In the early-morning hours of July 28, 2017, Republican Sen. John McCain of Arizona crossed the congressional aisle for a now-infamous huddle with Democratic lawmakers, the prelude to casting the decisive “no” vote on the Senate’s “skinny repeal” of Obamacare.

Political analysts tend to use the term “aisle-crossing” in symbolic terms, a metaphor for lawmakers' willingness to cooperate with the opposing party to get things done. But a new working paper by Bryce J. Dietrich of the University of Iowa finds that in strictly literal terms, members of Congress today are much less likely to physically cross the aisle of the chamber to fraternize with members of the opposing party than they were even 10 or 15 years ago.

To quantify lawmakers' aisle-crossing behavior, Dietrich combed through more than 1,400 hours of C-SPAN footage covering the years from 1997 to 2012. He calls it “the largest collection of C-SPAN videos ever compiled."

Dietrich used a computer algorithm to focus on what he calls the “ant farm” shots of the House Representatives — the high-angle shots showing lawmakers milling about the House floor during votes. He then had the computer key in on the telltale signs of motion that occur when a lawmaker physically crosses the aisle to mingle with members of the opposing party.

He finds that Republicans and Democrats grew less likely to cross aisles and speak with one another during the period he analyzed. “As Democrats and Republicans become increasingly divided,” Dietrich writes, “they have more incentive to display party discipline which translates to an unwillingness to physically cross the aisle.”

Lack of aisle-crossing was particularly prevalent both in the run-up to party-line votes, as well as in the aftermath of those votes. Polarization, in other words, isn’t just an ideological abstraction — it’s affecting how lawmakers move and operate in the physical world.

On some level that physical separation reflects what political scientists are observing all across the country: people increasingly sorting themselves into like-minded enclaves. Americans today are more likely to live and work with others who share their political beliefs than they were several decades ago.

“Just as Americans seem to be moving away from those who share opposing viewpoints,” Dietrich writes, members of Congress "seem less willing to expend the effort necessary to walk across the aisle, let alone work out their differences in order to find common ground.”

Dietrich’s measure provides a quantified confirmation of something that members of Congress have anecdotally noted: social interactions between lawmakers, once commonplace, are on the decline. It also tracks closely with a separate widely used measure of polarization based on how often members vote together.

Would more aisle-crossing and bipartisan fraternization lead to a less polarized Congress? Some observers believe it would, but Dietrich is doubtful: he writes that the social freeze between the parties is most likely a consequence of ideological sorting and polarization, rather than a cause of it.

At any rate, there’s nothing in the data to suggest the trend will reverse itself in coming years. Voting data show members of Congress continuing to drift away from each other ideologically, with survey data showing a similar split widening between voters.

Dietrich hopes, however, that other researchers in political science continue looking toward non-traditional metrics, like data derived from video recordings, to increase their understanding of the forces at work within Congress.