennedy and Eugene McCarthy on the left, George Wallace on the far right, and Hubert Humphrey more or less in the middle. One way or another, in the pre-nomination process the “extremists” were all filtered out. This nicely fit my story that the most centrist candidates (Nixon and Humphrey) would be nominated by both parties, and that there would be little choice on Vietnam.
The main thing I did for the thesis was to gather the texts of campaign speeches. All the speeches I could find by Humphrey and Nixon. (Humphrey later asked me, “Why on earth would you do that?”) And some speeches by candidates of other years, especially 1964 and 1972, that various scholars helped me out with. The whole motivation was basically to test the folk theorem on convergence that later became known as the “median voter theorem.” Nobody actually had such a theorem, except for Duncan Black, and that was just for the very simple, obviously unrealistic case of jointly single-peaked (one-dimensional) preferences and perfect information, everybody knowing everything.
Q: The research conflicted with your political beliefs, right? You were hoping for a more anti-war candidate?
BP: For sure. I was a Robert Kennedy guy, but he was assassinated. And my second choice, Gene McCarthy, was wiped out in the primaries.
Q: But your theory and the data pointed to centrists.
BP: That's a good point, because I was pretty liberal on domestic policy and definitely a leftist on U.S. foreign policy, much to the left of Hubert Humphrey. But at the same time, I did believe in small d democracy, and it seemed to me that Downs had an interesting model for its workings. I mean I hate to even say Downs, whose analyses were much more complex and insightful. But the simple spatial model. That was one way in which you could imagine two-party elections bringing about democratic control, by producing winners whose policy stands were preferred by a majority of voters to any other policies. I guess I was ambivalent.
Q: This was your dissertation research. It ultimately turned into the book Choices and Echoes. You focused on the 1968 election. But in the 1972 election, it wasn't the most centrist candidate who won the Democratic nomination, was it?
BP: See, that's the great thing about it. When I carefully read McGovern’s speeches, it turned out that he was very centrist. He was subject to a really dreadful defamation campaign, and Hubert Humphrey was partly responsible for that. McGovern proposed some smart ways to cut the military budget. And a mild income-support, “demogrant” program, not very different from Nixon’s and Moynihan’s. But Humphrey said, Oh, this will be, I forget the number, a five-trillion-dollar Treasury transaction. And everybody said, Oh oh, that must be radical. Plus, McGovern was associated with his student supporters, many of whom were anti-war protestors, smoked pot, had long hair, and so forth. McGovern himself was not at all like that. He was a moderate prairie populist. So actually, McGovern’s nomination was an especially important case to show that most major party nominees for president – other than Goldwater, who lost in a landslide – were taking pretty centrist positions. Based on my rather crude measurements, the overall pattern seemed to be that anytime you have a two-party election in a big, diverse constituency, there's going to be a force pushing candidates toward the center. Even Trump in 2016, though he governed very differently than he had campaigned.
Q: This work drew strongly from rational choice theory and particularly Downs’s spatial theory. Was this controversial at the time?
BP: It was new and exciting. I was one of the earliest fellow travelers with the rational choice people. After Downs, I read back to Hotelling, Duncan Black, and others. Then I got into the work of the Carnegie Mellon people, Davis, Hinich, and Ordeshook, with their clever linear-algebra-style mathematical models of two-party elections. I guess that ended up going essentially nowhere, because the “chaos” results by McKelvey and others seemed to show no hope of finding an equilibrium – so the theory made no empirical predictions. But at the time I dipped into it, in the middle and late 1960s, this was exciting stuff. A new way to think about electoral competition and voting in spatial terms. That's where our seven-point self-rating and candidate-perception scales came in. Of course there were plenty of political scientists who didn’t like it, too.
Q: How in general has rational choice theory influenced your work?
BP: Quite a lot. But I have always found limitations in the whole economic way of thinking about the way the world works. Especially when manifested in narrow-self-interest and perfect-information theories of politics. There are a lot of other things going on. Back then I was not only reading Downs, I was also reading E.E. Schattschneider and Grant McConnell, who talked about biased pluralism, about distortions in American democracy. And C. Wright Mills, who had an all-out elite-power view of how things work. I have always thought there was some truth to both of those camps, as different as they seem to be. (Downs actually started to bring the two perspectives together, in his chapter on uncertainty producing persistent political inequalities. And Mancur Olson, as I read him, did something similar about interest groups.)
