As part of our continuing, Goldilocks-inspired effort to identify what in early election-cycle polls is and is not arguably relevant to reality, I’ll begin by citing John Sides latest please-don’t-read-the-early-polls post at The Monkey Cage, wherein he acknowledges a couple of exceptions. The most obvious is the simple fact that certain Republican primary horse-race polls are being used by Fox News to cast six or seven candidates into the outer darkness of a kiddie-table secondary “forum” on the day of the first presidential candidate debate. You can deplore the fact that Fox did this all day long, but it undoubtedly matters.
Sides goes on to suggest that early polling on issues or voter mood can be relevant reading material, too, but that’s as far as he’ll take it.
I personally think favorable-unfavorable ratings in early polls are worth noting, at least at the extremes. And even primary horse-race polls can supply some data, if looked at from a considerable distance and via averages, not the sort of hopped-up this-is-the-end-for-Hillary spin of last week’s Quinnipiac polls of CO, IA and VA.
But here’s one relatively small item on which I flat disagree with Sides:
I would tend to ignore trial-heat match-ups for the general election. A Clinton-Bush or Clinton-Walker poll now doesn’t tell us much about November 2016. If those match-ups are discussed, I would pair it with more historical context that demonstrates how early polls have often been wrong and also venture to explain why they are wrong and, possibly, what direction they are likely to move.
Yeah, sure, but there’s a reason these trial heats can matter than has little to do with whether they wind up being accurate predictors of general election matchups: they are very useful with primary voters who are interested in a candidate’s electability. If Sen. Fudd is trailing the other party’s likely nominee consistently by 20 points in general election trial heats, that’s worth knowing, particularly if Fudd keeps claiming he’s the strongest general election candidate. The closer we get to decision time for primary voters, the more this sort of data could matter, even though we’ll still be many months away from November of 2016.
If I had the technical capacity, I’d tape a Blue Oyster Cult cover song entitled “Don’t Fear the Data” to serve as a counterpoint to all the lalalalala I can’t hear you reaction to discussion of early polls. Yes, I understand that some reporters and pundits are, well, idiotic on the subject, while others are deliberately spinning relatively meaningless numbers as though they are bulletins from God. But used carefully, most data is worth something.
