Consultants and staff hear the refrain regularly from members: “I know my district/state. I know what they think.” But do they really?
It’s hardly a new question. In 1963, University of Michigan survey research pioneers Warren Miller and Donald Stokes published a seminal paper arguing, in part, that members of Congress have “very imperfect information about the issue preferences of” their constituencies.
The two political scientists had asked members of Congress about their perceptions of district opinion in three policy domains and measured public opinion in their districts on the same questions. They found some correlation between members’ perceptions and actual opinion on civil rights issues, but almost none on foreign policy or on economic and social welfare issues.
Miller and Stokes’s methodology was subject to critiques that called into question, but never disproved, their conclusions. Recent research by David Broockman of Stanford and Christopher Skovron of Michigan has put the ambiguity and uncertainty to rest.
They collected data from nearly 2,000 state legislators and legislative candidates probing their perceptions of district opinion on three issues. Then, using a survey of 100,000 and an array of new statistical techniques, they compared actual district opinions to legislators’ beliefs about their districts’ views. Here, the issues were same-sex marriage, universal healthcare and federal welfare programs.
On average, legislators were more than 10 percentage points off the mark in estimating what percentage of their district held which opinion.
Districts have to support “a universal healthcare program to guarantee coverage to all Americas, regardless of income” by 2-to-1 before legislators start to believe a majority of their constituents would back that proposal. On same-sex marriage, legislators’ perceptions reflected what district opinion was probably like a decade ago.
Equally important, the only factor that seemed to predict inaccuracy was ideology. Politicians of every stripe overestimated the conservativism of their districts on these issues, but conservative politicians were much more likely to do so. Indeed, on average, conservative legislators estimated their constituents were some 20 points more conservative on these issues than they actually were.
Put differently, more than half of conservatives believe they represent what would be in fact one of the most conservative districts in the entire country.
For example, only 13 percent of Americans favor “abolish[ing] all federal welfare programs.” Liberal politicians tend to believe that over 25 percent of their constituents agree with position, while conservatives believe 40 percent take that view. This leaves conservatives exaggerating the right-wing bent of their constituents by about 25 percentage points.
Politicians are free to ignore public opinion — and some consider it a virtue to do so — but they frequently misperceive it, misjudging the views of their own constituents in meaningful ways. They cannot help but distort the decision-making process in legislative bodies.
The fact that legislators on the right make consistently bigger errors in assuming the conservatism of their constituents may account for some of the dysfunction evident in today’s GOP.
They may actually believe their districts want them to defund Planned Parenthood, or shutter the federal government or allow some gun buyers to evade background checks — even though the data clearly indicate their constituents disagree with them.
The rhetoric of members on the right self-righteously bristles with references to what they “were sent here to do.” In fact, they may well completely misunderstand what voters sent them to D.C. to do.
Of course, the big lesson is, if you really want to know what voters think, trust your pollsters’ data, not your gut.
Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the minority leader of the Senate and the Democratic whip in the House.