Self-described “political operative” Les Francis, with real-world experience as former executive director of the Democratic National Committee:
I don’t need any polls to tell me that Republicans will do well in November. The “out” party almost always shows significant gains in the first midterm election of a new President.
Political scientists Joe Bafumi, Bob Erikson, and Chris Wlezien, from elite, out-of-touch, ivory-tower institutions Dartmouth, Columbia, and Temple Universities:
P.S. I tried to keep this one short because sometimes a picture is worth more than a thousand words. But, after reading the first few comments, I think I need to explain a bit. So here goes:
Les Francis wrote, “I don’t need any polls to tell me that Republicans will do well in November.”
Bafumi, Bob Erikson, and Wlezien’s graph show that the generic ballot polls (these are what’s shown on the horizontal axes of the graphs above), even months before the election, yield a very good prediction (via a linear model that Bafumi et al. fit to past elections, as shown in the graphs) of actual congressional voting (these are what’s shown on the vertical axis). So, whether or not Les Francis “needs any polls,” they can be very useful for the rest of us. See here for further thoughts on the political implications of the predictability of early generic ballot surveys.
Francis’s anecdotes and insights about Obama, Pelosi, etc., can be valuable but he’s doing himself no favors by dismissing the polls–just because he personally doesn’t understand their value, not having seen, perhaps, the Bafumi et al. graph displayed above–or by making statements such as “Barack Obama’s 2008 winning margin was somewhat out of synch with the political alignment of the country at the time,” which is unsupported by any data I’ve seen. (See, for example, this discussion, from nearly two years ago, on how the Democrats’ gains in House voting in 2008 were in fact as large as their gains in the presidential vote.)