Do members of Congress shortchange their constituents in favor of donors? That accusation was flung in one congressional race recently — but it’s a question worth asking of districts nationwide. About a quarter of the members of the House of Representatives raise a large share of their campaign funds from donors outside of their districts. And my research suggests that the more they depend on those donors, the less they represent the voters in their districts.

Could this be one reason why Congress is so polarized?

Every election cycle, nearly 100 House members raise at least a third of their campaign funds from outside the district

Here’s the allegation. Last month, Rep. Scott Garrett (R-N.J.), who represents the 5th District, accused his opponent, Democrat Josh Gottheimer, of trying to “buy this congressional seat” with money from outside the district. According to reporting by the Bergen Record, only 20 percent of Gottheimer’s contributions had come from inside the district. The accusation implied that Gottheimer would pay more attention to out-of-the-district donors than to his constituents. It also implies that the focus of his fundraising is atypical.

But Gottheimer isn’t alone. A large percent of the money Garrett has raised has also come from outside the district — almost as much as Gottheimer’s. Between 2006 and 2012, about 100 House members per election cycle raised a third or more of their total campaign funds from people who live outside their districts. The average member of the House received just 11 percent of all campaign funds from donors inside the district.

Did those contributions change the representatives’ votes?

So how exactly do campaign contributions affect representation? It may be useful to think of representation as a Venn diagram. The ideology of the House member makes up one circle, and the ideology of constituents makes up the other circle. When a member’s ideology perfectly overlaps with that of her constituents, the two circles are exactly aligned. But the more a member deviates from constituent opinion, the less overlap between the two circles there is.

To examine whether that overlap is affected by out-of-district fundraising, I used data from the Cooperative Congressional Election Study from 2006, 2008 and 2010 to calculate an average ideology score for each U.S. House district. I did this by using district residents’ responses to a question asking them to place themselves on an ideology scale. This tells us how liberal or conservative a district is on average.

I then compared that score with the corresponding House member’s NOMINATE score, which uses the member’s entire roll-call voting record to estimate a member’s relative conservatism or liberalism. I compared each member’s score with the district’s score, estimating the overlap of the two representational circles.

Then, using Federal Election Commission data, I examined whether the degree of overlap between constituent and member ideology was affected by how much money the member of Congress received from donors outside the district.

House members follow their donors, not their districts

The more money a House member gets from people outside the district, the less reflective the member’s ideology is of his or her constituents’ ideology. House members are following the money, not their voters.

Interestingly, members who receive the most money from donors outside their home districts also tend to be more ideologically extreme than the rest of their party’s members of Congress. That might be because they represent very liberal or very conservative districts — but not all do.

The members who depend on outside money and are the most out of alignment with their districts represent moderate districts. If the member is among the top 25 percent of outside fundraisers in Congress — which means raising $353,700 or more in outside funds — that member of Congress is representing the party’s ideological wing rather than his or her moderate district.

Would donors from inside the district counteract this effect? No.

I examined the relationship between the total amount of in-district contributions and member ideology. Paradoxically, constituent contributions do not bring members closer to their districts.

That’s probably because in-district donors have more in common with other donors — wherever they are — than with their fellow constituents. The CCES data reveal that, on average, in-district donors’ ideology is closer to outside donors’ ideology than to the average ideology of voters in each party.

And political science research in general has revealed that people who donate to political campaigns are more ideologically extreme than most Americans, and hold different policy preferences.

The more money a member of Congress gets from donors outside the district, the less that member represents his or her constituents’ preferences. And all that outside funding may be leading to a more polarized Congress, as it appears to encourage members to pay attention to donors whose ideologies are more extreme than voters’.

Anne Baker is an assistant professor of political science at Santa Clara University who studies money in U.S. politics.