Statistical analysis of factors in change 2004 - 2008

[CARLHAYDEN]

With near final results in the 2008 Presidential election, let me offer three statistical factors which seem to largely explain the election result.  Specifically, the change amounts to 9.62% (the difference between the bush win in 2004 of 2.46% and the Obama win in 2008 of approximately 7.16%)

First, in the four years the party preferences of American voters moved favorably in the direction of the Democrats and unfavorably for the Republicans.  While the changes are relatively small, it seems to me this accounts for 2.60 of the 9.62. 

Second, the 2004 election had one of the highest turnout rates for Republicans in recent history (partially because of the superb turnout effort by the Republicans that year and partially because they sponsored a number of state votes on banning gay marriage which increased their vote in that year in a number of crucial states) while in 2004 there was essentially no real Republican get out the vote effort, and the Democrats had a better effort than in 2004.  This seems to me to explain 2.66 of the 9.62 differential.

Third, looking at the candidates, Obama was far more favorably viewed by the voters than McCain, and I estimate this accounted for 4.10 of the 9.62 differential.

This leaves 0.18 of the 9.62 unaccounted.

[Alcon]

What kind of statistical calculation are you doing to determine these?

[CARLHAYDEN]

What kind of statistical calculation are you doing to determine these?

Its extremely long, but, what I did was take the basic data for party identification, turnout rates, and candidate selection in the two elections, then held two of the three factors constant to determine an estimate of each factor.

Interestingly enough, Brookings had a seminar a couple of weeks ago in which the panelists opined that Hillary would also have beaten McCain largely as a result of the first two factors I cited.

 

[Alcon]

If you're willing to upload the spreadsheet (or whatever) so we could take a look, I'd appreciate it.  It's hard for me to comment on it without seeing the methodology.

[CARLHAYDEN]

If you're willing to upload the spreadsheet (or whatever) so we could take a look, I'd appreciate it.  It's hard for me to comment on it without seeing the methodology.

As a critical part of the data was generated by a friend, who has not yet published,  so I cannot provide the equations or data at this time (since most of the data is already available, if I provided the equations, then I'll bet your close friend Benjamin could derive the missing data).

[CARLHAYDEN]

If you're willing to upload the spreadsheet (or whatever) so we could take a look, I'd appreciate it.  It's hard for me to comment on it without seeing the methodology.

As a critical part of the data was generated by a friend, who has not yet published,  so I cannot provide the equations or data at this time

When it's published, could you post it?  Again it's kind of impossible to make a substantiative comment without knowing the methods.

(since most of the data is already available, if I provided the equations, then I'll bet your close friend Benjamin could derive the missing data).

haha?

Not sure whether to be flattered, creeped out, confused, or all three!

First, when he publishes, it should be no problem.

Second, not suprised you are confused. 

[Alcon]

Cool.  PM it to me when it's published, por favor

So, what's with the random dropping of my name?    You've piqued my curiosity.  I've posted it here many times, but with that kind of dramatic tone in your post, I assume that's not how you figured it out.