A Representative Sample of One
With people dropping their landlines and screening all their calls with caller ID these days, it’s hard for lonely pollsters to find people to talk to, and harder still to make sure the people they talk to are a representative sample of the electorate.
Since I still have a landline (and for some reason still answer it even though most of my friends and family don’t even know the number), and I don’t have caller ID, and I apparently don’t have anything better to do than talk to pollsters, I end up answering telephone surveys from time to time. I’m generally happy to do it—since I voraciously consume poll results during contentious elections, I figure that it would be hypocritical for me not to answer questions when pollsters call. So on Friday, 20 minutes after I happened to have posted comments on some of the Oakland mayoral candidates, I found myself answering questions about the upcoming local elections. And less than 48 hours later, I found myself answering another set of questions from a different polling outfit.
If I try to remember all the questions verbatim, then I will inevitably get things wrong, but since I had an inquiry about who had taken the polls and whether they asked about all 10 candidates for mayor, I’ll try to summarize the surveys as best I can from memory. The first survey was about 5-10 minutes long (it felt like 15, but I wasn’t looking at the clock, and it probably wasn’t quite that long), and was conducted by a woman identifying herself as being with Mountain West Research (I thought she said Madden West, but when I googled for information about the firm, I discovered that Mountain West is a big telephone survey operation, so I assume that I misheard her.)
The poll seemed pretty standard: In a neutral tone of voice, the woman would read something (a candidate’s name, or a summary of a ballot initiative, or a general statement about fiscal policy) and then ask me to say whether I strongly agree, somewhat agree, somewhat disagree, or strongly disagree. The questions addressed everything from my ideological self-identification (I chose “very liberal”) to my assessment of the tax referendums (despite my self-identification as “very liberal,” I said I “somewhat opposed” most of the tax proposals on this fall’s ballot), to my opinions about some of the Mayoral candidates. (Quan: “somewhat unfavorable;” Perata: “somewhat unfavorable;” Kaplan: “somewhat favorable.”) I believe Joe Tuman might have been included also (if so, “somewhat favorable”), but frankly, I can’t remember for sure. I’m fairly sure that I was not asked to evaluate any of the other 6 candidates. I was also asked the approve/disapprove question about the Mayor and some of the city council members who are not running for mayor (I believe I was asked about Kernighan, who represents my district, and De la Fuente, Brooks and Reid, but not about Nadel or Brunner).
Other questions were more general, getting at the same issues addressed by the ballot measures, but slightly more abstractly. I can’t remember the exact wording, but there were questions along the lines of “The city of Oakland is expected to face a large budget deficit for the next fiscal year. Which of the following statements best describes your position: ‘I support balancing the budget by increasing revenue through additional taxation,’ or ‘I support balancing the budget by cutting spending’?” Another question was something like, “The city of Oakland has laid off about 80 police officers, and city officials say that they may have to lay off 150 more if they do not raise additional revenue. Would you support additional property taxes in order to prevent further layoffs?”
I was a bit surprised that the Mountain West survey also asked about my district’s city council race, just because I never expect much polling to be done in city council races. In addition to the question about whether I approve or disapprove of Councilmember Kernighan (“somewhat favorable”), I was asked about whether I approve or disapprove of her opponent, Jennifer Pae (“somewhat favorable,” although if “no opinion” had been a stated option I probably would have gone with that one, since I know almost nothing about her except that she’s a young mover and shaker, and comes across as very cheerful and enthusiastic on Twitter). I was asked which of the two I would vote for if the election were held today, and that was one of the few questions that I wouldn’t answer at all, because I really don’t know—I am not unhappy with Kernighan, but if I learn anything that makes me think Pae would be better, then I am open to persuasion. (If any readers care to weigh in below, then I’m all ears, since I assume that a lot of people are more familiar Pae than I am.)
That was about it for the Mountain West survey. After it was finally over, I asked the interviewer who had commissioned the poll, and she said (as they always do) that she didn’t know. All in all, it seemed like a typical, professionally done survey, and I couldn’t tell from the questions whether it was being done for a candidate, or a media organization of some kind (Can local media even afford to do expensive polling on local elections? Beats me), or a union, or…who knows.