Q: Another finding in Choice and Echoes seems to be that candidates tended to take a clear stance on issues.
BP: Well, no, not exactly. There is a chapter about ambiguity in the book, and I did an article on an “emphasis allocation” theory of ambiguity that was supposed to counter Kenneth Shepsle’s risk-acceptance theory. Yes, if you look carefully at campaign speeches, you can usually find some sort of a stand on pretty much any policy issue. But in fact candidates do not generally emphasize policy specifics. Why offend anyone? They talk about stuff that everybody agrees about, motherhood and apple pie. And they talk about the incumbent’s supposedly wonderful or awful performance. Candidate ambiguity makes policy voting harder and makes certain kinds of democratic control harder, though it has less effect on retrospective voting and the electoral-reward-and-punishment type of democratic control.
Q: How would you characterize the main results in the book?
BP: Number one is the tendency towards centrism, trying to respond to the typical voter. But number two, there's a chapter about party cleavages and systematic differences between the major-party candidates. I showed that such differences could be predicted from poll data on what rank-and-file Republicans and Democrats disagreed about. But I was pretty sure that was a proxy for centripetal influence on the parties by money-givers and activists. That’s where a seed was planted for later studying the wealthy. Back then it seemed likely that there was something happening to make the parties different from each other that wasn’t just ordinary voters.
I guess the third main finding is that there was clearly retrospective voting going on. Nixon definitely encouraged it, with his talk about “sirens in the night,” riots in the streets, and so forth. The idea was to get voters to punish those damned Democrats, without having to say what he himself would do to solve those problems. So, that’s three. And then the fourth is ambiguity.
To me, the broader contributions of the book might be two. One would be thinking about voting behavior in the context of a whole political system, in which the choices that people are given are going to have a lot to do with how they behave. That's where the title comes from. Second, I tried to sketch out three different ways by which (in theory, anyhow) a two-party system might lead to democratic results. “Responsible parties,” with sharp policy differences and policy voting. Downsian electoral competition, with convergence on what the typical voter wants. And politicians’ anticipations of V.O. Key-type electoral reward and punishment, which works through retrospective voting. Each of the three has normative pros and cons. Each probably describes a bit of what goes on, but none works anywhere near perfectly.
Q: Was there a conventional wisdom or authors that you were arguing against?
BP: I had a bone to pick with the responsible-party thinking of my hero Schattschneider. But I think my main target was the early Michigan school, which was a huge force in political science at that time. They led the “behavioral revolution,” and their 1950s and 1960s books really dominated the field – as did the ANES data that Warren Miller worked so hard to produce. Very useful. Lots of good stuff there. But I thought their social psychological view of individuals was not very political, and they tended not to think of voting or elections in democratic theory terms. For example, Philip Converse’s “Mass Belief Systems” article is very smart and clearly tells some important truths in it. It's true that the average person does not have a complete, well-worked-out political belief system. But I think it's a mistake to give the impression that people don't know a thing about politics, which implies that officials don't need to pay any attention to them. That's really what I was most strongly against. I was on the side of V.O. Key, especially his idea of party identification as a sensible “standing decision” about policy. And his Responsible Electorate book, which offered a new take on rational voting by “switchers.” I was on that side, basically as opposed to what I saw as the Michigan social psychology side that seemed to attribute everything to mindless party ID.
Q: If you look at subsequent presidential elections, do you think your early findings have held up?
BP: I think surprisingly much so. Take the 2016 Trump election. I have paid a little attention to Trump's speeches. In fact, he was surprisingly centrist on policy as a candidate. He made a big point that he was not going to muck with Social Security or Medicare. He was going to provide lots of jobs through a giant infrastructure program. And it was only in December, after the election, that he said he'd been told that big infrastructure spending was not really a Republican idea. That sort of campaigning is part of how he beat all the other Republican candidates. They were way out of touch with public opinion, on social welfare issues as well as immigration and trade. And for better or worse (it turned out to be worse), Trump sounded much closer than other Republicans to where the average voter was.
