I have to agree with Thad Hall’s posting on California’s switch to paper ballots.

Residual (under and over) voting rates are highest on “paper” balloting systems. It will be an interesting research question (one which I hope to be positioned to answer, subject to funding) if “absentee” paper balloters show higher residual voting rates than “day of election” paper balloters. And, from a research perspective, California provides me the perfect opportunity to demonstrate lower residual voting rates on DREs, since the DREs will be in operation–albeit only for disabled voters–in the same precincts.

Thad noted the ultimate “cool” outcome for political scientists (warning to elections officials: when someone like Thad says something is “cool”, it probably means there will almost certainly be voting anomalies for us to study)–but he failed to notice another “cool” requirement.

The DRE paper records will have to be 100% manually recounted–but not the paper ballots! But what happens when, with almost 100% certainty, some of those “recounts” don’t match the DRE total? Will you believe the human-based manual recount or the machine total? Will you re-run the count? Re-print the output?

What I’m less sure about Thad’s claim that the institutional format–the absentee ballot–causes lower confidence among these voters. Perhaps they choose absentee balloting because they have lower confidence in elections overall, but the most confidence in the absentee ballot. Or maybe it is something else about those voters that make them less confident. (I suspect a paper draft is soon to be winging my way.)