In a research paper, we compared their responses with our best guesses of what the public in their districts or states actually wanted using large-scale public opinion surveys and standard models. Across the board, we found that congressional aides are wildly inaccurate in their perceptions of their constituents’ opinions and preferences.This has to be taken with a grain of salt. All good politicians know that what people tell pollsters about issues and how they vote are two different things. A good example is that most Republicans will tell pollsters they care about the deficit, even though hardly any have voted as if they did since Eisenhower's time. Likewise many democrats support integration as long as it it doesn't involve their own kids' school. So this could be a case of politicians understanding the complexity of the links between policy choices and votes better than pollsters do.
For instance, if we took a group of people who reflected the makeup of America and asked them whether they supported background checks for gun sales, nine out of 10 would say yes. But congressional aides guessed as few as one in 10 citizens in their district or state favored the policy. Shockingly, 92 percent of the staff members we surveyed underestimated support in their district or state for background checks, including all Republican aides and over 85 percent of Democratic aides. . . .
Aides usually assumed that the public agreed with their own policy views.
But my dealings with politicians have impressed on me that they are not analytical people. They are people people who thrive on fact-to-face interactions. The opinions that matter to them are the ones their friends and supporters express when they meet. Many of them seem to have serious attention deficit issues when it comes to policy, especially reading about policy. Some are obsessive readers of polls, but not all are; as you may have noticed some of them regularly denounce polls as silly and say the size of their rallies or their mail bags are better signs of what the public really thinks. I don't think they are all just trying to change the subject; I think many of them really form their view of "the public" much more from what they see and feel than what they read.