July 08, 2004
Rural voters revisited

Bill Bishop catches the New York Times in a dumb mistake regarding rural voters.

Last week, the New York Times had a front page story headlined “Bush and Kerry Vie for Support of Rural Voters.” Both campaigns pushed out into the heartland over the July 4 weekend and the story was about how rural America was a strong supporter of President Bush.

That’s true. However, the Times reported “the rural vote, which amounted to 23 percent of the electorate in 2000, broke decisively, 59 percent to 37 percent, for Mr. Bush four years ago.” Well, that’s wrong. The rural vote is a little over 20 percent of the electorate. (Lasso is counting all nonmetro America, the widest definition of rural.) And the spread in 2000 wasn’t 23 points. It was 16 points. Still a healthy difference, but the Times report is off by 44 percent.

What happened? The Times based its “results” of the 2000 election on exit polls. Pollsters asked people where they lived and how they voted. Based on that, the newspaper reported how rural America voted. (It turns out some people like to say they live in rural America even when they don’t; and these people are more likely to vote Republican.)

But why depend on a poll? The actual real-enough true-blue results are right there for everyone to see. And based on that poll (which included 100 percent of all voters), the Times was wrong.

The Times was also wrong when it reported in a graph that Republicans won rural America in 1992 and 1996. They didn’t. When you count the actual votes in those elections (what a concept!!), you find that Clinton won rural voters in both years.

In 1992, Clinton won 40 percent of nonmetro voters; George H.W. Bush won 38.9 percent. In 1996, Clinton won 44.4 percent of all rural votes; Bob Dole won 43.5 percent. Close, but hardly the rising tide of Republicanism the Times (and Republicans) would have you believe.

I emailed Bishop about this, and while the results are there to see, they're not there to see for free. Bishop got the data from Dave Leip's Election Atlas, but the county-by-county numbers are only available for pay. Not that this excuses the NYT - it's not like the Statesman has the budget for this sort of thing and the Times doesn't, after all.

One corollary of what Bishop found is that Ross Perot overperformed in rural areas in both 1992 and 1996, as his total share of the vote in each of those elections (18.9% and 8.4%, respectively) was less than the remaining share of the rural vote (21.1% and 12.1%). I don't really have a point to make about that, I just thought it was interesting.

Posted by Charles Kuffner on July 08, 2004 to The making of the President | TrackBack

Of the numerous significant problems at the Times, this one seems like a minor one to obsess over. Not that they all shouldn't be pointed out, but this doesn't seem like the sort of transgression that took place regularly under Howell Raines. Still, it's very useful that there are watchers of these sorts of things, and watchers to keep the watchers honest.

Posted by: kevin whited on July 9, 2004 12:02 PM