Friday, August 18, 2006

conflicted Pennsylvania voters?

Let's consider the conflicted Pennsylvania voter.. It's something of a mystery why Pennsylvania votes for diametrically opposed candidates, sometimes during the same election season. One example is the 2000 general election where Gore won Pennsylvania over Bush, but at the same time Santorum won over Klink. Yes, I know people have strong opinions about those particular candidates, but it's just one example of a larger pattern. There must be a sizable core of voters who never vote for one party consistently. A sign of independence for sure, but its hard to figure out what that group is really voting for. So here is a map from the 2000 election with a compiled statistic of the difference between the percentage voting for Bush in the Presidential election and the percentage voting for Santorum in the Senate race. Note I have ignored all the other candidates in this computation. If people just voted for candidates in the same party, then the difference in the two races would be near zero, which are the districts in white on this map. The few green areas are where Bush outpolled Santorum and shades of red/orange/yellow indicated areas where Santorum outpolled Bush. Now I know that the major statewide campaigns going on currently have modeled this to death internally. I am not sure that Allegheny County really helps explain the statewide results much anyway. Note that this map does not tell you anything about which candidate did better in different districts, the colors only represent the difference (in percentage points) between the two Republican candidates..

If I were to obsess on this a bit (too late I suppose) I would run the results through a simple regression with income, age and some other things to see if you can see what is driving this 'discrepancy'. I have not looked too hard, but I would not be surprised if there is not something in the public domain describing the mythical Pennsylvania swing voter. Anyone know of anything out there?

but if someone out there is interested in this some more. What I would need to do this for real are statewide results by voting district for these particular races or others. I do have statewide demographics by voting district. If someone wanted to match up their election returns data with this, let me know. I am sure there is a paper in it somehow.


Blogger Amos_thePokerCat said...

Ya, this is something I breifly ruminate over from time to time, but I really only worry about it when the phenomena cuts against my own muddled right of center anarcho-liberarian nonpracticing RC views.

Basicly, I write it off as party vs. ethic/religion politics. An Italian Catholic is always going to get the tribal Kennedy edge in SW PA. (Might that be Bloomfield as one of the dark orange dots?) Of course, at the tribal level, this race pits the Italians vs. the Irish, with no national Rorschach candidates to compare against.

A state wide, if only county by county, map just of this quick and dirty data, would also be interesting. Ya, it would be interesting to run that voting district data. Don't know if I have the time, or energy. I know I don't have the tools to generate the fancy maps, at least easily.

Tuesday, August 22, 2006 12:28:00 PM  
Blogger C. Briem said...

right of center anarcho-libertarian nonpracticing RC views

wow.. what a taxonomy. You need to come up with something pithier than that to describe yourself. Something like "soccer moms" or "nascar dads" or else the politicos will not know how to target you.

but ethnicity exlaining this? I dunno. Bloomfield is at most a light orange. I am not sure that would hold up across the state but as I said, I have not looked that deeply.

Tuesday, August 22, 2006 6:17:00 PM  

Post a Comment

<< Home