Now it's six years later and we're about to witness an important primary in New Hampshire, and there are ongoing investigations about apparently illegal push-polling [which seems to have targeted Mitt Romney] that has gone on in this primary. Do you know anything about that?
Well, some things never change. I don't have direct knowledge about that, because I'm not involved in politics. But what I can tell you from just reading about it is that it's a fairly sophisticated program the attorney general in New Hampshire is investigating. It demands money to do that type of program.

Now, the New Hampshire AG has determined which firms are behind this push-polling, one in Utah and another in Portland, Ore. What do you make of that?
The most important thing to make of that is if they don't appear in the FEC reports somewhere for some campaign, then it's been purposely concealed. So whoever did it knew enough to know that they didn't want to be linked to it. So in order to pay for it, they needed to structure that transaction in a way so that information did not appear on the Federal Election Commission disclosure reports. So while I don't know who did it, the fact that they took such great strides to conceal it says a great deal.

Your firm did a fair amount of push-polling. For the layman out there who doesn't understand the practice, can you explain how push-polls differ from legitimate surveys?
Yeah, a survey is a very legitimate thing. Candidates do it to understand what the salient messages are and what issues voters are predisposed to respond to. They collect data, the data is processed, and the results get filtered into the campaign. So those are valid. A push-poll is designed to be disguised as a legitimate survey, but it's not that. Its mission or intent is really to push a message under the guise of legitimate data collection. That's why it's a dirty trick, because you're not really very interested in the data you collect. You may generate data, but the intent really is to drive home a message in the guise of something else.

Can you give an example?
Say you were targeting evangelical voters in Iowa. You might say, "Would you be more or less likely to vote for a candidate whose faith states that the devil and Jesus are brothers [trying to cast Mormonism in a bad light and tarnishing Romney in the process]?" You know how the voter is going to respond. They are going to have a negative response to that question. You are linking the question to a certain candidate, aiming to leave a certain impression—generally a negative impression.

One of the tricks you used in push-polling was to use callers who had particular accents. Can you explain that?
Sure. Say you're calling a Democratic household, perhaps a union household. And this isn't necessarily a push-poll. You're just calling them with a [fake] advocacy message. You've got a script, say, about NAFTA, and you're describing a candidate's position about NAFTA. But maybe you do it with an accented voice. In this case, maybe a Hispanic or Latin accent would be effective. What you're doing is not only telling that household what a certain candidate's position is on a certain issue, when you know that household is predisposed to be opposed to the candidate because of that position. You're also adding in this additional layer of bringing this message into the household with an accent. Essentially, you're tapping into what you assume is latent bigotry.

So, for instance, you might use a Hispanic voice in targeting households you knew had strong opinions about illegal immigration. Then you'd have a Hispanic voice call up on behalf of a particular party, talking about the importance of legalizing immigrants, knowing that would be a turnoff for whoever was in that household?
Yes. Immigration is a very good example. Rather than the phones, let's talk about the mail. You might mail an immigration piece—a piece that says [in the name of a Democrat], "I'm the candidate for tougher immigration laws," with a picture of someone climbing through the border fence. Then you might send that piece of mail to a household with a Puerto Rican surname. The purpose of that is to incite that household that might otherwise vote Democratic on that issue.