Mystery Pollster on the Ohio reform poll discrepancies

There have been concerns raised in the blogosphre (e.g. me in John Cole's comments somewhere) about the marked difference between the polling and the voting for the recent Ohio ballot initiatives. There was one Republican-sponsored measure and four election-reform measures from the liberal RON. The latter did much worse than predicted. Is this evidence of dirty tricks? Mystery Pollster analyzes the poll and is highly skeptical of that claim, while suggesting a way to investigate.