Friday, January 05, 2007

Of Stats and Thefts

From Congressional Quarterly:

Of the 202 Republicans sworn in Thursday as members of 110th Congress, 15 maintained GOP control of their seats by margins of just 3 percentage points or less. On the other side of the aisle, just two of the 233 members of the new Democratic majority were winners of contests in which they retained their party’s control by similarly razor-thin margins.

This means that 7.426% of republicans were elected by 3% or less. Only 0.858% of Dems won by 3% or less. Republicans outperformed Dems by 11.5 times in these sorts of races. That is highly improbable statistically. I can think of two reasons:
  1. Congressional Quarterly cherry-picked 3%. Often sports announcers say that the team has 8 wins in its last nine games, or something similar. sports fans know that this means they lost the game, ten game ago. they are 8 for the last 9 and also 8 for the last 10. perhaps had the congressional spread number been 6% or 1%, or something else, the disparity would have been minor (and as a result not newsworthy).
  2. This is indicative of electoral fraud. If republicans continued their corrupt machinations, the election would look like this. The procedure for cleverly stealing an election (i assume) is that you first count the votes in secret. If your guy is ahead, you count the votes publicly and shake her hand and announce her the winner. If your guy is behind, you find out how many votes he needs to win and then make him a narrow victor. i imagine that the more you need to fudge the results the more likely you are to be caught. all this to say that a small set of the 15 republicans who won by small margins, probably should have lost by a small margin and that i am highly skeptical of the republican behavior. For a more specific example check out the story in FL-13.
If the numbers were not cherry-picked then they may very well indicate fraud, i haven't looked at them enough to have a substantial argument about whether they were presented well in the CQ article.


At 1/05/2007 , Anonymous Aharon said...

So out of 17 contest won by 3% points or less, 15 (88%) of the contests were won by Republicans, and 2 (12%) were won by Democrats. Help me with my statistics, but from what I recall, so long as the sample size is below 30, drawing any sensational conclusions is rather suspect. Are you saying that when races are won by less than 3%, the division of wins should be more subject to random luck, and so fall closer to a 50/50 split? In at least one of these district races, Democrats performed very well in what were overwhelmingly Republican districts -- and lot by narrow margins. (I'm thinking of the 2nd district in Ohio, Hamilton Country where the detestable Jean Schmidt-R won.)

At 1/08/2007 , Blogger Ruby K said...

I'm not sure when they added this, but CQ did make this important clarification... because I believe I was on a < 3 point ride myself:

larification: The contests analyzed in this story are those in which the incumbent party held onto a seat, but by a margin of 3 percentage points or less. on Dec. 4 published a roundup of House contests in which Democrats took over seats that had been held by Republicans — including eight in which the Democrats won by 3 points or less.

At 1/10/2007 , Blogger ZT said...

i'd think a nearly 50/50 break would be expected assuming the number of defended seats was normalized.


Post a Comment

Subscribe to Post Comments [Atom]

Links to this post:

Create a Link

<< Home