Wonkblog at the Washington Post has a fascinating post on a study which sought to show politician's estimates of conservatism in their district vs. actual measures of those attitudes.
What they found was that both Democratic and Republican politicians thought their district was more conservative than it is, Republicans estimating it more conservative by about 20 points.
The article is well worth reading, but what I want to do is launch into wild speculation about why.
The first thought, and probably the most accurate, is that politicians probably judge constituent interest by letters and phone calls they receive as well as volunteers and political donations. Since people engaging in these activities tend to be both older and richer, characteristics that skew conservative, this is likely to leave politicians misinformed about the true interests of their conservative.
A second issue is since the 1980s a belief has set in that America is a conservative country. While this perception is widespread self-identified conservatism doesn't match up conservatism measured by other metrics. Politicians are probably hearing from people identifying themselves as conservative, but who claim to diverge on just one issue, and feeling that these individuals are conservative on other issues as well, when they likely aren't.
A third issue is that many people just don't understand how each side is identifying with issues. The most popular example of this is the "get the government hands of my Medicare!" meme. People claim to hate entitlement programs while expressing support for Medicare, Social Security, Medicaid, and food stamps. Propaganda has led them to believe that someone, somewhere is getting an unfair deal and living high at the taxpayer's expense but they like and support the actual programs they have experience with, not realizing these are the programs meant when entitlements are talked about by the media; creating a further disconnect between politician's perception of their views and their actual views when surveyed.
A fourth reason is that movement conservatism is very well organized and has a long history of grass roots organization. This means that they tend to make themselves more visible and concerted campaigns make their numbers look larger and more popular than they actual are. Such are the benefits of political mobilization.
And a last reason, it seems there are more very loud people on the far right. While it adds little to the conversation people tend to remember those that shout loudest and make a scene. If these individuals are a bit more common on the right this would tend to shift perceptions as well.
As a closing note, I'd like to point out that this working paper is a great example of why political science funding is so important. It hints that even the most conscientious politician concerned with accurately representing his constituent's interests may lack the tools and training to assess what those interests may be. An obvious follow on to this study would be to find ways for politicians to more accurately assess their constituents positions on issues to make our government more functional and our citizen's lives better. While a skeptic may note than many politicians may care little beyond campaign contributions we would do better even if only a handful of politicians are motivated by their constituent interests.