homer tries to vote

Fortunately, some of the opportunity for stealing the election has changed since 2004. In the last two elections a lot of Republicans were voted out of office at the state level. For example, whereas Ohio had a Republican governor and secretary of state in 2004, both of those posts are now held by Democrats (and the Democratic secretary of state is making sure there are enough voting booths in the college towns and poor precincts that had multi-hour lines in 2004).

Of the 12 most important battleground states, the Secretary of State is now a Democrat in 9 of them, and the governor is now a Democrat in 9 of them:

State: Secretary of State/Governor

Colorado: Republican/Democrat
Florida: Republican/Republican
Iowa: Democrat/Democrat
Michigan: Republican/Democrat
Minnesota: Democrat/Republican
Nevada: Democrat/Republican
New Hampshire: Democrat/Democrat
New Mexico: Democrat/Democrat
Ohio: Democrat/Democrat
Pennsylvania: Democrat/Democrat
Virginia: Democrat/Democrat
Wisconsin: Democrat/Democrat

Also, I hear a lot of friends continuing to express concern over what is know in political circles as the “Bradley Effect” – i.e., that there is latent racism that isn’t being picked up by the polls because many people who wouldn’t vote for Obama because of racism wouldn’t admit it to a pollster. (Named for Tom Bradley, an African-American who lost the 1982 California governor's race despite being ahead in voter polls.)

Harvard political scientist Dan Harvey just came out with a comprehensive study that concludes that since the mid-‘90’s the Bradley Effect has disappeared. This study is a must-read for anyone really interested in the subject (for anyone else, here is a summary). Fivethirtyeight.com undertook a particularly good analysis of the 2008 Democratic primary results that found that, on average, Obama actually out-performed the last set of polls before the primaries. Other articles debunking the Bradley Effect are here and here.

If there were a Bradley Efffect, it would suggest that polls overstate Obama’s support. The opposite concern arises from the cell-phone effect. This is the idea that a lot of young voters, who overwhelmingly support Obama, don’t have landline phones and communicate only with cell phones that aren’t sampled in the polls. There is some validity to this concern. Most of the major pollsters have been trying to compensate for that effect, with uncertain effectiveness. (Others don’t even bother to try.) Best efforts to quantify the effect (summarized here) suggest it may result in support for Obama being understated by as much as 1 or 2 percent.

In any event, let’s make the sure the race isn’t close enough where any of this matters.