Reply #50: I think chorti's ideas make a lot of sense

[seaclyr]

including the idea of using a neighborhood state for comparison. What I'm afraid of though is that there may be very few "clean" neighborhood states out there. It's just hard to find states where the numbers can be completely trusted. Look at the following, for example, where two measures of possible irregularities are listed for potential comparison states. The first, labeled "exit" is the percentage gain (Bush over Kerry) going from exit polls to "actual" votes and the second, labeled B+%, is percentage gain by Bush over Kerry in 2004, relative to Bush vs. Gore in 2000.

______PA___NY
exit    6.6    7.6
B+%   1.69  6.69
Looking at it this way, I'm not sure NY looks any better than PA, it may even be worse.

The same goes for AR and LA

_____AR___LA
exit   2.2    4.2
B+%   4.3   6.83

for TN and KY

_____TN___KY
exit   -3      2
B+%   9.4   4.73

and for NM and AZ

_____NM___AZ
exit   3.6     5.2
B+%  0.85   4.18

For similar reasons, I don't believe one can completely trust IN or MS, either.


So I think "clean" states may be hard to find. Doesn't mean to say one gives up on the analysis, but one may wind up comparing "awful" with just plain "bad".