Reply #183: This is interesting

[daleo]

After a recount the Bush vote drops nearly 2% with a consequent pickup by Kerry, yielding a nearly 4% change in the spread between them, and the Miami Herald calls that inconsequential? Talk about burying the real story.

Note: this is just using the 2 full counties and the 60% "sample" of the other one.

Bush pre-recount = 17099/23627 = 71.99%
Kerry pre-recount = 6618/23627 = 28.01%

Bush post-recount = 11985/17089 = 70.13%
Kerry post-recount = 5104/17089 = 29.87%

Bush drop = -1.86%
Kerry pickup = 1.86%

Change in distance between them = 3.71%

It is interesting that the percentage change was small in the two counties that they completed counting (Bush lost .35% in Union and .13% in Lafeyette), but large in the one that they only did a 60% count in (Bush lost 3.86% in Suwannee). It seems like they wanted to bury that fact, by lumping it in with the other two while mentioning small absolute numbers of votes, and not actually working out the percentage changes for the county where the change was large.

It looks like they intentionally soft-pedalled a result that shows significant changes after their recount.

On edit - it also looks like they stopped the count early, so that they could avoid reporting it in a manner that showed an absolute vote change.