The coming Democratic bloodbath in the House
We have good reason to expect the Republicans to gain at least 50 House seats in the 2010 Congressional election. But, the question arises, isn’t it too early to make such a prediction, and wouldn’t you need a separate poll for every vulnerable House seat? Evidently not. A survey question known as the “generic ballot” asks people something like: “If the elections for Congress were being held today, which party’s candidate would you vote for in your Congressional district?”
For example, see Question 2/3 here. The website Pollster.com gives us the results to date of how people have responded to this question, and presents a graph along with the data here. The latest numbers have the Democrats getting 42 percent and the Republicans getting 43.7 percent with the rest (14.3 percent) undecided. One might think that these numbers are so close, they indicate a virtual stalemate for November thereby preserving the status quo. Not so. As we shall see, these numbers indicate an absolute disaster for the Democrats. To understand why, we need some political theory, history, and statistics.
Posted by Lawrence Auster at February 27, 2010 02:05 PM | Send
First we have to adjust the polling results for undecided respondents. There are many ways to do this, and the following is not necessarily the best, but it’s rational enough. The Republicans got 51 percent = 43.7/(43.7+42) of those who could make a choice, so give them about half of the undecided respondents. This bumps up the Republicans to 51 percent versus 49 percent for the Democrats, making the original 1.7 percent difference into a 2.0 percent difference. If we express the difference as a ratio, then the Republicans come out at 1.04 meaning the Republicans got 104 votes for every 100 votes the Democrats got. I re-express the polling results as a ratio so the readers can understand the graphs of past generic ballots. The axis labels on these graphs are somewhat confusing unless you understand they are actually ratios. On this ratio scale, the Democrats are 4 points behind, remember that.
This graph shows the relationship between the generic ballot vote and the actual election vote between 241 and 300 days before the election. The lower (blue colored) points on the graph show what happened to the Democrats when they held the White House. For example, in 1998 the Democrats were ahead by about 4.5 points (remember the ratio scale) on the generic ballot, but came out about one point behind in the House election. In 1994, the year famous for huge Democrat losses in the House, they were ahead two points in the generic poll. In 1946 the Democrats were again ahead in the generic poll, but suffered horrendous losses in the House elections, the worst in modern times. Now you can see how serious their current generic ballot results are. Note that the two points labeled “best” and “worst” cases for the Democrats don’t match the four point difference discussed above. He recalculated the generic ballots, dropping all polls that have been accused of favoring Republicans which includes Rasmussen. He’s really playing it safe, and he gets only about a two point disadvantage for the Democrats. You can read his whole analysis here, which explains everything in more detail, but is somewhat more technical than my discussion.
How do we go from the overall vote in the House to a result in terms of actual seats lost or gained? That gets really technical and goes way beyond the general discussion here. You can read this paper for how this works. Harry Joe applies the theory and predicts at least a 50 seat loss for the Democrats.
Michael Barone also predicts a disaster for the Dems based on the current generic ballot. He uses polling data from Real Clear Politics and gets different numbers. But the big picture is clear; unless something really unusual happens between now and November, the Democrats should suffer an absolute bloodbath in November. Of course, changed demographics and other factors might render the historical patterns obsolete, but I doubt it. The elections in Massachusetts, New Jersey and Virginia seem to confirm the reality of the crimson tide.
However I feel the need to warn conservatives. The real crimson tide, also known as agal bloom (red tides) are often toxic. Putting the Republicans back in control could also be hazardous to our health.