Jobs and Colleagues
Q: Going back a bit, how did you get your first job at Dartmouth?
BP: That was largely because Denis Sullivan had visited Stanford a year or two before and I was a teaching assistant for him. He studied political psychology; very smart and very broad. I assume he arranged the Dartmouth offer to me. It was an appealing idea to join Denis, and nice to do so on a leafy New Hampshire campus that reminded me of Exeter. A highlight of my Dartmouth experience was Denis, Jeff Pressman, and me taking some twenty-five undergrad students down to Miami for the 1972 Democratic convention. We had fun writing a little book, now lost in the mists of time, called The Politics of Representation.
Then, after a year at Dartmouth, I got a postdoc from the Social Science Research Council to study economics. It was a great interdisciplinary program that I think they abandoned a while later, but I was there at the right time. And so we went to Cambridge and lived off Central Square, halfway between Harvard and M.I.T. I shuttled between them by MTA.
This is just amazing to look back at. I audited courses with Kenneth Arrow, Paul Samuelson, Franklin Fisher (the great guru of identification in econometrics), Jerome Rothenberg, who knew a lot about welfare economics. Michael Spence on decision theory. And I’m almost forgetting Herbert Gintis, the great dissenter, with his radical left kind of economics.
This was maybe the pivotal year of my career because I learned so much from all these guys. Arrow’s seminar was basically a politics seminar, a sort of an expanded social choice seminar. And I was sitting in a little room with just six or seven other faculty and students. That was quite something.
Q: That’s three Nobel Prize winners.
BP: Yes!
So then while I was at Dartmouth, we were perfectly ready to settle there for the rest of my career. We thought about living in classic Norwich, Vermont. But the bright lights called, I got an offer from Chicago and came out to the U of C.
Q: What was the Chicago department like at the time?
BP: Oh my. That was an exciting department. That was sort of my second graduate school, because Stanford, despite the youngsters, it had limitations, especially in its downgrading of political theory. Chicago had all theory you could ask for. The most central thing was – to use Fred Block’s distinction – instrumental Marxists arguing with structural Marxists. At that time, Adam Przeworski was a fierce structural Marxist, and the instrumental Marxists were led by Ira Katznelson. Everyone was a theorist. Lenny Binder was a phenomenologist. Joe Cropsey led the Straussians. Mort Kaplan wrote a book about Rawls. David Greenstone, my leader in American politics, was studying Wittgenstein and the thought of Abraham Lincoln.
So it was just an exciting place to be. I co-taught a democratic theory seminar with Brian Barry. Unexpectedly, just a day or two before class he asked me to stand up at the blackboard and prove Arrow's theorem. That was harrowing. But I faked it well enough so that nobody seemed to notice that I didn’t know what I was doing.
Q: Just to finish the course of things, you didn’t stay there. You ended up moving to Texas.
BP: Yes. There were complicated issues involved in that move, and my family was not totally happy with it. For me there was some push as well as the pull of a nice UT chair, because all those Chicago theorists weren't really much interested in my sort of data-grubbing American politics. Even my friend and mentor David Greenstone cared more about the thought of Wittgenstein, Lincoln, and Jane Addams than about survey data. Although I did spend considerable time at NORC and learned a lot about survey research from James Davis, Andrew Greeley, and others.
But basically, I decamped to Wisconsin, starting with a summer at the Institute for Research on Poverty. The economists there were really great, especially the poverty economists Irv Garfinkel and Sheldon Danziger. So I took a teaching job with the Wisconsin department, where I had an end-of-the-hall office right between Murray Edelman and Gina Sapiro. Lovely. And Richard Merelman and Joel Grossman were just down the hall.
But Chicago came back with a tenured offer. I had left for a tenured appointment at Wisconsin and Chicago came back the next year with a tenured offer. For better or worse, we went back.
Q: That was Chicago to Wisconsin and back to Chicago. Then you went to Texas.
