Updated and revised, 9:35 AM
Wow, this year’s ballot initiative campaign gets stranger and stranger. The pro-Ref C Dead Governors have declared that their beloved tax increase is winning by 1 to 9 points in the “polls.” Their source? I quote directly:
No, we aren’t going to tell you who did these polls, and no, we didn’t just make this up.
Trust them… right. Trust but verify. Let’s be intellectually honest for a moment and admit we don’t know exactly where public sentiment is on Ref C & D. If the pro-tax increase crowd had a credible survey showing momentum on their side, the results would be trumpeted on all the major news pages. Their silence speaks in abundance.
My gut tells me the election will be close and the ballot measures will lose. But I can only trust my gut so far, and I won’t mislead anyone by saying my inclinations are based in hard, empirical evidence. Besides, one must consider the old cliche: “The only poll that counts is the one on Election Day.”
Meanwhile, the folks at Mile High Delphi have run a statistical model suggesting that Ref C has a 28.6% (or 2-in-7) chance of winning and would lose 53-47 if ballots were cast today.
I have no way of knowing if their insights into the election results are more credible than the Dead Governors’ (if anyone out there is qualified and wants to critique the statistical analysis, please share), but at least their method and sources are transparent.
We’ve had some insights already on the model.
Matt ran the model on the 2003 elections, here are the results:
Ref A (30+(.2 * 18) = 33.6 Actual vote=33
Amendment 32 (14 +(.2 * 38) = 21.6 Actual vote=22
Amendment 33 (19 + (.2 * 8) = 20.6 Actual vote=19
Overall the average error was .6%.
The problem is that since we only have complete information for 5 elections, the standard deviation is larger than we like.
However, I also calculated that the no on c allies outspent the Vote Yes on C and D campaign by about $200K last month. I’ve posted on it at the deadguvs, polstate and mile high delphi, no one (including the msm) seems to even have a clue that the vote yes side is being outspent.