Oct 18 2006

Where Is The GOP House

Published by at 9:04 am under 2006 Elections,All General Discussions

Update: One really has to questions polls which ignore the Battleground Poll’s finding that the country is 61% conservative verses 34% liberal. Without an accurate model the polls are junk. – End Update.

If you look at the top 40 GOP at-risk house seats at RCP one would think the Democrats have a good chance of controlling the house. The problem is RCP uses the poll numbers as is, as if there is no bias in polls – something we all know is not true simply by recalling the 2004 election. My experience has been a bias of 3-5% against Reps in the polls over the years. So we need a way to detect the bias and account for it. What I look for is to consider the VA model, where the state runs about 5+% in actual vote margins compared to the early polls (for some reason the polls become very accurate the last week – and differ from the ones taken running up to the last week).

So let’s assume a poll bias (5%) will tip a race in any district that voted clearly for Bush, and will not have an effect in a district that was close or for Kerry. Then we can baliwick the Top 40 into four categories: Probable Rep, Toss Up, Probable Dem and Special Factors (the poll bias model won’t apply). The last category is for races where the candidate’s individual issues are driving the polls, not any inherent bias. Those I will address individually. I have kept the RCP ranking in the listings below so folks can see where I differ with the experts. With that in mind, this is AJStrata’s warped sense of the field of play this week:

Probable Rep

6. CO-7: (Open – R) – The polls have been trending towards the Rep with the last two being down 6 and a tie, giving a 3% advantage to the Dem. The district voted 52-47 for Bush so the bias correction can be applied making this a probable Rep (due to the poll trend line).

12. IN-2: (Chocola – R) – The Trend is definitely heading towards Chocola from a high deficit of 16% to the last two polls averaging 2.5%. The District voted for Bush 56-43 so applying the bias correction puts this at +2.5% for the Rep. Combined with clear trend lines this is probably a seat the Reps hold.

16. IN-9: (Sodrel – R) – Bush won by 59-40 in this district. Polls trending towards Sodrel quite nicely. Last two polls show an average of 5% for the democrat. Applying the bias correction would put this at a tie, but the trend is clear where the momentum is heading. I could use the last poll and have Sodrel up 3% after the correction.

19. IL-6: (Open – R) – Polls show between a tie and a 2.5% Dem (if you average the last two polls). District went for Bush 53-47. Applying the bias correction results in a Rep hold.

20. MN-6: (Open – R) – Ignoring the Star Tribune poll (these polls are always way off) we see the average of the last two polls is 1% for the Dem. The district went for Bush 58-42. Applying the Bias correction gives the Rep a 4% lead. This seat should stay in Rep hands.

21. FL-13: (Open – R) – The Democrat pollster poll is suspect. The last poll showed 2% lead for the Dem in a district that went for Bush 56-43 in 2004. I would either apply the bias correction to the 2% and come up with a 3% lead for the Rep, or say the trend line is heavily towards the Rep. Either way I would expect the Reps to hold this seat.

22. CT-2: (Simmons – R) – Simmons is leading in a District that went heavily towards Kerry. My guess is Simmons will hold.

24. OH-2: (Schmidt – R) – RCP has no details on this race. If I average the last two polls it is the Rep up by 3%. If we add a bias correction it is a clear win. So it goes in the Rep column.

25. WA-8: (Reichert – R) – Polls show an average rep lead of 2.5%. Kerry took this district so a bias correction is risky to apply. If we did the Rep would have a clear edge. So we keep it in the Rep column.

26. FL-22: (Shaw – R) – No polling is showing any real movement towards the Dem and Shaw was elected by a 65-35 margin in 2004. Stays Rep.

27. OH-1: (Chabot – R) – No indication of shift towards dem.

Toss Up

2. AZ-8: (Open – R) – Dem Giffords has been leading, but that lead is shrinking from 12-25% to 8%. The district voted for Bush 53-46%. Applying the bias correction puts it at +3 Dems. Because of the trend in the polls it may be even closer now.

8. OH-18: (Open – R) – The district voted 57-43 for Bush. Polls show small trend towards the Rep. Applying the bias correction results in a 2% lead for the dem. This one is a toss up.

10. NY-24: (Open – R)– No polling at RCP. Open seat contest. Split Bush-Kerry vote. Toss up.

15. CT-4: (Shays – R) – The polls show a tie, the District went for Kerry, I am hesitant to apply the bias in this case. If I did it would push Shays into the win column and a keeper for the Reps. One to watch for the national trend. If Shays pulls ahead the Democrat tide will have ebbed.