BP: Yes. By then people were calling me a “rolling stone,” and they didn’t mean a rocker. I did not stay in Chicago very long the second time. A glimpse into why might come from one anecdote. An unnamed colleague said, “Ben, welcome back! How does it feel to be taken seriously now?” I thought gosh, I had worked hard at the U. of C. for five or six years. I had hoped I was taken seriously then. But there was also a pull. Texas had a ton of money at that time. They claimed to have 120 endowed professorships and to be building a Harvard on the Pecos. And I found that idea exciting, especially since I was still very enthusiastic about rational choice. Two of the Carnegie Mellon people were there, Peter Ordeshook and Melvin Hinich. And two young Caltech whizkids, Gary Cox and Matthew McCubbins. Plus Tom Schwartz. A remarkable group that Chuck Cnudde put together. And then I persuaded them to hire Tom Ferguson, who is totally different, a macro political economy person, money and politics. I became a close colleague and friend of Tom's, and he was a key part of my vision of these two camps sitting on opposite ends of a log and engaging in rational discourse. But instead of civil discourse we got civil war. It was pretty bad.
Q: There were some conflicts within the department?
BP: Looking around the country I have seen a fair number of political science departments in conflict. Sometimes it’s one subfield against another. Often it has to do with methods and/or ideology. And it's particularly bad when conflicts about methods and ideology coincide and reinforce each other. Our UT rational choice people were mostly very conservative. Peter Ordeshook would go off to Pittsburgh to fight against the teachers’ unions. Tom Ferguson wasn't exactly advocating violent left-wing revolution, but he did not sound keen on capitalism. So that ideological difference reinforced methodological disagreements about macro historical methods versus formal theory and so forth.
Q: Is that part of the reason you moved back to Chicago, to Northwestern?
BP: That's right. And part of the pull of Northwestern was getting back to Yankee land, frankly. And back to the Chicago area. But also, my good friend from Hyde Park days Jane Mansbridge was there, and I liked her work a whole lot. And Jay Casper, who I’d known back at Stanford. So I really liked the idea of going up to Northwestern, and in fact settled down there for the whole rest of my career.
Being a slow learner, at NU I initially tried to recreate my Texas dream of rational choice/ formal theory dialoguing with macro political economy. But John Stephens and Meredith Woo and Cathy Jo Martin (along with Tom Ferguson from a distance), and later Kathy Thelen, and then Dan Galvin and Chloe Thurston, drew me toward what is now called American Political Economy. Outside the department, important colleagues included Susan Herbst and Bob Entman in communication studies, and – at the Institute for Policy Research – Fay Cook and Rebecca Blank.
Policy Preferences and Responsiveness
Q: At some point back at the University of Chicago, you began working with Robert Shapiro and gathering a large dataset of public preferences. How did this project begin and what was your initial intention?
BP: I think the core idea came from a conference that John Ferejohn set up, where some young scholar I can't remember -- I wish I could – said in passing, ”Well, one way to test the median voter theorem is just see if policy changes when public opinion changes.” And I said, “Oh, that sounds like a good thing to try.” So Bob Shapiro – a terrific grad student – and I applied for a bunch of NSF money, which surprisingly we got. We put together this enormous team of outstanding research assistants, notably including Kathleen Bawn and Paul Gronke. I think we had 28 research assistants at one point, which was possible only because the U. of C. had an admissions policy of letting in a ton of graduate students. So, lots of talented research assistants were available. We started out with that idea about opinion change leading to policy change, but the project pretty quickly grew into collecting all the time series of identically repeated items we could find on collective policy preferences – that is, percentages of Americans who favored one thing or other – and seeing how those percentages went up or down over time. Bob Shapiro supervised that effort, supervised the grad students. He did most of this huge data collection thing. All the data were on paper. Nothing could be done electronically. I believe Bob still has piles of those old data sheets in his Columbia office. Bob did a lot of the thinking and writing, too.
Q: When did you start to realize that the public looked “rational” in the sense you used that word of having stable, coherent preferences that responded to events in reasonable ways?
BP: Well, I have to admit I went into the whole thing thinking that's probably the way things were. That's the V.O. Key influence again, as opposed to Converse. So I wasn't surprised to find the stability. I guess I was a little surprised that it was possible to think in terms of a coherent collective belief system shared among most Americans. But when we looked at those percentages, the prevailing opinions on many specific issues did fit together into a pretty coherent whole. I began thinking about how that could come about. It clearly has something to do with the statistics of aggregation. If you take a thousand people who each have a wisp of opinion surrounded by uncertainty and randomness, and then you measure a thousand opinions that they express at a given moment in time, averaging them will tell you something about the average underlying opinion among them all. That’s no “miracle” of aggregation, it’s simple arithmetic. The central limit theorem.
