For all the drama that might unfold in Thursday night's debate, history tells us that it's unlikely to have a huge impact on the polls — though it could matter for a candidate or two.

The question of whether debates really matter in presidential politics has been examined before (University of Missouri, Washington Monthly Magazine, The Washington Post, FiveThirtyEight). The general consensus is that their effect on polls is, if any, minimal — especially late in the general election, when polarized voters are largely decided and hard to move. That's not necessarily the case during primary season, though, when  debates can serve as an introduction for lesser-known candidates.

We wanted to put all these ideas to the test to try to gauge whether Thursday night's debate could have any significant polling impact on an incredibly fractured and broadly unknown Republican field.

Our analysis looks back at the “winners” and “losers” of each 2012 Republican presidential debate, according to The Fix,  and cross-references their performances with how each candidate was polling before and after each debate. (We are all about accountability here, after all.)

Some appeared to gain momentum after a strong showing, while others lost ground after a Fix-rated losing performance. But in sticking with the running theory that debates don't really matter, most candidates saw no change at all. And even for those who did see a change, it's impossible to say that it was definitely because of their debate performances.

With that said, here's a broad overview of where all the candidates stood during the 2012 primary:

Where debates may have mattered

Herman Cain — “Winner” of Oct. 11 debate

The Fix called Herman Cain a “winner” of the Oct. 11 debate. His catchy 9-9-9 economic plan was mentioned 12 times in the debate. And indeed, he earned his highest polling numbers of the primary shortly after the event.

Rick Santorum — “Loser” of Feb. 22 debate (and Mitt Romney as “winner”)

Santorum’s late-January surge ended days after a poor showing in a Feb. 22 debate. He led the Republican field with an average of 33.6 percent in five polls leading up to the debate. Santorum’s average dropped to 25.6 percent in five polls following the rough debate – an eight-point fall.

The Fix also said Mitt Romney was a “winner” of the debate, and the eventual GOP nominee's poll numbers from the time reflected that description.

Newt Gingrich — “Winner” of Jan. 16 and Jan. 19 debates

The former House speaker strung together a series of well-reviewed debate performances in the middle of January. His national polling average briefly rose toward the end of the month before falling for the rest of the race.

Mitt Romney — “Winner” of Jan. 8 debate (and Newt Gingrich as “loser”)

Romney’s national polling average jumped by more than six points following his “winner” performance in the Jan. 8 debate.

This example is imperfect, though; Newt Gingrich, declared a “loser” of the Jan. 8 debate, held his polling numbers mostly steady over the days following.

Where debates didn't matter

Newt Gingrich — “Winner” on Dec. 10, “loser” on Dec. 15

Newt Gingrich’s slide in December doesn’t appear to be related to his debate performances. His numbers were decreasing before Dec. 10 — and all four polls shown between the debates include responses taken before Dec. 10.

Rick Perry — “Loser” of Nov. 9 debate (and Newt Gingrich as “winner”)

Rick Perry had his “oops” moment in the Nov. 9 debate. The Fix declared him a “loser” of the event, but his polls — which had already plummeted — actually stayed fairly steady afterwards.

Gingrich was a “winner” in the Nov. 9 debate. His numbers did increase after the debate, but they were already on the rise (that Nov. 10 poll you see marked below began fielding responses the day before the debate).

To sum up, we found that the running theory on debates’ effect on the polls is probably right: A massive majority showed almost no change or changes within standard polling margins of error.

To be sure, there are other factors at play here — including the relative irrelevance of early polling, the media’s role in polling surges — e.g. picking debate “winners” and “losers” in the first place (cough, FIX, cough) — and a reminder that correlation does not necessarily mean causation.

But — and there’s always a but — there does appear to be evidence that some primary candidates, in some very specific cases, can gain traction after a great debate performance. Whether that’s due to the debates themselves, though, is far from certain.

Note: We’ve only included polls from trusted firms conducted via live phone. We also excluded polls of “likely voters” since this can be difficult to forecast so far from Election Day. Charts show polling end date (the last day “in the field”).