|So here are Obama and Thompson's absolute performances in the city.
Obama's performance we already knew about, but a few striking aspects of Thompson's performance:
- Upper East Siders lurve them some Bloombo.
- Whites in the Bronx voted for Bloomberg.
- Hispanics voted mostly for Thompson (though not to the levels they voted for Freddy Ferrer, I would posit).
- Blacks stayed strongly loyal to Thompson, with slight drop-offs visible in Brooklyn and East Queens.
- Staten Island stayed Staten Island.
More interestingly, here is a comparison of Obama and Thompson's absolute performances. A more intense blue indicates a stronger Obama performance; a deeper red indicates a stronger Thompson performance.
Obviously, most of the map is some shade of blue, since Obama's margin was 63.65% greater than Thompson's. Even given this, there are still two visible clusters of red in Brooklyn: Williamsburg and Borough Park. Thompson still lost these precincts by a decent margin, but he improved over Obama despite the tide moving 64% in the other direction. I took this as evidence of the Hasidic Jewish community's growing dislike of Bloomberg, which had been mentioned a few times before the election.
On the flipside, as you would expect, Manhattan and Brooklyn Heights are home to the Obama-Bloomberg voters, especially on the Upper East and West Sides, in Midtown, and down in the Financial District.
The lighter shades of blue are in East Queens and Central Brooklyn - mostly majority-black precincts that went strongly for both.
Another interesting cluster of these, though, is on the South Shore of Staten Island; Thompson's performance didn't fall all that much off from Obama's (admittedly already weak) performance there. It seems there, though, that the voters are more reflexively Republican than those in Southern Brooklyn, where Obama seemed to be a particularly bad fit. (backup evidence: Stephen Cymbrowitz and Carl Kruger are elected from those areas in Brooklyn. Southern Staten Island elects two Republicans to the Assembly/Senate, Lou Tobacco and Andrew Lanza).
Alternatively, this can be shown in graph form. Obama's margin on the x-axis; Thompson's on the y-axis.
Now any monkey could have told you generally a stronger Obama performance is correlated with a stronger Thompson performance, but the exceptions to that general rule are evident here as well. The large cluster of green on the bottom right are those previously mentioned Manhattan precincts, while the dispersed red dots towards the middle and lower left are the Brooklyn precincts in which Thompson actually improved. (Incidentally, yellow represents Staten Island, orange for the Bronx, and blue for Queens). The bright green line is the even-performance line.
Now two more maps of interest, each candidate's performance relative to their citywide cumulative total (Obama first, then Thompson).
Obama did well throughout the city, a strong Obama performance was the norm. You don't see many places darker than light blue, simply because you can't get more than 100% of the vote! Where Obama underperformed, he really underperformed. You see this in Suburban Queens and also Middle Village/Maspeth, and of course Southern Brooklyn and Staten Island.
Thompson's performance really varied much more. He overperformed in many places, and underperformed in many places as well; these deviations are of much more equal magnitude. Again, as we've realized, Thompson's weakest area was the Upper East Side.
So all this poses the question, what happened?
Well, in three words, Thompson's turnout problem.
Conventional wisdom dictates that minorities (who are actually a majority in NYC) turn out less in general. While this may or may not be true, I normalized and considered 2009 turnout as a percentage of 2008 turnout.
The results, first at the precinct level. The same color codes apply as before for borough. Turnout as a percentage of 2008 turnout is expressed on the x-axis; the Thompson-Bloomberg margin on the y-axis. (Turnout dropped most in the Bronx, in case you're wondering.)
You see a general effect of center left to lower right, suggesting stronger Bloomberg performances being correlated with greater turnout. This effect is even more pronounced when we consolidate to an assembly district level:
It's not pretty. For you stats geeks out there, the correlation on that bad boy is -0.77. Ouch.
Incidentally, that one AD with the lowest drop-off? None other than Dov Hikind's 48th AD. Turnout there was lower there in 2008, but those that voted in 2008 were most likely to have voted again in 2009.
As a parting thought, take solace (or anguish) in this: if turnout had dropped to the 44% figure I mentioned at the start equally across the city, Thompson would have won, 49.16% to 48.00%.
Having arrived at Brooklyn, I'm going to sleep. I realize I owe you a proposed set of New York Senate districts. I just need to write the diary. I'll get around to it...eventually.