On Polling

Voting is hard to explain, since it is behavior that ought to suffer from what Mancur Olson famously dubbed the "collective action problem." That is, given that my vote is unlikely to make any difference in the outcome (true in the overwhelming majority of cases), why should I go to the trouble when I know that others will do it? Olson wrote that we can overcome such problems either through the distribution of benefits that reward an activity (like voting), or by the introduction of penalties that punish failure to act. Political scientists, political theorists, and philosophers (all often very different kinds of creatures) have often explained voting as either an expression of narrow self-interest (Madison?) or (perhaps more grudgingly) as an expression of some sort of cultural norm (Tocqueville?). Some people vote because they want things, others because it's their civic duty. Those who do not vote, by extension, if we presume that there is a rational, conscious decision-making process at work (perhaps a heroic assumption), either do not see that they have anything to gain from their participation (notice that we have not really confronted the collective action problem head on), or that they have not adopted voting as a social/cultural norm. That brief lumpy introduction should give you enough foundation to think about the following short excerpt from the first Chapter of David E. Campbell's Why We Vote:

To make the case that the communities in which we spend our adolescence affect whether we vote in adulthood first requires establishing that

a. communities shape the civic and political engagement of the people who live within them, or what you do now depends on where you are now
b. the engagement of adolescents in particular is shaped by where they live, or what you did then depends on where you were then
c. adolescents’ engagement links to their engagement as adults, or what you do now depends on what you did then

Together these claims lay the foundation for the book’s central argument: the civic norms within one’s adolescent social environment have an effect on civic participation well beyond adolescence: what you do now depends on where you were then.

Why am I subjecting you to this? Because so much of our discussion has danced around questions of why person (or group) x will vote for candidate y or candidate z without attending too much to questions of who it is who will vote in the first place.

In most recent presidential elections, this actually hasn't mattered too much because the electorate has been relatively stable -- all else equal, wealthier, better-educated, older, whites are those most likely to appear at the polls and poorer, less-educated people of color the least likely. Is this a product of socialization, as Campbell would have it, or of some belief held by the former group that their are benefits to be derived from participation and, conversely, a calculation by the latter that it matters little? (And what's the relationship/correlation between an adolescent's "environment" and her family or community's economic status?.)

And, again, I hear you ask, why am I belaboring this, and what does this have to do with polling?

Reason one: Because this election contains the possibility that Obama's candidacy will persuade significant numbers of the latter group to vote either for the first time or for the first time in a while -- and it is an open question as to whether opinion polling is adequately capturing these "new" voters. I know -- that was a long way to go for a dig at the trustworthiness of polling, but another (among many) things to keep in mind as you (inescapably) see and read poll results. Use many, many grains of salt, please. . .

Reason two: This is more fodder for thinking about the endogeneity/exogeneity problem: where do preferences come from?