During the ups and downs of this U.S. recession, Americans' faith in small business has grown, while their faith in big business has not. Three times more Americans now say they have "a great deal" or "quite a lot" of confidence in small business (66%) than say this about big business (19%).And yet...I'm not seeing a whole lot of people rejecting Amazon and Barnes & Noble for independent bookstores, or replacing Olive Garden with real Italian joints. Nor do I see Walmart shutting down while neighborhood grocery stores thrive, or local hardware stores replacing Lowe's. And, you know what? I'm willing to bet that people go to McDonald's, among other reasons, because they have enormous amounts of faith that they'll get exactly what they expect, while they have no idea what they'll get at that new place that just opened in the strip mall they just drove by. In other words, it's not hard at all to find actions that appear to be the exact opposite of what you would expect from the survey results.
Does this mean that the poll got it wrong? No, not exactly. We can be confident that the poll accurately reflects how the entire nation would answer the questions Gallup asked. We can also guess that the result indicates something about what people think about big and small business, although I suppose one could argue for a variety of interpretations. What we can't do, based on only this result, is to predict behavior. And I'm even reluctant to say that we can say anything in particular with great confidence about what the answer means, beyond the fact that if you ask people those questions, you'll get those answers.
Now, sometimes we can in fact predict behavior from survey results. It helps when the question is closely tied to a clear action (i.e., do you intend to vote for Smith or Jones?). It helps when we have a long record of survey results and corresponding actions so we can compare them (we know that people lie about whether they've voted or not, but if I recall correctly there's a clear enough pattern that we can establish useful estimates of how many people falsely claim to have voted). But sometimes, the survey results just don't correspond to any particular action, or even any meaningful underlying position.
Again, just a reminder to avoid overreacting the next time you see, for example, survey results that seem to indicate that people agree (or disagree) with your position on some issue of public policy. I'm not telling you to ignore such results, but only to remember all those people who claim to hate big business and love small business yet flock to chain stores.