Huffpost Politics
THE BLOG

Featuring fresh takes and real-time analysis from HuffPost's signature lineup of contributors

Mark Blumenthal Headshot

Party ID: Random Error Happens

Posted: Updated:
Print

There is something about party identification -- the question that asks if voters think of themselves as Democrats, Republicans or Independents -- that always seems to inspire comment and conjecture. Last week, almost as an afterthought, I ended a post noting that the recent Gallup panel survey had a few more Republican identifiers among all adults (36%) than what the standard Gallup telephone surveys had been averaging since July (30%). One valued Pollster.com reader emailed to ask if this might be the sign of a bit of "buyer's remorse," a sign that some voters who felt unsure earlier in the year have come to regret that the Democrats have taken control of the House and Senate. In this case, it looks to be most the usual random variation to which polls are prone.

One obvious argument against any sort of buyer's remorse is that the Bush job rating continues to fall (see Professor Franklin's update from earlier today). Another is that, at least as of a month ago, the favorable ratings of both Nancy Pelosi and Harry Reid were increasing on the Gallup Poll.

However, to look more directly at the reader's question, the simplest approach is to see if any of the other national polls have shown any similar shift in party identification. Both Newsweek and AP/IPSOS, two of the organizations that conducted new surveys in December, routinely provide results for party identification. They show no increase in Republican identification. If anything, the opposite may be true. Among registered voters, both show small (and probably statistically insignificant) decreases in Republican identification since the election.

12-11%20party%20id.png

The table also includes the party identification results among adults released by Gallup on their standard telephone surveys in October and November (it does not include the recent panel survey results -- note also that the Newsweek and AP/IPSOS results are among registered voters).

The most likely explanation for the slightly higher number of Republicans on the panel survey discussed last week is an extreme example of the usual culprit: random variation.  A swing up to 36% Republican among adults would be extreme, but it happens.

PS: I frequently link to reports on GallupPoll.com.  Until recently, Gallup's analytical reports were available for free for the first 24 hours and then to subscribers only thereafter. It looks as if Gallup has recently shifted to more of an advertiser-driven model. Now, if you click on an archived Gallup reports (like the one referenced above), by watching a short advertisement you can get a time-limited "free premium pass" that provides access to everything on GallupPoll.com but the "Gallup Brain" data archive.