With the preliminary numbers pointing in favor of a large disparity in punch card vote errors, will he retract his criticism of Dr. Brady?
Along these lines, I received the following e-mail from Matt Wall:

it was my note to a pal of mine at Slate and my comments on my blog that started the thread at Kaus’ blog at Slate, I think, and I read your follow-ups on your site. I wanted to give you very anecdotal evidence to supplement the statistical evidence.
(One has to bear in mind that the very counties that had punch cards were among those more likely to vote against Schwarzenegger. I’m not sure what if anything this says about the results. But I digress)
In my county, at my polling place, I can personally attest the punch card format was at fault for erroneous voting. I’ve worked elections in several stats, using lever-machines, optical scanner fill-ins, and good old paper check-off ballots. And now that I’m in California, I arrived at the high technology of 1888, the punch card.
My precinct was educated, upper middle class to wealthy, native English speakers with only a few exceptions, and we registered only 123 polling-place voted ballots (144 voted absentee!) with the vote-o-matics. We had two dozen spoiled ballots! And those are just the voters who asked! I have to assume there were voters who messed up and either didn’t notice or didn’t ask (we tell voters if they mess up they can get another ballot, but it’s not always clear the message gets across).
If we had 25 of 123 voters _catch_ their mistakes, then, having all the linguistic and voter skills necessary to presumably successfully register their votes…how many didn’t? How many in other counties?
I personally assisted two voters who were having difficulty trying to cast their votes for Schwarzenegger, one of whom had the punch OVER Schwartzman’s name and was asking “this is going to vote for Schwarzenegger, right?” This was an apparently well-spoken thirty-something white woman. (The other one simply had no clue and asked to be shown where to register his vote for Arnold.)
We rotated assistance duties, and while I did not ask my fellow election board members specifics about names (inappropriate to do so), we did confer at one point about making sure voters were able to find the name of their candidate, and it was clear the other members of our election board had had the same experience with voters being unable to clearly line up the card with the appropriate punch.
I know it’s not rocket science, etc., but it’s clear that cognitively asking people to find one name amidst 135 in fairly small type and line it up under plexiglass with a tiny punch tool is bound to produce errors. And again, anecdotally, in my experience this kind of problem did not exist with lever-machines or boxes. (It was also a problem with the optical scanner sheets, but that was in an Ohio election over 20 years ago.)
In fact, I suggest that were not the local citizen-election officials as diligent as we were (pat self on back) there probably would’ve been many more Schwartzman votes.
It’s all moot, of course, as you legal types say from time to time, since clearly the effect did not affect the results of this particular election. But the evidence of my own participation to me is quite strong that the punch card machine effect is very real.