Fear of Clowns

Yesterday, the LA Times reported that Kerry nationally trailed Bush 46% to 49%, losing his lead of 2 percentage points from late July: "resident Bush heads into next week's Republican National Convention with voters moving slightly in his direction since July amid signs that Sen. John F. Kerry has been nicked by attacks on his service in Vietnam, a Times poll has found."

A new USA TODAY/CNN/Gallup Poll also reported a slip in Kerry's advantage nationally partially due to "a sharp drop in people who said Kerry's military service would make them more likely to support him."

It's been my contention that the Swift Boat Vets for [Bush]'s effort has likely only served to further polarize the electorate and driven voters away from Bush if anything. To confirm the hypothesis, one would have to show that since the group's efforts have become popularized, states in which Bush already lead had seen an increased lead and states who already supported Kerry had shown a like increase.

The method I used is described in detail below the table. In summary, I found the difference between Bush and Kerry preference numbers on a state level and compared them to the change between the most recent poll and the most recent poll on or prior to August 5 - the day Swift Boat Vets for Truth aired their first ad. The polls had to be conducted by the same entity.

There are 31 states for which poll data is available to make the analysis:

• Kerry gained ground or stayed even in 11 states in which he currently leads: OR, NY, WA, NH, PA, IA, NM, WI, MN, TN, and NV
• Kerry lost ground in 5 states which Kerry currently leads: MD, IL, CT, MI, CA,
• Kerry gained ground in the two states which are currently tied: CO and MO.
• Bush lost ground in 5 states in which he currently leads: FL, VA, NC, IN, OK
• Bush gained ground or stayed even in 8 states in which he currently leads: AR, OH, WV, AZ, SC, KY, GA, AL

In other words, my hypothesis bore out: Bush is generally more ahead in hates which favor him, the same for Kerry. Colorado went from a Bush lead to a tie.

The results of the method of analysis indicate that Kerry may be in trouble in Michigan and California for a total of 72 electoral votes, and that Bush may be in trouble in Florida and Colorado for a total of 27 total electoral votes. California matters lots. If the sharp drop reflected in CA is close to accurate, Kerry should feel quite nervous about the state.

As these numbers only reflect the differences between two polls and that the states which appear to be tossups are close to or within the margins of error, I've put together a page of the same results sorted by different columns. I'll conduct a similar project in the near future if enough state level polls change so as to warrant it. Here are the results sorted by current difference.

// posted by Erik @ 8/26/2004 11:01:00 PM
permalink | email post: