Tuesday, June 12, 2012

Is the PPP Poll Beginning to Resemble Zogby?

I was very excited about the prospects of Ted Cruz making a runoff with David Dewhurst in the GOP primary in Texas as almost every poll showed Cruz closing the gap to 9 or 10 points, with Dewhurst well under 50%. Then literally days before the runoff, here comes a poll from the Public Polling Policy that made me sick to my stomach. From the PPP:  "David Dewhurst leads Ted Cruz 46-29 in the Texas Republican Senate primary and the main question on Tuesday will be whether he can sneak over 50% and avoid a runoff...5% remain undecided and if Dewhurst can pick up enough of those folks it might get him over the finish line."  So, with every other poll showing Dewhurst well under the 50% mark and Cruz only 10 points behind, here comes PPP saying Dewhurst was up by 17 and he had the momentum to maybe win without a runoff.  So, what were the final results.  Dewhurst had just over 45% and Cruz had almost 35%.  A ten point gap, just like all the other polls had, with Dewhurst well under the 50% mark.  I thought at the time, how could the PPP poll be so far off and so much of an outlier from the other polls?

Then there is the case of Wisconsin.  From the Public Polling Policy [PPP] web site they announced days before the recall election in Wisconsin their final poll:  "PPP's final poll on the Wisconsin recall finds Scott Walker ahead, but also a race that's tightening. Walker leads Tom Barrett 50-47."  They also, found if there was a heavy turnout nearing 2008 levels, Barrett actually led Scott Walker 50-49.

I was really worried about this poll, even though almost every other poll had Walker up between 5-9 points during this same time period.  Then I thought about how far off the PPP poll was on the Ted Cruz race just days before the election, and that gave me comfort that PPP is just a bogus poll and I shouldn't take to much stock into it.  As it turned out, I was right not to be concerned.  In spite of a huge turnout, the kind of turnout that PPP said would favor a Barrett win, Scott Walker won impressively by 7 points; that was right in line with the other polls and right in line with RCP projections.

So, in two of the biggest races so far this year, the PPP poll has been highly inaccurate [and both times  inaccurate against the most conservative in the race].  A coincidence?

If you go to the PPP web site you will find there proud pronouncement that The Wall Street Journal has rated them as one of the top polling sites. I thought that was curious, how inaccurate they have been the last couple of years.  Then I looked at the date of the Wall Street Journal's rating: November of 2008.   They may have been very accurate then, but how accurate have been since then?  If you look at the state polls they are doing now, they all seem to lean in favor [compared to other polls] of the Democrats.

I remember the Zogby poll at one time was rated highly, and then it fell to disfavor when they have since been complete outliers to most other pollsters.   Some have even questioned their polling methods.  You hardly ever see a Zogby poll mentioned anymore by anyone; and the Real Clear Politics average does not include them in their list of polls anymore.

I am not saying that I think Real Clear Politics should drop the PPP poll completely from their site, as PPP does much of the state polling others won't, but I am saying I think when they are complete outliers, like they were in the Ted Cruz primary race and the Scott Walker recall election, they should not be included as they have been proven to be so far off.  And when they are included in some of their state polls, I think the public should consider them, but then take them with a grain of salt.

No comments: