We keep hearing the pundit class braying about the coming Democratic win in November, based largely on three polling figures. One is the low approval score of President Bush, who isn't on the ballot. Another is the "generic ballot," which has never been a strong predicter of the outcome. The third is the "right track/wrong track" figure, which measures voter satisfaction with the general direction of things.

I missed his analysis of the "generic ballot" question, although it is linked in the text, but RealClearPolitics' Jay Cost has a strong piece up today on the "right track/wrong track" numbers:

Pundits will usually point to high negatives in both 1982 and 1994 as evidence of the predictive power of the "right track/wrong track" statistic. However, the plural of anecdote is not data. If we are going to use "right track/wrong track" as a guide - we have to be a little more systematic than that.

As I have mentioned in the past, the predictive power of any given indicator is measured by its coefficient of determination. This is a percentage value between 0 and 100. It answers: "what percent of change in phenomena P is anticipated by a change in cause C?" The higher the percentage, the better explanation C provides.

Gallup has been asking "right track/wrong track" in every election year since 1982 (the actual wording of the question is "are you satisfied" or are "dissatisfied"). If "right track/wrong track" is an important factor in congressional elections, we would expect it to have a high coefficient of determination when compared with the share of the two-party vote of the President. This would mean that, when people are satisfied, they reward the party of the President. When they are not satisfied, they punish the party of the President.

As it turns out, this is not how it works -- at all. Changes in the final Gallup "right track/wrong track" only anticipate 11.6% of changes in the President's party's share of the two-party vote between 1982 and 2004. They anticipate only 11.25% of changes in the President's party's share of House seats. Why is that the case?

Read all of this excellent analysis at the link above. You can also get Cost's piece explaining the generic ballot via hyperlink.

Cost does not address the question of WHY media sources continually rely upon poll numbers which have not proved predictive of election results in the past, but I think we all have a pretty good idea of the answer research would yield.