Every once in a while you trip across something to which you can only react by saying aha, that is something I completely didn't know and it sure explains a lot.

Over at Democracy, the journal I edit in addition to writing for the Beast, we have a little blog, and it's mostly about policy and social-science research, in keeping with the journal's mission, so we often fill the blog with papers from the outstanding Scholars' Strategy Network, a kind of umbrella organization/clearing house for the best and most interesting social science research being done in the country. If you are interested in substance, on virtually any topic under the political sun by America's leading academics, you should be visiting SSN regularly.

Now. Here's something we posted last night, from David Broockman and Christopher Skovron. They thought of asking a question I've never seen anyone think to ask before. Last year, they asked more than 2,000 state legislative candidates from around the country what they thought the political leanings of their constituents were. Specifically, they asked the candidates to estimate what percentage of the voters in the districts where they were seeking office supported: same-sex marriage; a government-run universal health-care program; the abolition of all federal welfare programs. Then they matched those to existing polling.

Answer? From the authors:

When we compare what legislators believe their constituents want to their constituents’ actual views, we discover that politicians hold remarkably inaccurate perceptions. Pick an American state legislator at random, and chances are that he or she will have massive misperceptions about district views on big-ticket issues, typically missing the mark by 15 percentage points.

What is more, the mistakes legislators make tend to fall in one direction, giving U.S. politics a rightward tilt compared to what most voters say they want.

Not surprising, in a way. But startling. The typical conservative candidate in their survey overestimated the district's conservatism by 20 points. The typical liberal candidate overestimated the conservatism by around 5 percentage points.

The authors didn't really get into why candidates have these perceptions beyond saying that politically active citizens tend to be older and more conservative, but I think it's pretty obvious that a whole set of factors in most places creates this misperception. Conservatives are often more vocal. Liberals, especially outside cities and university towns, are probably a little cowed. In most places the local social establishment that dictates the agenda will tilt right. The local newspaper is probably conservative. Rush Limbaugh and all those other fat mouths are on the radio. There's often a local public-affairs radio host who is just about as conservative.

So here's another organization some rich liberal ought to fund: a group that tells legislators what their constituents actually think, and that organizes a district's progressives to make sure the accurate perception is reinforced. Legislators react to perceived pressure, and if they think all that pressure is from the right, that's how they're going to behave.