Beware Poll-Spinning PolsPolitically Aware Thursday, October 13th, 2011
Commentary: Politically Aware
Polling may be a science, but reporting polls is the art of spin, and reading them can be as confusing as a complex physics equation.
I hadn’t planned to discuss polling again so soon, but the contrasts in two recently released surveys demand an explanation, and offer a lesson. To test your spin-sulation, scribble your answer to the following before reading further:
1. Did Congressman Filner’s support: (a) increase by 11 percent or (b) nearly double?
2. Did Councilmember DeMaio’s support: (a) increase by 3 percent or (b) nearly double?
3. Whose support increased more: (a) DeMaio or (b) Assemblymember Fletcher?
4. If Filner, DeMaio and Fletcher all saw increased support, did District Attorney Dumanis’ support: (a) increase or (b) decrease?
When we last discussed polls, the candidates stood thus: Councilman DeMaio at 22 percent, District Attorney Bonnie Dumanis at 15 percent, Congressman Bob Filner at 14 percent, and Assemblymember Nathan Fletcher at 7 percent (SurveyUSA and KGTV, June 8). Almost four months later, everything has changed. Or almost nothing has.
1. Both. SurveyUSA and KGTV released another poll Sept. 27, showing Filner at 25 percent. 25 percent-14 percent = 11 percent, what we’ll call the “absolute” increase. But compared to his original 14 percent, that 11 percent absolute increase is a “relative” increase of 11 percent/14 percent = 79 percent. Sure, it’s closer to 80 percent, but any spinmaster worth her salary would say his support “almost doubled”.
The remainder of said survey showed: DeMaio at 25 percent, Dumanis at 16 percent, Fletcher at 10 percent, “Other” at 9 percent, and “Undecided” at 16 percent. Feel free to change your answers on 2-4.
2. Both. In the SurveyUSA poll, the absolute increase in DeMaio’s support was 3 percent. The relative increase was only 13 percent. So how can you say it nearly doubled? Enter our second poll, released to the sdrostra.com blog by Brown (and Bailey) Sept. 13. Their study showed DeMaio at 39 percent, Filner at 28 percent, Dumanis at 19 percent, Fletcher at 10 percent and “Other” at 5 percent.
Spin-glish would allow a DeMaio to claim the doubling (22 percent to 39 percent) as much as Filner. The GOP elephant in the room, though, is that comparing polls with different methods is a touchy business. (To their credit, Brown and Bailey don’t suggest this.) The Brown poll notably does not include “Undecided”, arguing that voters don’t have that option on Election Day. That’s not technically true, as they could skip the questions, but it’s probably fair to assume that the mayoral race will be the ballot’s marquee match-up, or a close second to the presidential primary.
Why does DeMaio do 14 percent better in the Brown poll, while no other candidate gains more than 3 percent? One might expect that eliminating “Undecided” would help all candidates equally, or in proportion to their share of the vote. Both polls involved calls to likely voters from a similar mix of political affiliation, making it unlikely that Brown simply sampled more DeMaio voters (Bailey admits to being a DeMaio supporter, but left the voter selection to his partners.)
Instead, DeMaio’s bounce is due to his high profile ballot initiative. Early in the electoral cycle, undecided voters are often unengaged voters. When pressed, they preferentially choose the name most recognizable or associated with the race. Given the media exposure of his Comprehensive Pension Reform (CPR) initiative, most voters’ first answer to “Who is running for Mayor of San Diego?” would probably be “Carl DeMaio”.
By now, you probably realize the answer to “3” depends on which poll you read. Using the best comparison, the two SurveryUSA polls, Fletcher’s rise from 7 percent to 10 percent equals DeMaio’s 3 percent absolute increase, and his 43 percent relative increase blows DeMaio’s out of the water. As to “4,” like the other candidates, Dumanis’ support increased; a conundrum possible because the June poll included candidates who chose not to run, like Christine Kehoe and Steve Francis.
If you need proof that these numbers will change, consider this: Fletcher and Dumanis are each the choice of 15 percent of self-identified Republicans. So is Filner, the only Democrat in the race (SurveyUSA). If that holds up, Filner will survive the primary, and win the general in a walk. It won’t.
Dumanis has barely started her public campaign, but has quietly built an impressive endorsement and donor network. As DeMaio’s media coverage fades after the submission of CPR signatures, Dumanis’ could increase should there be a high profile law enforcement issue. Anyone who doubts her political skills should remember the Gardner case, where she got closure for the DuBios family by taking the death penalty off the table, and looked tough on crime by doing so. Filner and Fletcher will also move in and out of the spotlight as each of their legislative votes are over analyzed for clues to their mayoral agendas. If his initiative makes the ballot, DeMaio could rise with it, or seem less necessary should it pass.
In fact, the only sure thing is that a lot will happen before the only poll that matters … June 5, 2012.
Short URL: http://lgbtweekly.com/?p=15991