But it’s not really all just arithmetic. It makes a difference exactly what is being aggregated. To get the results we were seeing, you also need individuals to have some scraps of belief systems, and you need them to do some reasonably sensible information processing. You also need some of what we called “collective deliberation.”
For collective opinion to be real and stable, all that’s required is for a lot of people to have those wisps or tendencies of opinion, and for the central tendencies of each of their wisps to be fairly stable over time. But for collective opinion to be consistent and coherent, you also need lots of people to have wispy belief systems, with some logical or ideological connections among their tendencies of opinion. And you need those proto-belief systems to be similar among individuals, so that they add up to one big collective belief system. Moreover, for collective opinion to react sensibly to events, and for the collective belief system to accurately reflect the underlying needs and values of the citizenry, individuals have to be able to process information reasonably well. And they need to get hold of some good, unbiased information. That depends partly on the whole social process of information production and distribution. The whole “information system,” as we put it. That includes deliberation by elites, information transmission through the media, peoples’ use of cue-givers like friends and family and their favorite politicians, and so forth. A tall order!
I think Bob and I did a pretty good job of describing the felicitous characteristics we found in collective policy preferences. (Though I wish we had presented more tables to summarize our findings and help readers who are impatient with narrative.) But we did not completely work out the distinct conditions that must be met for each separate characteristic of collective preferences to manifest itself. If we had done so, our theory would have been better able to deal explicitly with future changes in the information system. Big changes were coming.
Q: Today the topic of misinformation and manipulation has returned to the limelight. You were mostly skeptical that opinion was manipulated in The Rational Public. Was that less of a worry at the time?
BP: Yes and no. This is an example of incompleteness in our statements of the theory. Much of one chapter of The Rational Public discusses opinion manipulation, which fits into the theory in terms of biased inputs from the information system. That chapter describes several really blatant cases of manipulated opinion in foreign policy. We make the argument that the foreign policy realm is somewhat different, in that the government has a tremendous amount of information control. Much more easily than in domestic policy, officials can bring people along on questionable adventures.
I suppose some people have an image that I've been complacent about American politics during most of my career and that I think democracy's been working perfectly. Actually, that's not the case. Back at that time, democracy probably was in fact working better than it does now. But there's always been some non-responsiveness and some manipulation of opinion. There has been a thread in my work about those phenomena. The foreign policy manipulation cases in The Rational Public. Way back in the party cleavages chapter of Choices and Echoes I tried to make clear that some kind of elites were really affecting the electoral process, probably money givers or party activists.
Q: To what extent do you think it still makes sense, thirty years on, to describe US public opinion as rational?
BP: I would really like to know. Somebody's got to do a new study. Somebody should replicate our data, and it'd be far, far easier now with electronic archiving of poll data.
I can only guess at what the results would be. One guess is that the “real and stable” collective opinion thing is probably still true, because the conditions for it are so easy to meet. I also imagine that the idea of the information environment feeding into collective opinion will remain correct, because there is a lot of political information out there and surely people are processing some of it. But it's a very different information environment. There are a lot of half-truths and outright falsehoods in the media – especially social media, talk radio, and Fox – and quite a few of those falsehoods originate in high places. I would say that’s a real change. Yes, we had old cases of politicians and officials lying about foreign policy, the Tonkin Gulf incident, the so-called “missile gap,” the “bomber gap,” various U.S.-backed coups and invasions. But now information channels are much more fragmented, which makes a single coherent belief system less likely. And some of those channels are conveying outrageous misinformation about both foreign and domestic matters. A whole political party is implicated. It seems likely that opinion is more often manipulated now.
We need to know how many people believe those falsehoods, and to what extent collective policy preferences may no longer always reflect people’s needs and values. With highly polarized parties leading their partisans in very different directions, the “parallel publics” phenomenon has probably changed too.
Q: Another extension that I've thought about a lot is how well the rational public idea applies outside the U.S. I'm surprised that not many people have applied it to other countries. I know of papers on Canada and European foreign policy, and I wrote one on the Czech Republic. But there are not many on other countries. Do you have any thoughts on why?
