Chronicle poll on district 22

The Houston Chronicle has released a survey by Professors Murray and Stein. The full list of questions is here. While a few (but not exhaustive, by any means) crosstabs are here. Meanwhile, here is the graphic from the news article.

1/10-12. 560 REGISTERED voters, =/- 4.1% margin of error
party ID: 42% GOP, 27% D, 23% independent, 8% other

General election:
DeLay 22%
Lampson 30%
Stockman 11%

General election (including leaners):
DeLay 25%
Lampson 32%
Stockman 14%

The wording of this question very substantially overstates Stockman's support ("if the election were held today and the candidates were [rotate choices] Congressman Tom DeLay, former Democratic Congressman Nick Lampson and former Republican Congressman Steve Stockman, for whom would you vote?"). The way the question is worded makes it much more likely for Republicans to choose Stockman on this survey than on election day.

A campaign pollster would be unlikely to choose wording such as this, because the wording leads to the results being less predictive than they might otherwise be.

On to the primary numbers...

Primary:
DeLay 39%
Tom Campbell 6%
Mike Fjetland 2%
Pat Baig 1%

Where is the support for the numbers immediately above? This is from the graphic that accompanied the news article (the 39% is also mentioned in the news article), yet I can't find these numbers anywhere in the poll results or crosstabs.

What I can find is this:

Four candidates are running for Congress in the Republican primary being held in the 22nd Congressional District. If the election were held today, which of these candidates would you most likely vote for?
Tom DeLay (120) 21%
Pat Baig (13) 2%
Tom Campbell (34) 6%
Mike Fjetland (12) 2%
No Answer/Refused (381) 68%

Did this poll really ask EVERY person surveyed (including Democrats) who they were voting for the in Republican primary? That's what the results seem to indicate. If they asked every Democrat surveyed who they were voting for in the GOP primary, then that's sloppy surveying and makes the results of the primary question worthless.

It's totally bizarre. It's almost unfathomable for them to have asked every person surveyed who they were voting for in the Republican primary.



Things I've heard (mis)reported, based on this poll:
Channel 11's news report said that 91% now had a lower opinion of DeLay. Channel 11 also said that 47% think DeLay should resign. This is sloppy reporting by Channel 11. From the poll:

Has your opinion of Tom DeLay changed over the past year?
Yes (213) 38%
No (332) 59%
Don't Know/Refused (15) 3%

Do you now have a higher or lower opinion of Tom DeLay than you did last year?
Higher (7) 3%
Lower (194) 91%
Don't Know/Refused (13) 6%

Should Congressman Tom DeLay withdraw as a candidate for the 22nd Congressional seat?
Yes (263) 47%
No (226) 40%
Don't Know/Refused (71) 13%

So, based solely on the survey, only 35% (91% of 38) have a lower opinion of DeLay. Also, "withdrawing as a candidate" is much different from "resigning." It may seem like pedantry, but the two are definitely not equivalent in a poll.

Undersampled Republicans?
This survey appears to have undersampled Republicans. The district is about 62-63% GOP. According to the party breakdown in the survey, 42% is GOP, 27% D, 23%. When the independents (according to a question in the survey) are asked which way they leaned, then you end up with 36% D, 50% GOP, 10% "strict independent," with the remaining 4% being other. Given that the district has voted around 62-63% Republican in the past, this seems like a rather low number of Republicans.

There are a few ways to take this: 1) the survey undersampled Republicans, 2) the district is trending Democratic demographically (there is some very mild support for this), 3) district 22 voters are upset with Republicans and thus are identifying more with Democrats right now.

My guess is that the discrepancy is best explained by some combination of the above 3 possibilities, with option 1 being the most prevalent, and option 3 being also part of the reason.


Big picture points:

1. This isn't good for DeLay, although it's not really unexpected. The survey puts DeLay's favorable/unfavorable at 29%/60%, and it shows that many voters who have voted for him are considering other candidates because of his legal situation.

2. Unless DeLay's legal situation deteriorates, this is very likely the nadir of his poll numbers. There's no doubt that after the last few weeks, DeLay's poll numbers would be bad. But is this really the best time to take a poll, if you're attempting to predict the outcome of upcoming elections? No.

3. As far as I can tell, unless the Chronicle or Stein/Murray choose to clarify the primary results, it's very hard to take the primary results seriously.

4. If DeLay wins legally, he'll win re-election. If he loses legally, he won't win re-election. This guy, in the Chronicle's article, put it well:

"I go a lot on the way somebody talks and he doesn't give me warm and fuzzy feelings," said Robert Jones of Pearland, who has supported DeLay but rated him unfavorably in the Chronicle survey last week. Jones, who voted in the 2004 Republican primary and considers himself an independent, said he hasn't decided whom he will support this year.

"A lot depends on what happens with Tom DeLay," he said.

Yup. What happens in court will decide this election.