As we move into the election season, polling is the grist of political blogging. Polls will drive media coverage and create horse-race narratives about who is up and who is down and why. Most of it will be meaningless. At its very best, political polling is luck because of the difficulty of constructing a representative sample of the election day electorate before the event. Often polls are contrived to help a party. If you believe anything you see published by PPP between May/June and late-September you deserve whatever bad fate befalls you because they will be generating results to show a strong Democrat candidate right up until they think their results will be included in post-election accuracy analysis.
You are look forward to months of people who could not pass high school algebra — which is why they became political reporters and pundits and blog commenters — lecture you on the relevance of party ID, time of day of poll contacts, day of the week of poll contacts, outliers, sample weighting, margin of error, “skewed” polls, the virtues of robo-calls versus human interviewers, polling patterns from previous years, undecideds breaking for the incumbent, etc., etc. much in the way a pagan shaman would lecture you about the spots on the liver of a black rooster sacrificed at midnight under a full moon. Most of what you hear will be old wives tales based on nothing more than belief and superstition. In short, where at one time GIGO meant Garbage In Garbage Out, technology and the endless stream of political punditry has stood all of that on its head: in polling GIGO means Garbage In Gospel Out.
One of the most enduring of these myths is the “Bradley” or “Wilder”. The premise is that a minority candidate will always show stronger support in polls than at the ballot box. It comes from the 1982 California governor’s race between the black mayor of Los Angeles, Tom Bradley, and the very white George Deukmejian. Bradly showed a consistent 10-point lead throughout the race but lost. Instead of looking for the obvious solution (the polls simply weren’t very good) the theory developed that large numbers of white voters told poll interviewers that they were voting for Bradly, ostensibly so they wouldn’t feel racist, but actually voted for the pasty white guy. If such an effect actually existed — and if you believe it did you have to simultaneously believe that voting against a black candidate is thought to be racist by virtually everyone — it had certainly vanished by the time 2008 rolled around.
The LA Times is now asking if Donald Trump if a similar effect is now forming around Donald Trump.
The firm polled 2,397 potential Republican voters earlier this month, randomly assigning them to one of three different methods — a traditional telephone survey with live interviewers calling landlines and cellphones, an online survey and an interactive dialing technique that calls people by telephone and asks them to respond to recorded questions by hitting buttons on their phone.
By randomly assigning people to the three different approaches and running all at the same time, the researchers hoped to eliminate factors that might cause results to vary from one poll to another.
The experiment confirmed that “voters are about six points more likely to support Trump when they’re taking the poll online then when they’re talking to a live interviewer,” said Dropp.
The most telling part of the experiment, however, was that not all types of people responded the same way. Among blue-collar Republicans, who have formed the core of Trump’s support, the polls were about the same regardless of method. But among college-educated Republicans, a significant difference appeared, with Trump scoring 9 points better in the online poll.
The most likely explanation for that education gap, Dropp and his colleagues believe, is a well-known problem known as social-desirability bias — the tendency of people to not want to confess unpopular views to a pollster.
Blue-collar voters don’t feel embarrassed about supporting Trump, who is very popular in their communities, the pollsters suggested. But many college-educated Republicans may hesitate to admit their attraction to Trump, the experiment indicates.
Color me skeptical. The underlying thesis here is attractive to many that being that Donald Trump is so odious that no respectable person would vote for him. And there are many, some here at RedState, that regularly argue that point. But there is a lot we don’t know about this study. For instance, what was the response rate per polling method? Were all 2,397 respondents equally reachable by phone and online or was the telephone sample formed by including people who could not be contacted online? What did the live interviewers know about the purpose of the interview, i.e. did they think it was a legit opinion poll? Or were they in on the experimental nature of the project? What was the male/female/college grad/non-college grad breakout? And what was the response rate in each of those groups? Were there differences by age?
While unable to give this thesis a comprehensive test, a couple of data points are available that let us hazard a guess. Just this week there were two national polls released. Quinnipiac showed Trump with 28%. CNN/ORC has Trump with 39%. Both used live interviewers via landline and cell phones.
There is an old saw in journalism called Betteridge’s Law of Headlines which, simply stated, says that if a headline asks a question the answer is “no.”
*Note. The reader might ask why I post on polling given my low opinion of political polls. I think that a group of polls can reveal trends and directions but if you ever find me arguing about who is in the lead in a poll showing a 51-49 race, you have my permission to hunt me down and kick my ass.