Asymmetric Misperceptions and the Conservative Construction of ‘Public Opinion’

There’s been a lot of buzz about an excellent (but not yet peer-reviewed) working paper by David Broockman and Chris Skovron, “What Politicians Believe About Their Constituents: Asymmetric Misperceptions and Prospects for Constituency Control,” which looks at state legislative candidates’ perceptions of their constituents’ opinions.  The findings are striking, but unlike many others, I don’t find them all that surprising:

Actual district opinion explains only a modest share of the variation in politicians’ perceptions of their districts’ views. Moreover, there is a striking conservative bias in politicians’ perceptions, particularly among conservatives: conservative politicians systematically believe their constituents are more conservative than they actually are by over 20 percentage points, while liberal politicians also typically overestimate their constituents’ conservatism by several percentage points.

One of the things that is interesting about this paper is the data analysis itself.

We first investigate the general relationship between elite perception of public opinion and actual public opinion and show that it is remarkably linear on average – an increase in support for same-sex marriage of 10 percentage points in a district is associated with an equally-sized 10 percentage point increase in politicians’ perceptions of their districts’ opinions on average. The same holds true for districts’ support for universal healthcare and politicians’ average perceptions of this support. Yet although politicians’ perceptions do vary directly with their constituencies’ actual views on average, the correlations between public opinion and politicians’ perceptions of it are at best modest – 0.43 for universal health care and 0.51 for same-sex marriage. These relationships are far stronger than those reported by Miller and Stokes (1963) but still suggest that lion’s share of the variance in politicians’ beliefs about their constituents’ cannot be explained by the reality of their constituents’ views.2

What is most notable about the relationship between actual public opinion and elite perceptions of this opinion is not its strength nor its average slope, however, but its intercept (e.g Achen 1978). Politicians consistently and substantially overestimate support for conservative positions among their constituents on these issues.

So it would be easy to simply look for the relationship between “actual public opinion” (more on that term later) and elite perceptions, find it, and stop there. Interestingly, this isn’t about complex techniques, it’s simple a matter of looking carefully at the data. In fact, I think this problem helps explain the dispute over how well public opinion is reflected in policy.  If you are looking for a relationship, you will surely find it. But if you go issue area by issue area and ask if there is a disconnect, you will find that too. The existence of the relationship is what social scientists are trained to look for, but it can be highly misleading if it is not put in context.

This leaves us with two questions. First, what is driving this?  A natural answer is that the donor class, or elites more generally, are more conservative, and politicians are more attuned to them than the general public. I generally think this is true, and that it matters a great deal, but it can’t fully explain the results, since it asked not just about “economic” issues, but also about “social” issues, specifically marriage equality.

Interestingly, there has been little attention in this discussion to the mechanism whereby candidates would learn of opinion. Remember we’re talking about a situation where polls of the district generally don’t exist. Part of the project involved estimating public opinion at the district level. Broockman and Skovron mention elections as having a possible educative effect, while noting this does not happen. Elections are pretty blunt instruments for learning about issue specific opinion anyway. They also look at whether candidates who spend more time interacting with others might have a more accurate perception of district opinion. But again, this doesn’t make a difference. It  seems likely that candidates interact mostly with those who are supporters, which wouldn’t necessarily help.

Broockman and Skovron borrow this chart from a classic article by Miller and Stokes to map out avenues of constituency control.  What process is supposed to be happening between the constituency’s attitude and the legislator’s perception of those attitudes?  If you don’t stop to think about it, this may seem unproblematic. But if you do, it’s unclear.

That said, there have been studies of this question. Susan Herbst’s Reading Public Opinion: How Political Actors View the Democratic Process used interviews to examine how three sets of state-level actors–legislative staffers, political activists and journalists–sought to understand public opinion.  She found that the staffers largely relied in the media and interest groups in constructing ‘public opinion’.  In fact, these three groups held contrasting ideas about what public opinion was and how to decipher it, largely for pragmatic reasons. That is, it depended on the nature of their jobs. Knowing what majorities of people believe isn’t terribly useful for these actors. As summarized by Lance Bennett, “these elites all share one startling perspective: none regards publics as broad aggregations of citizen viewpoints that should be counted routinely in their political or journalistic through political situations.” More important, for example, is knowing whether those who follow legislative events closely might turn something into a public controversy, and how that might be received by those who pay attention. To the extent that elites form ideas about what the public thinks more broadly, it is likely happening incidentally.

Broockman and Skovron take a more orthodox approach, equating “actual public opinion” with survey results. But for Herbst, it is more complicated.  For me,  ‘public opinion’ is a claim, a justification used to support or undermine policies or persons. ‘Public opinion’ will be used to justify all sorts of decisions in a democracy, in the same way that ‘god’s will’ is used to justify things in a theocracy.

(It’s worth noting that politicians at the national level do have access to polls and pay a great deal of attention to them. But the question is for what purpose.  It’s still unlikely that they care about simply knowing ‘what the public thinks’.)

So the answer to why this misperception is happening likely can be found in the organizational advantage held by conservatives, and by higher levels of mobilization in service of conservative causes.  Or at least, that is where we should look first. Anyone interested in these issues should read Herbst’s fine book. The larger point is that we’re better off looking at what political actors actually do rather than imposing an abstract, formal view of democracy on them and asking how well they measure up. The claim of public opinion is the flip side of the claim of representation, not something that should be taken as fact.

The second question is: what to do about it?  This isn’t a matter of simply telling legislators that they are wrong, or providing them with better information.  Liberal mobilization, greater organization at the state level, more persistent demand making–these are the things that should ensure legislators get a more accurate picture of public opinion, and more importantly produce policy that is less disconnected from the public.