BP: Don’t know. Perhaps what's thought of as a big question worth studying in the United States doesn't seem like a question in some other places. Elsewhere it may just be assumed that citizens know what they want and need, and that the government ought to be heeding them.
Q: At about the same time that you were working on and publishing The Rational Public, Samuel Popkin was writing about the reasoning voter and James Stimson about the public mood. Were you aware of and in contact with these other projects? What did you think about them?
BP: I knew Sam a bit. I thought of us as kind of doing the same big project. We had both been interested in “floating voters,” the rational electorate, V.O. Key type stuff. So, I liked his work and thought he was doing a good thing.
Jim Stimson's a more complicated case. I was very pleased that he was paying attention to the macro level, to collective public opinion, because most of the public opinion field has always been focused on individual psychology. That's important scientifically, but to me the macro side is more interesting. Stimson and his co-authors have this well worked out, opinion-driven model of the whole political system. Unfortunately, it probably does not get very close to describing real-world politics, for much the same reasons that Bob Shapiro and I got in trouble: the omission of crucial independent variables. But it's a really clever, interesting model. Within the model’s own terms, the chief shortcoming may be that they lump all issues into this one “mood” continuum. That can seem plausible in the United States because with two parties we observe unidimensional voting in Congress. But that's not what’s actually in most ordinary people's minds.
Most people's thinking about public policies is not nearly unidimensional. And there are really interesting differences over time in how collective opinion changes on one issue versus another. All you have to do is mention gay rights. That has had a far different trajectory over time from other social issues, like gun control or abortion, let alone Social Security or aid to education. So, one problem I have with this Stimson line of work is that it scrunches a complex reality into one dimension. The other problem, which I can be very critical about because I had it too, is the spuriousness problem: not taking enough account of the ways in which you could get public opinion connected to policy without public opinion being the cause of the connection. That, to me, is maybe the biggest problem in my past work.
Q: Are you thinking of endogeneity – policy affecting public opinion?
BP: Endogeneity is part of it, but also just plain spuriousness. If there's something that affects both your independent and your dependent variable, then your variables can correlate positively with each other without having any direct causal relationship at all. And if you do a bivariate study, like Bob’s and my responsiveness study, where the only variables are changes in collective public opinion and changes in policy, the model can be very wrong. You're leaving out the possibility, for example, that wealthy people or interest groups are influencing both public opinion and policy making. You’re also leaving out the possibility of endogeneity: that when policy changes, there's a feedback effect and opinion changes to go along. We did try to deal with that, but not very well.
Q: Ultimately, Christopher Wlezien argues that opinion is thermostatic.
BP: Clever theory, and interesting. I believe it's true that the general public does tend to react to what the government's doing. And if current policy looks like it's extreme, they react in a negative direction. In other words, there is such a thing as negative feedback of policy on opinion, and the homeostatic idea gets at that to some degree. But it I think it leaves out the stimulus part of it. You want to know what the government is actually doing. Is it doing something that is right where the median voter wants, or is it doing something that's far left, far right?
In my view of how politics works, the parties have always had some real differences, and public policy has always lurched between one direction and the other depending on which party controls government. And surely there's oversteering a fair amount of the time. So when there is oversteering, people say, Wait, wait. A great example of that is the Reagan defense budget. Through the late 1970s and early 1980s, public opinion favored spending more on defense. Reagan takes office and spends a ton of money on the military. Very quickly, people want to spend less. But that doesn't necessarily mean that there was any change in what ideal level of military spending they favored. What it means is that the status quo changed. Wlezien’s survey questions, I believe, are almost always “more” or “less” type questions, in which the status quo is the baseline. So, if the status quo changes, then survey responses about it change meaning. And I think that's one way to reconcile his findings with ours.
Q: Using your collected data, you and Shapiro published an influential paper on responsiveness that assessed whether changes in preferences were followed by corresponding changes in policy. You found that they were, about two-thirds of the time. This was a quite positive result. How do you look back at that responsiveness result today?
BP: The first point is that our basic congruence finding was correct then and quite possibly remains correct today. There is some tendency for the government to do what the people want, even if it is only “democracy by coincidence” in which affluent elites are making that happen.