18. PA-6: (Gerlach – R) – The last two polls show an average lead for the Dem of 4%, which if the bias was applied would leave a Rep +1%. Given the district went for Kerry I see this as an obvious toss up.

23. PA-8: (Fitzpatrick – R) – No way to tell on this one!

Probable Dem

4. IN-8: (Hostettler – R) – The Democrat is so far ahead the bias correction will not change this race.

7. IA-1: (Open – R) – The district voted 53-46 for Kerry, so the bias would not apply. The Dem also has a solid lead in the polls.

11. OH-15: (Pryce – R) – The district split Kerry-Bush 50-50. Pryce won two years ago 60-40. There is only one poll out (+12 Dem) making the poll risky to use (need various polls from various sources). But since this analysis is simply to apply the poll bias correction and re-assess, I must put this race in the Dem column for now (+7). I have my doubts on any race that is tied to Foleygate as the reason a popular incumbent will lose.

14. NC-11: (Taylor – R) This race is marginally in the Probable Dem category. The District voted for Bush 57-43 so applying a bias correction brings the race down to maybe 4% for the popular Ex NFL QB (and Dem) Schuler). But Talylor won 55-45 last time so I am again suspect on why a popular incumbent would be tossed out for a famous name in a Republican district which would abhor a Democrat led House. If a later poll comes in showing a much tighter race I would change this assessment.

17. NM-1: (Wilson – R) – District went for Kerry and the Dem has an 8 point lead in the latest poll. Even applying the bias correction would leave this a likely democrat pick up.

Special Factors

1. TX-22: (Open – R) – This being Delay’s old seat which voted 64-36 Bush I still need to give this to the Reps. If there is a Dem wave I will be wrong. But I am not seeing a Dem Wave. Goes Rep

3. FL-16: (Open – R) – This is Mark Foley’s old seat. The district voted for Bush 54-46. The average of the last three polls is 6% for the Democrat. Applying the bias correction would make this a toss with a measely +1 Dems.

5. PA-10: (Sherwood – R) – One thing I do not like about the RCP method is it puts in any poll out, which can allow one poll to push the average out of skew if there are not a lot of polls out for a race. This is one where I am not sure about the lates poll. But assuming a 9% lead and observing this very conservative district voted 60-40 for Bush means we can apply the bias and determine a likely 5% lead for the Dem. The personal issues with Sherwood make this a Probable Dem pick up. If later polls show a fear of a Democrat Congress tightening this race I may move it.

9. NY-26: (Reynolds – R) – This race is being rocked by the Foley non-scandal. Why Reynolds is even mentioned in the scandal is beyond me. So while polls showed an initial hit for Reynolds, and unless there is something else in play, I would expect the Foley fallout to subside and bring this race back into the Rep column. The District’s 55-43 vote for Bush tells me it probably will not give Democrats the House over Foleygate. I agree this is a toss up, but I could just as easily put it into the Rep category. Will keep it a Toss Up until another poll comes out.

13. PA-7: (Weldon – R) – Weldon’s race just got very complicated with the late breaking FBI raids in his family. The polls show a Sestak lead of around 4.5%. The District voted 53-47 Bush so the bias can be applied – gingerly! My guess is the raid may cause a sympathy vote towards Weldon since the target is his daughter. Toss Up for now.

I only went through the top 27 races because after that number the idea these races were in play was getting quite thin. So the end result I see is this: Dems look to pick up 6, with another 7 as toss ups. Running the table the Dems will not take control of the House. But I doubt they will run the table of the toss ups. So I would say the dems could win 6-9 seats this year, and fall short of their fantasy. I understand why Rove would be so confident right now.

Update: What is bizarre is the news that the Democrats see pick up of 40+ seats! Something is happening this election cycle with the polls, there is no doubt. Why do I get the feeling the Dems are using the same pollsters and polling methods that convinced them Al Gore was going to cake walk over Bush? We are going to see a political disaster this fall, no two ways about it. But which side will it be on?

21 responses so far

21 Responses to “Where Is The GOP House”

  1. “For the others, their vote will be unchanged, if they NORMALLY vote republican they still will because they don’t know the difference.”

    Didn’t I read a couple of weeks ago that according to FL state law, a vote for Foley will go to the Republican?