Can we trust St. Pete Polls’ findings?

By on August 21, 2012

The influence of St. Pete Polls – a local, non-partisan, survey firm — appears to be growing by leaps and bounds.

A year ago, St. Pete Polls was under the radar, if not a mysterious organization whose methods and motives were unclear. Nowadays, an edition of the Tampa Bay Times rarely gets printed without at least one columnist or reporter referring to the organization’s findings.

This blog has welcomed St. Pete Polls to the public arena, if for no other reason than its findings make for interesting fodder.

I’m also a polling geek who learned the science of politics from the inestimable Dr. Susan MacManus. That makes me a junkie when it comes to crunching numbers, examining cross-tabs and reading polls. St. Pete Polls helps supply a fix.

In July, I questioned some of St. Pete Polls’ findings from a survey of pretty much every race on the primary ballot in Pinellas Its numbers ran counter to not only my own internal polling but to that of other consultants working those races.

My instincts also told me that some of SPP’s findings were waaayyy off, regardless of what its results showed.

Then again, some of St. Pete Polls’ findings were dead-on accurate, including in some races I predicted would go the other way.

So now that the Primary Election has concluded, I think it is worthwhile to go back and look at St. Pete Polls’ findings and compare them to the actual results. One important caveat: their poll is just that, a poll. It’s a snapshot. In fact, the poll’s results themselves may have influenced the actual results, i.e., coverage of Bob Gualtieri’s momentum as evidenced by his improving poll numbers may have created a multiplier effect.

In fact, it is the numbers from the race for Pinellas Sheriff of which St. Pete Polls should be most proud.  Not only did St. Pete Polls’ survey accurately forecast a Gualtieri win, it came within two points of predicting the final outcome. Moreover, SPP was tracking Gualtieri’s directional improvement through all of its polling.

Unfortunately, St. Pete Polls made a glaring, almost unforgivable error by polling a GOP primary in Senate District 8 which simply did not exist. The candidate St. Pete Polls forecast winning that election wasn’t even on the ballot. I realize St. Pete Polls is still working out the kinks in its operation, but I can’t imagine Mason-Dixon or Quinnipiac committing such a grievous error.

In the other State Senate campaigns it polled, St. Pete Polls accurately forecast the winner in several key races — including District 17 with John Legg, District 22 with Jeff Brandes and Jim Frishe and District 24 with Rachel Burgin and Tom Lee.  St. Pete Polls was just one point off from exactly forecasting the outcome of John Legg’s victory, which is very impressive, even if the firm overshot Brandes’ and Lee’s wins.

In my July criticism of St. Pete Polls, I wrote, “I don’t doubt Brandes is ahead of Frishe, but not fourteen points ahead.” Actually, that was exactly how much Brandes would win by, so not only was St. Pete Polls off by forecasting a twenty-eight point win, but so I was for believing the race was within single-digits.

A second criticism I offered of St. Pete Polls was concerning its findings in Congressional District 13, writing, “Only 48.6% of voters are decided on U.S. Rep. Bill Young? That’s highly suspicious.” It turns out my suspicions were well-founded as Young ended up with nearly 70% of the vote.

The rest of my criticisms, save one, of St. Pete Polls findings also turned out to be accurate:

  • I wondered why Ed Hooper was “only” up twenty-three points over Chris Sheppard in House District 67; Hooper won by almost forty points;
  • I did not believe Ken Welch was “only” up seven points over Maria Scruggs; Welch won by thirty-seven;
  • I questioned if Bill Burgess was really leading Andy Steingold by two-to-one; actually that’s by how much Burgess beat Steingold — St. Pete Polls nailed this.
  • Finally, I did not believe that Shelly Ladd-Gilbert was leading the race for the District 1 seat on the Pinellas School Board; a well-meaning candidate Ladd-Gilbert finished a distant third.

So, in summary, St. Pete Polls was right about the big-ticket items, such as in the races for Pinellas Sheriff and State Senate, but it had problems in some of the down-ballot races. I genuinely don’t know why that would be.

Still, it’s fair to conclude that St. Pete Polls can be trusted. Its results in the most important races outweigh its askew numbers in down ballot campaigns.  Moreover, the residents and voters involved in these races should be grateful for St. Pete Polls’ contribution to the public debate.

With a little fine-tuning, St. Pete Polls could be a real force in the years to come.

Comments

comments