When I was young, and modern liberalism dominated American politics, I recall periodically running across voting conspiracy kooks, usually conservatives, who were absolutely convinced that American elections were rigged. This type never completely goes away, but their numbers can rise and fall.

Of late, conspiracy theories about voting have been getting out of hand. A major problem is that a large element of the American political left has embarked on a campaign to convince Americans that their elections are fraudulently manipulated at the most fundamental level, and many people in positions of responsibility have been unwilling to throw water on the fire, thinking somehow that they can harness that energy for electoral victory.

A good example of this left-wing election paranoia is found at a site called, sadly enough, The Brad Blog. This particular post was published shortly after November's election, when Ohio voters overwhelmingly rejected a package of four proposed amendments to the Ohio Constitution, placed on the ballot by a coalition of left wing called "Reform Ohio Now." Brad Blog (relying heavily on the "analysis" of Bob Fritrakis, a leftwing Columbus political gadfly for many years) finds obvious fraud in the fact that the results did not comport with - indeed were vastly different than - the results of the Columbus Dispatch mail poll, a poll which traditionally is quite accurate. Furthermore, notes Brad Blog, it was accurate with regard to a 5th issue on the ballot, a state bond issue.

But as usual, there are in fact very average, normal reasons why the pre-election poll was so off the final results. The Dispatch poll (subscription required) was taken over a very long time, from Oct. 24 through Nov. 3, with election day falling several days further out, on November 8, always a potential problem. Other issues, identified by the Mystery Pollster Mark Blumenthal (a Democrat, for what it's worth), may also have contributed to the poor performance of the Dispatch poll. These include changes in the Dispatch methodolgy from past polls, including adding an "undecided" option; the fact that the Dispatch had not previously done an off-year poll featuring only ballot issues; the fact that the Dispatch poll itself showed support trending downward as its responses came closer to election day; a substantially lower response rate than in the past; the length of the ballot issues, which were not reproduced in full in the poll; and the fact that the Dispatch poll has always been least accurate at predicting statewide ballot issues.

Now Democratic pollster Celinda Lake (Lake Research Partners) has released a report on why voters in Ohio (and California) rejected redistricting and other initiatives, and it should (but surely won't) put the final nails in this little leftist conspiracy theory. You can link to a pdf file of the Lake report here.

Among other findings, Lake found that 75% of Ohio voters made up their minds close to election day (see Table 24), and 73% of them voted "no" on the redistricting reforms. Also, newspapers were the single most important source of information on the intiatives for Ohio voters, and especially for voters deciding in the last week (79%). This is relevant because eighty one of 104 Ohio newspaper editorials on the intitiatives urged a "no" vote.

We are living in bizarre times, indeed, when so many people insist that, though there is no direct evidence of fraud, the fact that exit polls and pre-election polls did produce corresponding results conclusively demonstrates that the actual election day results are invalid. But as usual, and as we see in the Ohio vote, there is an easy explanation.