ReallyClearlyPolitical Polling Data
Watching the favorable/unfavorable polling numbers on this WH has been quite the pastime for me. I, and I think a lot of political junkies, frequently visit realclearpolitics.com. Normally such a visit takes about a minute or so. I usually chuckle at the CBS, CNN, or NYT polling results. I almost never really try to analyze their numbers because what’s the point, they never do any analysis, so why should I. Obviously, sampling bias is a watchword for them. A normal trip to this site is a but brief layover before I go to Rasmussen’s site. Any polling data that rasmussen.com reports out has been combed through thoroughly and I’m fairly certain analyzed for outliers and anomalous data. All that said, let’s take a look at today’s RCP polling aggregates.
Today the RCP averages use: Gallup, Rasmussen, USA Today/Gallup, Fox News, CBS “News,” Democracy Corps, and CNN/Opinion Research. Of these none are current, Gallup’s last rolling polling data is from 11/28, and all of the rest are seven or more days old. Interestingly, the best results for the leader of the free world, coming from CBS, Democrazy Corps, and CNN/Opinion Research, are close to two weeks out of date. The problem for RCP is in throwing out the three best polls, some odious numbers result. The new results would certainly get headlines and RCP, apparently, doesn’t want to be driving the news cycle in quite that direction.
If one were to factor out the three oldest polls, as mentioned above, the averages change substantially. The approval numbers fall to 49.0%, the disapproval results drop to 46.3%, yielding a spread of +2.7%, a precipitous drop to be sure.
I’m not arguing with the notion that the dropping the oldest three polls is a fair reading of the correct average, though I’m personally given to that interpretation. I do believe that all of the polling data should be deleted, no averages be reported, and an explanatory note as to why this was done should be published. The alternative is to be far more honest about the numbers. I could even be convinced that it would be fair to use only the Gallup numbers, despite Gallup’s use of all adults vs. Rasmussen’s use of likely voters. Whatever the real numbers are, they are not the ones you see posted at RealClearPolitics.