For the first 13 decades of its history, the U.S. House of Representatives was an ever-expanding institution. From 65 members in its vintage 1789 configuration, the lower chamber grew steadily with each new census count, accommodating the growing population of the country.

But a bigger House also meant a more unwieldy House. And so in 1911, Congress somewhat arbitrarily decided that 435 was enough already and set the number down in a statute. The House had gotten as big as it was going to be.

And so it has been ever since, even as the country has more than tripled in size. The average U.S. congressional district now contains roughly 640,000 citizens, as opposed to about 200,000 in 1911.

With so many people to keep up with, do relationships between the representatives and their constituents inevitably begin to fray? And if so, does this mean that the so-called "people's house" isn't really living up to its name anymore?

Brian Frederick, a professor of political science at Bridgewater State College in Massachusetts, thinks that things are heading that way. His research shows that as districts get bigger in population, constituents are less likely to report that they had contact with their member of Congress, less likely to think their member would be helpful, and less likely to favorably evaluate their member of Congress (and more likely to see their member as out of touch with the district).

These findings confirm what most theory on representation already suggests: Members from larger districts should have a harder time connecting with and thus representing their constituents. But until Frederick combined National Election Survey data with district and member characteristics, there was no solid empirical evidence to back it up.

Based on such findings, Frederick said he is now convinced that increasing the size of the U.S. House would, on balance, be a good idea. "It would make it easier for members to serve fewer constituents and a more homogenous constituency," he said. "It would allow for smaller geographic areas. There is something to be said for that kind of connection between members and their constituents."

It would also likely lead to better minority representation in Congress, Frederick noted, because it would create more majority-minority districts. (The representation of minorities lags behind their percentage of the general population. The U.S. population is 12.8 percent black and 14.4 percent Latino, but 9.4 percent and 5.1 percent in the U.S. House, respectively.)

But if the U.S. House were to add seats, how many should it have? Arend Lijphart, an emeritus professor of political science at the University of California, San Diego, and a comparative scholar of democratic institutions, has argued for 650 seats.

That figure is based on the so-called "cube root law" of Rein Taagepera, who figured out that taking the cube root of a nation's population provided a remarkably good predictor of the size of that nation's lower house. By that logic, the U.S. was an outlier on the low side, with a House of 435 instead of the 669 that would now be expected given the U.S. population of 300 million. (Lijphart made his 650-seat recommendation in 1998, when the U.S. population was at 275 million.)