Wednesday, November 5, 2008

The Art of Predicting Races

The art of election predictions is getting people to remember the good predictions while forgetting the bad ones. My Obama electoral vote prediction (including that he would win Indiana) appears right on the money. I'll have to take a look at my predictions versus the map, but I'm not sure I missed a single state. There are some states still being decided, however, such as North Carolina, Missouri and Montana.

In the other statewide races I didn't fare so well. While Obama's coattails were present, they were not quite long enough to put the other statewide Democrats over the top. The state Democratic Party got very little mileage out of the fact that a Democrat presidential candidate won the state for the first time since 1964. It appears that the Indiana House is hardly unchanged in a year that they should have picked up several seats. Big winner this morning: the Indiana Republican State Committee that survived the Obama hurricane with barely a scratch.

I'm eating crow for breakfast this morning on the Daniels' 110,000 Marion-Lake Counties vote hole prediction. I knew I was making a mistake when I predicted it but wanted to use a race people knew rather than talk about abstract concepts like baseline votes. If I would have concentrated on the low profile race results in those two counties, rather than a high profile race with a very weak Democratic challenger. It is the baseline numbers in those two counties that are critical to state wide, lower profile races like Attorney General and Superintendent of Public Instruction. I didn't think people would know what I was doing in my analysis.

If you use the Marion County Surveyor's race as a baseline, the Democrat majority baseline in Marion County is 72,000. The Governor lost by about 48,000 votes in Lake County. The base line in that county is majority Democrat, let's say by 58,000. That yields 130,000 baseline for the two counties. I thought the Governor would exceed the baseline by some - but I never accounted for the enormous cross-over vote he received in Marion County and the state as a whole.

Looking at the results, in Marion County, Democrat AG candidate Linda Pence ran ahead of Republican candidate Greg Zoeller by 62,000 votes, while the Democrat candidate for Superintendent ran ahead by about 68,000. I don't know the Lake County numbers for those two races, but it is clear the Republican state wide candidates overcame a huge Democrat base vote in Lake and Marion Counties, probably substantially more than 110,00 votes.


varangianguard said...

Since you used the word "art", a lot of latitude can be given. :)

Personally, I thought Obama's coattails would drive the gubernatorial race to be a little closer. But, was I wrong!

Sad part is, that -I- could have run a better campaign than JLT (and her crew) did. When someone who is, at best, a dilettante feels that they could have done better, then something went seriously south.

Diana Vice said...

Don't be too hard on yourself. Your predictions for the statewide races scared me enough that it forced me to start sounding the alarm bells to friends, family, and neighbors to get to the polls! My son-in-law voted for the first time in his life, and that was because I told him if he didn't there would be NO Thanksgiving dinner for him. Of course I was bluffing, but it worked, didn't it?

Paul K. Ogden said...


I also thought Obama's coattails, whilc not closing the gap for JLT, would keep her loss to around 8 points. I thought the Obama turnout would give the other Democrat statewide candidates enough to push them over the top. It came close, but fell a little short. Frankly, if JLT had ran a better campaign, it would have probably been enought to put them over the top.

I thought Daniels' continual assurances that he had the race won (which I still think is a strange approach) could hurt turnout Republicans down the line. But Republicans still went to the polls anyway. See my comment on 100% turnout at Pike 28.

Paul K. Ogden said...


Well I try to be a realist. I'm keeping my hopes up about Zoeller. There is nothing more I wanted that a Republican AG, but I want one who is active and aggressive, within the confines of the power granted to that office under the Indiana Code, which are not insubstantial. Steve Carter has been a major disappointment as AG and the source of much rustration for many Republican officeholders and upperlevel appointees who have had to deal with the inactivity of that office.

Zoeller has a chance to take a fresh approach, more aggressive approach to the job. I hope he does more than just continue the status quo.