20121003_me_06.mp3?orgId=1&topicId=1007&d=440&p=3&story=162103368&t=progseg&e=162163970&seg=6&ft=nprml&f=162103368

Brett O'Donnell is a debate consultant who trains Republican candidates. He has worked with George W. Bush and John McCain, and for a short time earlier this year, he helped prep Mitt Romney.

O'Donnell is an expert on "the pivot."

If you have watched a debate, you have watched a pivot. "The pivot is a way of taking a question that might be on a specific subject, and moving to answer it on your own terms," O'Donnell says.

Take, as just one small example, a moment from the 2004 debate between President George W. Bush and Sen. John Kerry.

You could, by the way, just as easily use an example from Kerry — both Bush and Kerry used pivots roughly the same amount of the time.

In this case, the moderator, Bob Schieffer of CBS News, asked President Bush about job loss. What, Schieffer wondered, would Bush say to someone who has lost his job? Bush began by promising to "continue to grow our economy" and then, subtly, changed course. Suddenly, Bush was talking about education, specifically his signature No Child Left Behind legislation. "I went to Washington to solve problems," he explained. "And I saw a problem in the public education system."

In two or three sentences, Bush had moved from a question about lack of jobs to an answer about education and a promise fulfilled. That is the power of the pivot. Which is why, in the age of debate coaches like O'Donnell, candidates in both parties use them all the time — "frequently," O'Donnell says, "better than 60 or 70 percent of the time, I would say."

The question is, how good are viewers at identifying these "pivots" — or, in the language of my people (journalists who ask questions), "dodges"?

How good are you?

There's a man at Harvard who actually has an answer to that question.

The Pivot And The Brain

Todd Rogers, a behavioral psychologist at Harvard's Kennedy School of Government, got interested in looking at pivots, or dodges, or whatever you want to call them, after watching the 2004 Bush-Kerry debate I quoted earlier.

To him, the dodging on both sides of that debate was enraging, and he couldn't understand why others didn't feel the same.

To figure it out, he decided to do a study that tried to replicate what typical viewers see when they watch a debate.

He recorded a moderator asking candidates a series of questions.

In the first question, the moderator asked the candidates about health care in America, and the politician answered with a health care answer — a long disquisition on why Americans could not afford the care they needed.

Rogers then took that answer and used it as a response to a totally different moderator question, this one about the problem of illegal drug use. So one set of people saw a candidate answering a health care question with a health care answer, while another group saw an illegal-drug use question answered with a health care answer. Essentially, the second group saw a relatively subtle pivot, from drug use question to health care answer.

Finally, he had a third group view the moderator asking a question about terrorism, which was answered again with the exact same health care answer — a much more blatant shift.

At the end of this he asked the different groups two things:

Can you remember what question the person was asked?

How honest, likable and trustworthy is this person?

'Exploiting Our Cognitive Limitation'

What he found was that when a politician answered the health care question with a health care answer, viewers could recall the question and thought the candidate was likable, honest and trustworthy.

When the politician pivoted a little bit and answered the illegal drug question with a health care answer, viewers could not recall the question — but they didn't penalize the politician at all. "Listeners thought he was just as honest, trustworthy and likable as the guy who actually answered the question," Rogers says.

It was only when the politician answered the terrorism question with a health care answer that people could actually tell. "Everyone noticed, and they thought he was a jerk," Rogers says.

This led Rogers to the conclusion that people are capable of detecting dodges — but only if they're egregious. They don't seem capable of detecting subtle evasions.

Rogers believes this is because we have limited attention, and most of the time when we're watching debates, we spend that attention on social evaluation — Do we like this person? Do we trust this person? — and only generally monitor content.

It's only when the speaker is wildly inconsistent that some deep mental wire is tripped. "It raises some flags, and we direct our limited attention to assessing whether this person is doing something unusual by failing to answer the question and offering an egregiously different answer," Rogers says.

This, Rogers believes, is why politicians can get away with dodging questions as much as 70 percent of the time.

Copyright 2016 NPR. To see more, visit http://www.npr.org/.