OK, back to the wonderful world of Polling minutiae! Thanks to the Libertarian Lass for getting me to check “Survey USA’s” pre-election poll on last year’s Half Cent sales tax item, Prop. D. Summarized above is their Oct. 7, 2010 poll, the last one I can find. You can see the Crosstabs and Analysis at this link. Having seen polling guru John Nienstedt’s presentation, I believe there was a poll showing the race “too close to call,” but it may have been done for the sales tax supporters. Survey USA partners with Channel 10 on various local subjects of interest. If there was a later Survey USA release on Prop. D, please let me know. And Kudos to Channel 10 for taking the lead on useful Polling !
If you pro-rate the Undecideds listed above, Survey USA had Proposition D losing 62% to 38%. That is exactly what happened, as the Libertarian Lass and her cohorts drove home the No-on-D message with what President Kennedy would call “Great Vigor.”
Polling minutiae. How I love it!
Comments 1
John Nienstedt’s polling is superior to SurveyUSA, in that he figures out who the “likely voters” are. But the SurveyUSA data can be useful in making similar if cruder and less accurate projections.
Perhaps the single highest demographic indicator of “likely voters” is age (experts can provide more guidance on this factor vs. others). And the SurveyUSA mayoral poll includes this demographic — though they don’t use it in their analysis.
I’ll shortly post on Rostra an article I wrote on my blog on this topic — which is an expansion of an earlier comment I made on Rostra on the mayoral poll. But the bottom line is that DeMaio’s 22% lead is understated, given that he polls better with the older, voting crowd. From my blog post:
“While DeMaio gets 22% of the poll, he gets 28% of the age 65+ voters, 27% of the 50-64 voters, 22% of the 35-49 voters and only 8% of the flaky 18-34 voters. And we all know who shows up to vote — especially in primaries.”
_______
Your reasoning on this is valid, since those age groups are vital, especially in a primary election. In the Oct. 2010 poll, however, Survey USA did screen for “likely” voters, cutting the 900 overall interviews down to “577 likely or actual voters.” Actual voters were folks who had already balloted by mail when interviewed. But I can’t agree with you that voters aged 18 to 34 are ‘flaky’! (Jim Sills)