Gallup's bias continues

I continue to believe (as I've written before) that Gallup intentionally, even if not persistently, analyzes its poll results and titles its press releases in a way designed to bias readers' interpretations of the data against Republicans and for Barack Obama.

Another example came Thursday when I got an e-mail update from Gallup entitled "Republicans Less Enthusiastic About Voting in 2012."

Before you read the story, tell me what your initial reaction is...what you think that headline means or would mean to an ordinary American who read it.

To me, there are a few most likely things it could mean, in this order:

  • Republicans are less enthusiastic about voting in 2012 than Democrats are
  • Republicans are less enthusiastic about voting in 2012 than they were in 2010
  • Republicans are less enthusiastic about voting in 2012 than they were during the last presidential election

Of these, the first is the most likely implication of the e-mail and article's title, but it is false. The third statement is also false. The second statement is true, but not what Gallup means.

In fact, none of the three is what Gallup means; these, especially the first, are just what they want you to think they mean if you only read the headline.

What Gallup is really saying, though nobody less than Carnac the Magnificent could divine it from the title of the article, is that Republicans are less enthusiastic about voting in 2012 than they were in Gallup's last survey, just over two months ago. Big effin' deal.

Here's the lead paragraph from Gallup's analysis of their own poll: "Republicans' enthusiasm about voting in the election for president next year has decreased, with 49% of Republicans and independents who lean Republican now saying they are more enthusiastic than usual about voting, down from 58% in September. This narrows the gap between them and Democrats, 44% of whom are more enthusiastic than usual, essentially the same as in September."

Having watched the way Gallup's writers work for some time now, something jumped out at me here as well. Take a guess...

If you said "I wonder what they're hiding by saying 'essentially the same'" you get a gold star.

What "essentially the same" says to me is that I need to go look at the prior data because the change is likely a modest negative for Democrats that the Gallup writer is trying to avoid making plain.

And indeed that's the case here: In the current survey, 44 percent of Democrats and Dem-leaning independent voters are more enthusiastic about voting than they were in the prior election while 47 percent are less enthusiastic. In the prior survey, done in September, 45 percent of Dem and Dem-leaning voters were more enthusiastic and 44 percent less enthusiastic. In other words, there was a four point swing away from voter enthusiasm by Democrats during the last two months. Yes, that's much less than the swing among Republican and GOP-leaning voters, but it is still worth mentioning because it shows that at least some part of the GOP enthusiasm drop is part of a electorate-wide drop rather than particular to Republicans.

In the September article, Gallup mentioned that the data at that time was "nearly identical" to 2004 data...but they then included the 2004 data in parenthesis so readers wouldn't have to go dig for it. Of course in that case the 2004 data was one point stronger for Republicans, so the comparison, even if old and statistically insignificant, was a change for the worse for GOP voters. Apparently "nearly identical" or "essentially the same" data is only worth specifying if it can be taken as bad news for Republicans. Otherwise, Gallup prefers that you don't see the numbers, at least not easily.

In fact -- and I realize this may sound a little tin-foil hat-inspired -- the link from the current article to the prior article (that you'd have to check to see what "essentially the same" meant) did not work. It took copying and pasting the link, then taking out an erroneous part, to make it work. The average reader would therefore have not been able to read the prior article and see that Democrat enthusiasm has also dropped, albeit less than GOP enthusiasm, since Gallup's prior polling of the same question.

It turns out that the writer of the misleading article this week was none other than Frank Newport, Gallup's Editor-in-Chief. I suppose any note I write to Gallup about this apparent reporting bias will fall on deaf ears. To be fair, Mr. Newport has been the subject of criticism by Democrats, and often gives non-partisan responses to questions. Nevertheless, the poor work done by Newport and Gallup in this and other cases leaves me with no way to conclude anything but a pro-Democrat bias, even if a subtle one.

No feedback yet
Leave a comment

You must be logged in to leave a comment. Log in now!

If you have no account yet, you can register now...
(It only takes a few seconds!)