After reading the following from newspaper columnist David Brooks:
In times like these, deficit spending to pump up the economy doesn’t make consumers feel more confident; it makes them feel more insecure because they see a political system out of control. Deficit spending doesn’t induce small businesspeople to hire and expand. It scares them because they conclude the growth isn’t real and they know big tax increases are on the horizon. It doesn’t make political leaders feel better either. Lacking faith that they can wisely cut the debt in some magically virtuous future, they see their nations careening to fiscal ruin.
Henry writes:
Is it so hard to go to public opinion poll data to try to figure out what people think, or to acknowledge that you don’t know what they think when, like, your only source of information is what you think about the topic? No – it isn’t so hard. This is a particularly annoying pundit trick.
I see what Henry is saying–rather than assembling evidence on how various people would behave, Brooks is making assumptions based on his own internal logic..
But I disagree with Henry’s implication that poll data would answer Brooks’s questions, at least in two-thirds of Brooks’s examples.
– Brooks’s first example involves consumer confidence. Here, I agree with Henry that it would make sense to refer to some standard polling data on the topic.
– In Brooks’s second example, he is talking about business decisions. To start with business owners are a small percentage of the population, so you won’t get much on their opinions by simply taking the relevant subset of a national poll. Beyond this, a poll will tell you what people are thinking right now (or, at least, what they are willing to tell the interviewer), but it won’t necessarily give you a bead on what people will actually do.
– Brooks’s third example is how political leasers might feel. As Henry points out, Brooks’s views as a columnist are not necessarily representative of political leaders’ views (or even, for that matter, what Brooks’s views might be were he elected into such a position). But a survey of 1500 randomly sampled Americans won’t tell us much about this particular question either.
I guess what I’m saying is that I’d like to separate Henry’s criticism of Brooks for imputing his reasonable opinions into the minds of others, from Henry’s implication that it’s easy to answer the sorts of questions (about what businessmen might do) from polling data. Trying to guess what a fraction of the population are going to do in a new situation is quite a bit different from estimating what people in the aggregate feel about something right now.
As a public opinion researcher, I’m often impressed by what can be learned from surveys, but I think it’s important to recognize their limitations also.