President Trump spoke last year during an event on the South Lawn of the White House to acknowledge the final passage of tax overhaul legislation by Congress. (AP Photo/Evan Vucci)

Senior staffers in congressional offices hold highly inaccurate assumptions about what voters in their districts actually want when it comes to policy. They tend to believe that voters support much more conservative policies than they actually do. And this stunning misperception can largely be explained by the disproportionate attention lawmakers and their aides lavish on donors and special interest groups.

Those are the results of a new paper forthcoming in the American Political Science Review by Alexander Hertel-Fernandez of Columbia University and Matto Mildenberger and Leah C. Stokes of The University of California, Santa Barbara.

While voters have long believed that members of Congress are focused too much on the needs of special interests and out of touch with the general public, the paper provides some of the most convincing evidence to date that these perceptions are largely accurate.

In August 2016, the authors sent out a survey to the chiefs of staff and legislative directors of every House and Senate office. They targeted these senior officials because they’re largely responsible for setting an office’s legislative agenda: They play “a crucial role in the policymaking process, connecting the preferences of constituents with Members of Congress,” as the authors put it.

One section of the survey asked staffers to estimate public support in their own districts (or states, in the case of Senate aides) for five policy proposals: repealing Obamacare, raising the federal minimum wage to $12 an hour, enacting universal gun background checks, regulating carbon dioxide as a pollutant, and investing $305 billion in infrastructure improvements over a five-year period. One could reasonably expect that a congressional aide would understand constituents' opinions on these issues, given that Congress had voted on bills related to these issues in the previous year.

The authors compared the aides' responses to actual district-level public opinion on the issues, as measured by large national surveys like the Cooperative Congressional Election Study. The results showed a stark disconnect.

Aides' estimates of public support for the proposals were way off-base relative to the actual numbers. “In none of the five areas are staffers estimating their constituents’ preferences with any degree of relative or absolute accuracy,” the authors write. Legislative aides in Republican offices were particularly bad at estimating their constituents' support for various policies.

On the question of gun background checks, for instance, staffers in Republican offices underestimated public support by 49 percentage points. Democrats underestimated too, by 11 percentage points. Republican staffers also underestimated support for carbon dioxide regulation, infrastructure spending and minimum wage hikes by 20 percentage points or more. Democrats underestimated support for those policies by between five and nine percentage points.

There was one issue, however, where Democratic staffers were more off-base: Obamacare repeal. Democrats underestimated support for repealing the Affordable Care Act by 24 percentage points, while Republicans overestimated support by 10 percentage points.

Overall, the researchers found, aides in Democratic offices better understood public opinion in their districts by a margin of about 13 percentage points. But with the exception of Obamacare repeal, Democratic aides sided with their Republican colleagues in assuming their constituents were more conservative on the issues than they actually were.

“Overall, we find a conservative bias in staffers’ estimations,” the authors write.

What’s driving these legislative misperceptions? The paper offers some clues. The authors checked to see whether aides in more politically competitive districts did a better job guessing their constituents' preferences than aides in safer districts — they didn’t. Ditto for staffers' years of experience.

They did find, however, that, like all of us, congressional aides tend to believe that everyone else thinks like they do. If a given aide supported a policy, they tended to think that their constituents did too, regardless of whether that was accurate.

But perhaps the most significant factor the researchers identified was the role of interest groups. The survey questions yielded a number of key findings.

First, aides who reported more reliance for policy-making on business-oriented interest groups, like the U.S. Chamber of Commerce or the American Petroleum Institute, had a poorer understanding of constituent preferences than did aides who relied more on groups drawing their membership from the general population, like the Sierra Club or the League of Conservation Voters. Of note, those business-oriented groups tend to support conservative policies, which could explain some of the aides' conservative bias in their estimates.

Second, the researchers found that aides in offices receiving more money from corporate interests did a worse job of estimating constituent preferences. In fact, “45% of senior legislative staffers report having changed their opinion about legislation after a group gave their Member a campaign contribution,” according to the paper.

Money talks, in other words, and congressional staffers are listening — even if it means becoming less responsive to the needs of their voters.

Perhaps the most alarming finding in the survey, however, was that “staffers are more likely to interpret correspondence from businesses as being more representative of their constituents’ preferences than correspondence from ordinary constituents." The survey asked aides to imagine receiving letters about a policy issue from either “employees of a large company” in their districts or “constituents" and to consider how much those letters represented public opinion in their districts. Sixty-two percent of staffers said they’d view the employees' letters as representative of public opinion, versus only 34 percent who said the same of letters from ordinary citizens.

All told, the study paints a picture of a Congress that is out of touch with the American people -- with perceptions of public opinion skewed rightward by the influence of deep-pocketed lobbyists. The authors say that the best way to combat these distortions is by increased civic participation among the general public.

“Political action can’t end on Election Day,” co-author Matto Mildenberger wrote in a tweet on Thursday. “Citizens need to keep writing, calling and meeting with elected officials and their staffs long after the midterms.”