A Florida Polling Bonanza

By Jim Ellis

Feb. 9, 2018 — The 2018 Florida Senate race is on the cusp of becoming one of the top political campaigns in the country, but polling has been scarce … until yesterday.

Left: Sen. Bill Nelson (D) | Right: Gov. Rick Scott (R)

Left: Sen. Bill Nelson (D) | Right: Gov. Rick Scott (R)

Three different pollsters released data from their recent Florida electorate surveys, each testing the impending contest between Sen. Bill Nelson (D) and Gov. Rick Scott (R). Though the governor has not announced his candidacy, a loosely affiliated Super PAC has been spending heavily touting his accomplishments through various substantial statewide media buys. Since no other Republican candidate is even contemplating running, few doubt that the governor will make the race.

That being said, Florida Atlantic University, the University of North Florida, and Mason-Dixon Polling & Strategy all released surveys this week. Though each arrived at different ballot test results, and all three have some methodological flaws, it is clear that the overall conclusion tells us that the Florida campaign is already in the toss-up realm.

The FAU poll (Feb. 1-4; 750 registered Florida voters; 375 on-line, 375 via automated telephone system) returns the most surprising result. According to their sampling universe, Gov. Scott has a 10-point, 44-34 percent lead over Sen. Nelson. This seems far-fetched, especially in comparison with the two succeeding polls taken during the same time frame. Additionally, the polling sample contains too many Independents, and is a bit low for both Democrats and Republicans. This makes the ballot test response even more curious and suspect.

The University of Northern Florida (Jan. 29-Feb. 4; 619 registered Florida voters via live telephone interview) sees the race much differently. According to their sample, it is Sen. Nelson who holds a 48-42 percent edge. The problem here is the likely voter contingent is too small for a state the size of Florida. Of the 619 people sampled, only 429 considered themselves likely voters. This would be a good sample size for a congressional district, but not of a state comprised of 27 CDs.

Mason-Dixon Polling & Strategy (Jan. 30-Feb. 1; 625 registered Florida voters) arrives at a conclusion that’s about halfway between the other two, and consistent with the small number of earlier polls that were conducted last year. M-D reports a 45-44 percent ballot test in Sen. Nelson’s favor, again forecasting a virtual tie between the two men. In October, their statewide Florida survey found the pair tied at 44 percent, meaning virtually nothing has changed since that time. A year ago, in February of 2017, the Mason-Dixon ballot test gave Sen. Nelson a stronger 46-41 percent advantage, which is still within a consistent result realm when compared to their next two future polls.

The Mason-Dixon flaw is in the sample’s demographic split. According to the prospectus, the White/Caucasian number is 11 points higher than the Florida US Census count (66 to 55 percent). The Hispanic number is eight points low (17 to 25 percent), while African Americans appear to be three points under-sampled (14 to 17 percent). Additionally, females are two points over-sampled.

Aside from the last point, the demographic skew would seem to cut against Sen. Nelson, suggesting that his lead may be a bit bigger than the one-point edge M-D finds.

Taking account the methodological flaws found in the three polls, and thinking that the FAU survey (Scott plus-10 points) is an outlier, the preponderance of data still points to a highly competitive and virtually even contest against two well known political figures, both with high positives. But, a razor-thin statewide result is certainly nothing new to Florida voters.

Though this race so far has suffered from polling scarcity, we can be assured of seeing regular survey data coming from the Sunshine State throughout the rest of this year.

Leave a Reply

Your email address will not be published. Required fields are marked *