At a Federalist Society meeting regarding the recall that I participated in last month, Chapman law professor John Eastman raised the question of greater intentional voting by racial minorities. The suggestion was that members of minority groups might turn out to vote for certain minority candidates lower down the ballot, but fail to vote the top of the ticket. I indicated there that I did not know of any studies addressing the issue of intentional undervoting by minorities.
I have now come across a study that has examined this question and found that African-Americans do tend to undervote more than whites, but the difference is very small, a few tenths of a percentage point. See Michael Tomz and Robert P. Van Houweling, How Does Voting Equipment Affect the Racial Gap in Voided Ballots, 47 American Journal of Political Science 46, 57 (2003). The article is available to those whose libraries have an electronic subscription here.
Intentional undervoting by minorities certainly could not explain why there were about 11% of voters in Los Angeles that failed to cast a valid vote on the first part of the recall, compared to less than 1% in Alameda county, using touch screen machines, or the average 2.6% across the state (all of these are based on preliminary figures). (More on those figures here and here.)
And while we are on the subject of punch card errors, I am still waiting to hear from Mickey Kaus on his criticisms of Henry Brady. By now, I think Kaus would have to admit, Brady’s position has been vindicated. Kaus told me he was going to blog about this soon after the recall, but I don’t think he has.
UPDATE: John Eastman replies here. In response, let me note that I think there’s no question now that (1) touch screen voting reduces intentional undervoting and (2) that minority preferences cannot explain the voting disparities between punch card counties and non-punch card counties. The statistics comparing undervoting and exit polls across to voting methods, if they hold up on final review of numbers (we don’t have final numbers yet), will lead to the unassailable conclusion that punch card errors are responsible for this disparity across counties. The disparity is huge even taking touch screens out of the picture and just comparing to other voting technologies.