The second survey, done by McGuire Research, was quite a different thing. It was much shorter, and let’s just say that the interviewer had a much less formal mien. That’s not to say that it was a push poll (that is to say, a poll which is designed more to influence the opinion of the electorate than to measure it), but the interviewer had a much more relaxed style, and oddly seemed to veer almost into small talk at times. As for whether this was done intentionally in order to subtly guide my opinions in a certain direction, or whether it was done intentionally in order to elicit more honest responses from the interviewees, or whether it was done unintentionally due to poor training, I won’t speculate, but it was unusual.
The questions were much more limited than the first survey’s, and as far as I remember, they all related to the mayoral race. I was asked whether I was familiar with Don Perata, Jean Quan, Rebecca Kaplan, Joe Tuman, and I believe Terence Candell and Marcie Hodge. It’s possible that he also asked about Don MacLeay or Greg Harland too (my memory is a sieve, and I didn’t take any notes), but I am pretty sure that Fields and Young weren’t mentioned at all. I was then asked whether I had favorable or unfavorable opinions of some of the candidates (again, I can’t remember exactly which ones—certainly the first four listed above, and maybe one or two others). I was also asked whether I knew that Ranked Choice Voting was going to be used in the November election (“Yes”) and which candidates I would put as my first, second and third choices if the election were held today.
The interviewer had used a pretty casual conversational tone throughout the interview, but I had chalked that up to his being a native English speaker, whereas the woman doing the first interview sounded like she was probably not a native English speaker. But at a few points during the survey, the interviewer from McGuire seemed to go a bit overboard in his informality. For instance, when I said (as I did in my post on Friday) that I would probably put Rebecca Kaplan as my first choice if the election were held today, he said offhandedly as he was noting it down, “Okay, Kaplan…lots of people seem to be saying Kaplan,” or something along those lines. Excuse me? Since when do people doing surveys for political polls tell the interviewees how other people have been answering the questions? My gut feeling is that he was not trying to influence me in any way, but it seemed like such an unprofessional thing for a surveyer to say in the middle of taking a poll that it did make me wonder.
That was the most striking example, but there were other moments where his chatty style seemed to border on the unprofessional. For instance, when I told him that I might just leave the third choice blank if the election were held today, he said something like, “okay, so you wouldn’t put any of the other candidates there, Candell or Marcie Hodge or Quan or anyone else…?” Again, I really didn’t get the feeling that he was pushing me to give particular answers, but it seemed a bit odd for him to start throwing out names like that.
I’m generally a skeptic, but not a conspiracy theorist, and my guess is that his informality was just a personality or training issue, and not anything more sinister. It seemed as if he had a list of candidates in front of him on a piece of paper or a computer screen so that he could mark down my choices, and he was just sort of thinking out loud as he glanced through the list. It was a bit strange, however, and people can draw their own conclusions. (The main reason it struck me is because these interviewers are always so determinedly lacking in personality, as if they are doing their best to imitate a robot when they read through their scripts. Being surveyed by a guy who sounded like a normal human being having a normal phone conversation was disarming!)
As I do whenever I am interviewed, I asked him at the end who had commissioned the poll, and he replied, as expected, that they are never told who has commissioned the poll so that they won’t bias the results. So there you have it. I did a half-assed Google search in case I could determine who was doing the polling, but as far as I know such information isn’t always easy to find, at least not until quarterly campaign spending disclosures are filed, which might show what campaigns have hired what polling firms (and even that might not shed a lot of light on things, however, because often a pollster will contract out the grunt work to a surveying firm such as Mountain West, so Mountain West or McGuire might not have been directly hired by the organization who had commissioned the poll.) If anyone happens to know who paid for the polls, I’d love to know, just to satisfy my own curiosity. Better yet would be to hear some actual polling data. There seems to be a widespread assumption that Perata, Quan and Kaplan are the only candidates with a serious chance, and I have no particular reason to doubt that, but like most assumptions, it would be nice to have more data to back it up