But the second point is, we found that glass to be two-thirds full, but also one third empty. So many cases in which policy actually moved in the opposite direction from opinion seems rather alarming. And then there were many cases of no measurable policy change at all, which can hardly be called responsiveness. Moreover, our criterion for congruence between opinion and policy was an easy-to-satisfy ordinal one: movement in the same direction. Not necessarily by the same amount at all.
We made those points in the article. We tried not to oversell the democracy interpretation of it too much. We said that these findings don’t reflect perfect democracy. But our enthusiasm for the results probably outweighed such cautionary language. And the congruence results were descriptively correct. Where we went badly wrong, or at least gave a very wrong impression, was in the title about “effects” of public opinion on policy. That was not good.
Q: You mean the causal language?
BP: The causal language, much too strong. At one point some years later, after Marty Gilens and I had showed that that earlier causal argument was just not correct, Sidney Verba said, “Well, Ben, what are you doing to atone for your sins of the past?” Yeah, that was a mistake. But eventually we did get around to fixing it. Science marches on.
Q: On the policy preferences again, a little later, John Zaller came out with an influential theory of the survey response, the RAS theory. Did you view that as in tension or consistent with your thinking about how public opinion evolves?
BP: I've always assumed that elites play an important part in politics, including through information systems. I guess we’ll get to the deliberation book at some point. And so in principle I don't find it a terrible idea that John would work out a theory of collective public opinion in which elites play a major part in creating it. But I think he underestimated the ability of people to form opinions that reflect their own needs and values. Another problem, I believe, was that that he attributed so much opinion leadership to political parties. At that time I believe there had been rather few cases of this actually happening. Of course he should probably be congratulated for prescience about the future. The highly polarized parties of today may well lead their followers’ opinions more effectively.
I think his favorite case back then was Vietnam, when party control of government changed. it had been Lyndon Johnson's war. And, you know, the Republicans took over, and it became Nixon’s war. There was some switching of sides among partisans.
Q: I may not be remembering correctly, I haven't read it in a while, but I think that he said that when elites divided on Vietnam, then the public could go either way. When elites were united, so was the public.
BP: I agree with that point. Elites dividing and voicing an alternative elite point of view was key for change in opinions on the use of force in the Vietnam War. Walter Cronkite was the big symbol of that. When he turned against the war, that was a major thing.
Q: I think we're getting into this issue of the media deliberation. I’ll come back to your book in a second. But you first wrote a paper on what “moves” or drives public opinion, and that's also part of the rational public book. What did you see as the main drivers of public opinion?
BP: Back then, as you were kind of hinting a minute ago, there were really unified mass media. There was the AP wire, which went to all the local newspapers. There were the three television networks. Studies showed that they were all saying pretty nearly the same thing. So you could have a very simple model of what influences public opinion. At that time, you could just study what was on TV news and separate the messages by their sources – that turns out to be really crucial – and then predict opinion changes. We had these wonderful time series on public opinion. So Bob Shapiro and Glenn Dempsey and I turned the policy responsiveness study around, took a few of our shortest time series, and basically looked at what happened after the first poll that might affect opinion in the second poll.
And it turned out that supposedly nonpartisan experts and commentators were hugely important influencers. Presidents probably came in, in a subtle way that I'm not sure we ever conveyed well. But presidents – at least popular presidents – almost certainly influenced a whole bunch of actors who in turn influenced opinion. But the direct effects of the “bully pulpit” were not big.
Q: But the media was.
BP: The media were essential. I mean, as the Latin (plural) name says, they were the means of communication. They were transmitting these elite views, which changed public opinion. My rational-public view of that was that collectively, people were processing the news pretty well, given what they got. The big problems were and are with the information system, not the public.
Q: You mentioned the book about deliberation that you published in 1996, Who Deliberates, where you considered how the media treated several policy issues: the first Gulf War, the Rodney King riots, and the Zoe Baird case. What motivated that book?
BP: Well, there I spent a lot of time in the Northwestern library basement, reading microfilms. That was a different kind of research than I'd ever done before. And it was really fun to do.
