HD 96 poll results: Outliers and the margin of error

in Uncategorized by

Let me start at the beginning.

We conducted a quick brushfire poll to see where things were in the contested (or what we assumed was contested) House District 96 Democratic primary and the results were, well, a little crazy.

Of those who have already voted, Kristin Jacobs was beating, destroying, walloping (choose your term) Steve Perman by an unbelievable margin of 86 to 8 percent and 53 to 13 percent for the entire sample. Some simple calculations and it appears Jacobs could possibly end up in the 70 percent range. That’s impressive.

Or is it?

In polling we often quickly gloss over – almost like the audio disclaimer in a political radio spot – the standard, “this poll has an 8% margin of error at the 95% confidence level.”  What that last part — the “95% confidential level” part — essentially means is that the remaining 5 percent of polls will likely be outliers. By definition, they will (purely due to random chance) produce results that are not entirely accurate (a.k.a. “valid”). This is not due to anything other than the laws of bell curve (parametric) statistics and simple probability. We see it all the time and statisticians have several tools to correct even for these natural variants.

So what do we have here?

Is Jacobs really that far ahead? Is she the outlier? One could make that case. She has won consistently and by large margins of 75 percent in the district three times in the past; she has significantly out-raised Perma; and her campaign tells me her ground game is awesome.

Or is this poll one of 20 that fall outside the margin of error? Is the poll the outlier? One could also make that case. Perman is an experienced campaigner and, despite being heavily out-spent, was no slouch himself. He has raised nearly $70,000 and loaned himself another $10,000 – certainly he has enough to engage in serious voter outreach. He could easily drop five or six pieces of mail (he did a very interesting, if not odd, scratch-and-sniff attack piece that should get voters’ attention) and he should be running a real live campaign … certainly not the kind of campaign where he’s looking like a guy who barely raised enough to pay his own qualifying fee.

Here’s the beauty of this little experiment: We will know soon enough.

I can’t wait for Tuesday.

Peter Schorsch is the President of Extensive Enterprises and is the publisher of some of Florida’s most influential new media websites, including SaintPetersBlog.com, FloridaPolitics.com, ContextFlorida.com, and Sunburn, the morning read of what’s hot in Florida politics. SaintPetersBlog has for three years running been ranked by the Washington Post as the best state-based blog in Florida. In addition to his publishing efforts, Peter is a political consultant to several of the state’s largest governmental affairs and public relations firms. Peter lives in St. Petersburg with his wife, Michelle, and their daughter, Ella.