During a presentation I gave last week at ConvergeSouth I mentioned that, in my experience, making decisions based on survey results was a dangerous proposition. For instance, when I was working in a marketing department we'd constantly ask customers if they'd be willing to spend $x on a product they'd say "yes" and then we'd send them an offer with that exact price and a rare few would actually buy it. In other words we learned real fast to follow the money and largely ignore what people said they'd spend.

Here's the thing: we knew people weren't lying to us, but it was a lot easier for them to say they'd spend money than to actually spend it. In essence they were lying to themselves. According to this interview on Freakonomics most of us are very good at lying to ourselves:

DUBNER: I wouldn’t say you’re wrong there, but let me also say this: we lie to ourselves all the time. We’re constantly trying to predict how we’re going to behave in the future when something happens.  A tax hike.  A price change.  A Presidential election.  And we’re almost always wrong.  Take something as simple as driving. The American Automobile Association is constantly surveying drivers.  They’ll say something like, “if gas prices stay as high as they are now, or go up, will you drive less?”  And people always say, “oh, absolutely!”  And then you look at the data and they do not drive less.  Here’s Joel Weichsel with AAA:

Joel WEICHSEL: “I think there may be people who lie to themselves, or imagine that they’re doing something that they’re not.  But I think there are also people who maybe forget about things that they’ve done.  

RYSSDAL: “Forget?”  He’s being very polite and saying “we’re lying to ourselves.”  That’s what he’s saying.

DUBNER: It’s a synonym.  But I will say this: I don’t believe it’s necessarily intentional. One problem with any survey is that the power of suggestion comes into play.

The full interview can be heard here and it's worth a listen: