The Election is Upon Us

I've found myself more interested in this campaign season than any in a long time. I find the avalanche of public opinion polls and TV talk shows and gasping-breath news reports and cynical pseudo-news and unbelievable television ads to be overwhelming and massively entertaining.

I was just looking at the latest news articles about late movement in public opinion polls.
Bloomberg
ABC News
Washington Post – ABC News Poll

Some of the questions that are included in the polling are pretty interesting and they got me to thinking.

I went in and voted early this past week. When I went in to vote, I intended to cast votes in only a couple of races - governor, US representative, a local bond referendum, and a few others. Those are the only ones where I really know the candidates and issues and have a strong opinion.

A funny thing happened to me in the voting booth and that relates to my interest in the poll results. If you had asked me several of those poll questions before I stepped foot in the voting booth and then asked me after I stepped out, I would have given different answers.

When I first stepped in that booth I scrolled to the contests in which I planned to vote and skipped the others. Then I scrolled through again, just to make sure I didn't miss anything and didn't make any inadvertent mistakes. As I was scrolling along it really struck me that I had an opportunity and in fact a responsibility to make a choice and send a message. And that is what I decided to do. I wonder how many others will find that they feel stronger about issues when confronted with the reality of a ballot in front of them?

So, if I were going to answer some of these interesting poll questions:

1) What is the most important issue facing the country today? Iraq
2) Is your vote in the current elections influenced by your opinion of George W. Bush? Yes
3) Is this election a referendum on George W. Bush? Yes
4) Is the country headed in the right direction? No

All of those answers are the same today as they were a week or more ago with the exception of question 3. If you had asked me a week ago if I thought that on a national scale, this was a referendum on George W. Bush, I would have said yes. However, if you had asked me if I would cast my votes specifically to send a message about the leadership of George W. Bush, I would have said no. Before I stepped out of that booth last week, I changed my mind and cast every vote as a reflection of my opinion on the Bush presidency. I thought through those four questions above and made my decision and cast my votes. Is that an intelligent way to vote? I don’t know. You can criticize me all you want, but that is what I did.

It is intriguing for me to see those same four questions in one form or another listed in a number of the polls that came out today.

It will be interesting to see if the national opinion machine is really detecting an anti-Bush tsunami or just another weak Democratic Party attempt to take on a superior political machine. We'll start to know Tuesday night.

Because I have been so interested in all of this polling information, I started following the House of Representative projection polls. http://www.electoral-vote.com/ publishes a web site that uses all of the available public polls to project what will happen in the Senate and House races. Because new polls are released daily, the overall projections on that site also change daily. It seems to me that the Senate is pretty close to a tossup. It could end up anywhere from a Republican to Democratic majority with no more than a majority of 4 (meaning 52 to 48 for one side). The House of Representatives on the other hand looks much more predictable. The site mentioned above lets you download a spreadsheet with all of the poll results for each of the house and senate races where there are available public polls.

So, being a nerd, I decided to do my own analysis/projections based on the available polls. So, for each of the last three days, I have taken all of the house poll data from that site and plugged it into a couple of projection models. For each day, I assume that the closer we get to election-day, the lower the chance of movement in the polls (I assume greater accuracy). Further, I have one model that more heavily favors incumbents - it gives them the benefit of the doubt in many close races unless there is a lot of recent data to override that default assumption. The other model just takes the most recent poll and assumes that it is the best predictor. Well, whatever, I've been plugging in my numbers for the past three days and came up with these results.

TODAY - model that favors incumbents
Democrats - 219
Republicans - 206
Tossups - 10

TODAY - model that does not assume an advantage for incumbents
Democrats - 223
Republicans - 188
Tossups - 24

YESTERDAY - incumbent advantage model
Democrats - 217
Republicans - 206
Tossups - 12

YESTERDAY - no incumbent advantage
Democrats - 221
Republicans - 185
Tossups - 29

FRIDAY - incumbent advantage model
Democrats - 217
Republicans - 208
Tossups - 10

FRIDAY - no incumbent advantage
Democrats - 217
Republicans - 179
Tossups - 39


So, based on this, I'm willing to make a few generalizations.

1) It looks to me like the Democrats will take control of the House. That control could be as low as a razor thin 2 or 3. Or that control might be a pretty commanding 40+. Of course that assumes that these polls are an accurate reflection of what will happen on Tuesday. There are lots of flaws in polls and there are even more flaws in my predictor modeling (I put no weight on projected voter polls versus other general polls and I assume in some models that the most recent poll is the most accurate without examining the actual polling methods. In another of the models, I assume that a poll from 3 moths ago has as much value as a poll from last week. I don't claim that I have a good model at this point. I'm just trying to do some quick analysis.)
2) If it is possible to spot a trend in these few samples, it seems to me that the Republicans have been closing the gap in the past few days, but probably not fast enough.
3) In each of the models the Democrats seem to have a very solid number of seats around 217 to 223. It is Republican seats that seem to vary between being a tossup or a Republican projection. If those tossups split evenly between the parties, then the Democrats will have a majority around 30.
4) It looks to me like it would take a miracle for the Republicans to end up with control and it would be a worthy accomplishment for the Republicans to keep it really close. The most likely outcome is a strong Democratic majority in the House.

It is the question of what way these tossups will go that makes my experience in the voting booth from last week relevant. I think it is going to come down to the prevailing mood in the country and just how angry people are over Iraq and George Bush. I believe this vote is a referendum on George W. Bush and our policies in Iraq.