Public Policy Polling (a robo-call polling service) has released a poll showing KBH leading Perry by 25 points (56-31%) among likely GOP primary voters.
http://www.publicpolicypolling.com/pdf/P...as_224.pdf
BTW, PPP defended robo-call polling as being accurate by pointing to their very accurate results the day before Texas GOP primary where McCain took 51% and Huckabee 38% (their poll from 3/3 showed 50-38):
http://www.surveyusa.com/wp-content/uplo...030508.JPG
Here are my observations regarding automated phone polling vs personal phone polling techniques... this is completely non-scientific, but I feel fairly good about my conclusions:
Automated Polling Results:
Faster
More likely to result in non-participation
... as a result, more reliant on voters with STRONG opinions at the time of the poll
Thus, less representative of the electorate as a whole (with exception -- read on)
Personal Polling Results:
Slower
More likely to result in participation even among undecideds who will still show up on election day
Thus more representative of the electorate as a whole (with exception)
Exceptions:
As election day approaches, the automated polling results are going to be more precise. Further, if it is a tight race and it's going to fall to undecideds, as election day approaches, the personal polling results will be less precise -- because they're always a couple days old.
So, as election day draws near, the results from automated and personal methods will converge in precision; if it's a tight race, the automatic methods may even surpass the personal methods in the last couple days because people are still making up their minds and the automatic polling methods are better able to keep up with the dynamics such a race at that late stage.
But right now, more than a year out from the primary, I wouldn't put much stock in an automated phone poll. (One only need look at the undecideds in today's release to figure that out... only 13% of likely GOP voters undecided at more than a year out is absurdly -- and improbably -- low.)
As for PPP's defense of automated polling, they point to their numbers from the day before the 2008 primary as evidence that automated polling works. That doesn't say anything, however, about the relative value of automated polling months (in this case 12 months) prior to an election though.
I do think KBH will beat the tar out of RP come primary day, but this poll does little to reinforce that belief.
Regards,
Mike Duncan