Data from previous debates has come from polling after the fact or from small groups of voters using hand dials from the Nielsen Company. For example, Nielsen brought together 44 voters in 2008 to watch the debate between then-vice presidential candidates Joe Biden and Sarah Palin.

The students, lured by the promise of extra credit, will tap one of four reactions to what’s said: agree, disagree, spin or dodge. Beforehand, they’ll answer some basic demographic questions. They’ll be asked more when the 90-minute debate ends.

Researchers and participants both will be able to watch the responses in real time.

Boydstun said that if all goes as planned, the app will combine “the best of all worlds” — moment-by-moment responses, like those like-dislike Nielsen dials, the sense of engagement offered by Twitter and the demographic data of traditional polling.

The researchers also plan to follow up with questions until Election Day, to see if a change or boost in opinion brought about by a debate fades or strengthens with time.

They will have some basic data to share after the debate, like how participants respond to the question of who performed better on sliding scale. In short order, bar charts also should be available for participating classes that show the relative number of agrees and disagrees by different demographic groups.

The researchers plan to drill deeper in the weeks and months ahead.

They’re especially interested in how the audience responds when Obama and Romney use two often effective tools of political rhetoric: agenda-setting and issue-framing.

Agenda-setting is focusing attention to a particular issue, like steering the debate toward the economy instead of other topics. Issue-framing turns the spotlight on one aspect of an issue, like the cost of the war in Afghanistan, at the expense of other dimensions of that topic, like the safety of soldiers.

The opinions of college students may not be representative of voters as a whole, but the sheer numbers available — thanks to the help of 200 or so professors and instructors at 125 or more colleges — will provide lots of numbers to crunch.

Some colleges have invited community members to take part and others have larger numbers of older, nontraditional students. Plus, “We have enough classes from typically conservative states that I think we’ll have a pretty nice balance in terms of partisanship,” Boydstun said.

That’s not always easy, she said. Republicans are in short supply in Davis; political scientists at the University of Arkansas, meanwhile, struggle to find Democrats.

At UCD, about 400 undergraduate political science students have been invited to use the app while watching one of the three presidential debates or the vice presidential debate.

For the first face-off between Romney and Obama, about 100 law and business students will use the app during a watch party at King Hall. Watch parties with students using the app are planned for five other campuses across the country.

The researchers also will have two other pools of data.

The University of Denver, which is hosting tonight’s debate, plans an outdoor watch party at which upwards of 5,000 people will be invited to use the app. The researchers are attempting to recruit groups of more religious voters, too.

The year-old app project is also the work of Rebecca Glazier, an assistant professor of political science at the University of Arkansas, Little Rock; Matthew Pietryka, a lecturer and postdoctoral fellow in political science at UCD; and Philip Resnik, a professor of linguistics and computer science at the University of Maryland.

Resnik assembled the tech team that built from scratch the React Labs app, which Boydstun said also could be used to gauge the real-time response to other sorts of events — be it an Academy Awards ceremony, a State of the Union address or a football game. Resnik also has formed a startup company to commercialize the app.

Because of limited capacity, the app isn’t yet available for download by the general public.

Boydstun and her colleagues have been gearing up for the debate by testing their system. She has another worry, however.

“What if we went through all of this and find that debates don’t matter? This is an ongoing question. So far the answer is, ‘It depends.’ ”

History shows that debates don’t often boost a campaign — but a bad performance can sink one.

“I’m selfishly hoping we get to see some reaction to something unusual,” Boydstun said.

Wednesday’s presidential debate on domestic issues will air live on major networks and cable news stations and be streamed online at 6 p.m.

The remaining schedule: Oct. 11, vice presidential debate from Centre College in Danville, Ky.; Oct. 16, presidential candidate town hall on domestic and foreign policy from Hofstra University in Hempstead, N.Y.; Oct. 22, presidential debate on foreign policy from Lynn University in Boca Raton, Fla.

— Reach Cory Golden at [email protected] or 530-747-8046. Follow him on Twitter at @cory_golden