Was California’s Prop. 8 (Anti-Same-Sex-Marriage) Vote So Close Partly Because of Voter Confusion?

This L.A. Times op-ed reports:

It’s true that the official election results — 52% to 48% — appeared quite close. But the truth is more complicated. The data we analyzed show that the No on 8 campaign benefitted from voter confusion.

Polling suggests that half a million people who opposed same-sex marriage mistakenly voted against the proposition. They were confused by the idea that a “no” vote was actually a vote for gay marriage. This “wrong-way voting” affected both sides, but overwhelmingly it helped the “no” side. Our analysis suggests that the division among California voters on same-sex marriage at the time of Proposition 8 was actually 54% to 46% — not so close.

Of course, in many contexts, the difference between 54%-46% and 52%-48% is not great, but my sense is that in elections a 54%-46% result is seen as a much bigger gulf than 52%-48%. The underlying report is here; see especially PDF pp. 105-07. Thanks to Prof. Rick Hasen (Election Law Blog) for the pointer.

UPDATE: Some comments seem to view this as evidence of general voter stupidity, or in particular of conservative voter stupidity (since the errors seemed to be especially prevalent among voters taking what is seen as the more conservative view, that opposed to recognizing same-sex marriage). I’d be cautious about taking that view; the survey reports that only 11% of the voters cast a vote based on their misunderstanding of the proposition, and of those 57% were actually opposed to same-sex marriage (but voted in favor of same-sex marriage) and 43% actually supported same-sex marriage (but voted against same-sex marriage). So the data doesn’t tell us much about voters generally, or conservative voters generally. At most it might be an argument against popular votes on issues because even a small fraction of voter error can skew things dramatically — a potentially important point (though one subject to the obvious counterarguments) but not a general indictment of voter intelligence.

FURTHER UPDATE: Prof. Rick Hasen (Election Law Blog) has a follow-up; the bottom line:

My guess is if that there would have been considerably less confusion if the question was written to say: “Shall the California Constitution be changed to provide that only marriage between a man and a woman is valid and recognized in California?” This appears consistent with the full text of the statute. Indeed, the proponents sued Jerry Brown over the use of the word “eliminates” and lost. I’ve been involved with enough ballot measure litigation to know that the attorney general should not be entrusted with the responsibility of providing a neutral ballot title and summary, because he or she often has a political axe to grind and can be subject to pressure over ballot wording.

The error rate says a lot about ballot wording, but very little about voter intelligence.