My column for this week looks at the current trends in the generic ballot for the U.S. House and what we can say at this point about what that portends for the results in November. The short version is that current results point to a strong Republican performance, although whether Republicans retake control of the House depends on whether this measure shows any trend over the next 200 days. Please click through and read it all.
Earlier today, Charles Franklin posted two graphics that show trends in the[(http://pollster.com/polls/us/10-us-house-genballot.php over previous election years -- mid-term and presidential -- going all the way back to 1946. If nothing else, his charts give an immediate sense of the growth in public opinion polling over the last two decades.
The big and difficult question noted in the column is whether we can predict where the generic House ballot -- and control of the House -- will end up based on recent history. The most compelling effort to model the outcome based on the generic vote months before the election comes in a 2009
2 paper by Joseph Bafumi, Robert Erikson and Christopher Wlezien. They argue that the generic ballot moves in a predictable pattern during a mid-term election year toward the party that lost the previous presidential election. So, if those previous trends repeat, Republicans will be doing better on the generic House ballot in the fall than they are now.
As Dartmouth undergraduate (and Pollster.com intern) Harry Enten observed two months ago, the Bafumi, et. al. model currently points to a 50-60 seat gain in November. Nate Silver noted the same result this past Friday.
Finally, the column references another model in created by Emory University Political Scientist Alan Abramowitz in 2002. We should note that Abramowitz is currently publishing results for newer model on Larry Sabato's Crystal Ball that most recently forecast a gain of 37 seats for the Republicans in November.