But what motivated it intellectually was the puzzle that we were just talking about. You could call it the Zaller puzzle: where does public opinion come from? What part do elites play? Bob Shapiro and I had to have some view of that in order to talk about a rational public. We had our thing about media effects, but we hadn’t yet fit that well into a whole political system that affects what the commentators and experts and presidents are saying. They're part of some kind of system: who gets to be an influencer? what influences them? and so forth. As a small step along the way toward that, I wanted to work out the mechanisms by which elites deliberate among themselves and reach policy conclusions, which then get transmitted to what I think of as the rational public through these more or less neutral communicators.
The idea was that in various different ways, talk radio was one channel, mainstream newspapers were another, letters to the editors of the New York Times were still another (a minor but easily studied one.) There are several channels through which elites can argue about stuff, and there are cases in which that argument goes quickly enough that you can actually watch people get convinced and come to a conclusion. One conclusion elites came to was that poor Zoe Baird couldn't be Attorney General because she had, gasp, not paid taxes for a nanny. I thought that was rather stupid and sexist (not that I loved Baird), but it was still an interesting deliberation case. On the New York Times letters to editors, I had a fascinating conversation with the Letters editor at the Times. Back then, they already got about 300 submissions a day! By now, who knows. I imagine it's in the thousands. The editor told me a little about how they picked letters to publish. And it was it was fun to read those letters and watch how ideas worked their way through them over time, in that case more as an indicator of what various elites were thinking than as an actual site for deliberation by them.
Q: In the book, you find some evidence of media slant, but you also argue that the marketplace of ideas seems to be working, that there is deliberation. Is that a correct interpretation?
BP: I suppose that’s mostly right about that book. Except for the grossly U.S.-centric slant of letters to the Times about Iraq, I didn’t find much media slant. But I did some work more directly on slant and bias around that same time. Bob Entman and I studied the broader media and the first Gulf War, finding a big pro-U.S. bias in nearly all the U.S. media. Later I worked with Jim Snider on media owners’ control over what their outlets say. And I have always thought that slant is a problem. Again, especially in foreign policy, but sometimes in other realms too. But I think it's fair to say yes, the Deliberation book mostly made the argument that a marketplace of ideas was working fairly well, in the sense that the public was getting the results of reasonably serious elite deliberations. At that time, I think that was probably true.
Q: At the conclusion of the book, you call for more media diversity, so that there would be more deliberation. Would you say that your call has been answered but not with positive results?
BP: Oh, Lord! Well, it's seems likely that media diversity is a necessary condition for a well-informed public. But it's certainly not a sufficient condition, if the diversity includes a whole lot of misinformation and lies and propaganda. I think that's a horrible feature of our present political reality. A big question is how much impact that has on collective public opinion. To what degree do many or most Americans now believe things that are nonsense, not true? We know that’s partly a partisan matter, that there are an awful lot of Republicans who still believe the 2020 election was stolen, and so forth. When the leaders of a whole major political party are saying things that aren't true, that seems likely to upset the marketplace of ideas.
Textbooks
Q: Along the way you wrote a couple of textbooks. The first was on the presidency. What was your intention with that textbook, entitled The American Presidency?
BP: One aim of the presidency text with Mark Petracca was just to summarize what was known about the presidency. Mark was super at gathering that together, an omnivorous reader. But the thing that made our text a little different was that we – influenced by David Easton – tried to locate the president within a political system in which his actions are dependent variables as well as independent variables. Getting away from superficial interpretations of Neustadt or Barber in which presidents look like unmoved movers. Why are presidents doing what they do? What influences what they want and what they attempt to do?
Q: You also wrote another textbook, an American politics textbook.
BP: That must have started in the middle 1980s because we were in Austin. Ed Greenberg had the idea for The Struggle for Democracy and asked me to join in. We were partly driven by normative concerns of the same sort that show up in my other work, in favor of a majoritarian kind of democracy. Ed and I wrote about the obstacles to winning democracy in the U.S., and the historical struggles, and so forth. This was our frame for a standard American politics textbook that did all the usual things too. We aimed for a contrast with the other texts, most of which seemed complacent or dull or both. The Struggle for Democracy went through many editions, with Ed taking over and doing all the work. Some forty years later it still keeps on selling copies.
Part 2 of this interview can be found here.