Questions about political polls
October 19, 2020 6:57 AM   Subscribe

Do live interviewers in phone polls know who's paying for their poll? Would they tell me if I asked? Do pollsters know anything about my political affiliations or how often I've voted in the past?

If I get a call from a pollster, they generally say who they work for near the beginning of the call, but the names of polling organizations are usually pretty bland and opaque, e.g. [Surname] Research Strategies, which doesn't tell me anything about who I'm giving the information to, and has on occasion led to me getting halfway into a push poll before I realize that's what's going on. But: a lot of the calls I get are likely being commissioned by specific parties or candidates. Is there any harm in asking the interviewer who's paying for the poll before I start answering questions? How likely are they to know? If they do know, how likely are they to tell me? If they do tell me, is there any strategic advantage to accepting polls from some groups and not others, or to accepting all polls but lying to some of them?

I know that in some cases, the pollster got my phone number by dialing randomly, but are there any polling organizations that connect my number to my voting history or party affiliation (both of which are public in my state), or past responses to polls from the same pollster? For computerized polls without a live interviewer, if I'm actually a 20-year-old Democratic woman, will the pollster know I'm lying if I say I'm a 75-year-old Republican man, and discount my responses accordingly? (If so, why do they ask for age, gender, party affiliation, etc. anyway?)

Bonus questions:

Why do so many live interviewers have very strong accents? It seems like the companies would want the questions to be as clear as possible, and yet I struggle to understand about two out of three live callers (sometimes they also struggle to understand me, especially if I don't respond with exactly the words they're expecting).*

If I realize I'm in a push poll, should I continue to the end or hang up right then? Is there any way to respond that makes people less likely to do future push polls (even if only incrementally less likely)?

-

* I'm assuming the answer is some version of, people with accents take the jobs they can get, same as anybody else. It feels worth asking anyway because it sometimes seems like pollsters are actively selecting for people who are difficult to understand, and that doesn't make any sense.
posted by Spathe Cadet to Grab Bag (5 answers total) 1 user marked this as a favorite
 
Best answer: As with most things, it's a cost/benefit ratio and comes down to money:

1. These jobs are outsourced to overseas call centers, as many companies do.
2. They are going with very inexpensive centers that employ people with thicker accents.
3. There's no quality control because this isnt a customer service job where poor understandability could lead to loss of business. So the polling company has no motivation to improve the situation for the people who are called, because this is about collecting as many responses as possible for the lowest amount of money. This means minimal training and oversight of the workers to keep costs down. If they can get enough responses because of the volume of calls, the cheaper labor offsets the risk of having someone not complete the poll because they cant understand the caller.
posted by ananci at 7:31 AM on October 19, 2020


Best answer: If I realize I'm in a push poll, should I continue to the end or hang up right then? Up to you.

Is there any way to respond that makes people less likely to do future push polls (even If they do tell me, is there any strategic advantage to accepting polls from some groups and not others, or to accepting all polls but lying to some of them? if only incrementally less likely)? No

are there any polling organizations that connect my number to my voting history or party affiliation (both of which are public in my state), or past responses to polls from the same pollster? No, or at least highly unlikely. Saving and tracking data is expensive.

will the pollster know I'm lying if I say I'm a 75-year-old Republican man, and discount my responses accordingly? No, its automated. They are not going to pay people extra to guess if a response is untrue at an individual response level. With the volume of data they get, outliers of bad data are not generally enough to skew results.

why do they ask for age, gender, party affiliation, etc. anyway?) demographics are key to understanding and analyzing the collected data. This is also how they can track outliers.

Is there any harm in asking the interviewer who's paying for the poll before I start answering questions? No, why would there be?

How likely are they to know? They will probably not know who is commissioning their employer. They may even be a sub-sub contractor. Even if they did, the name of the commissioning org may not mean much to you as many of them are big data collecting firms who arent going to tell a call center who's giving them money, and the call center certainly isnt going to tell their workers.

If they do tell me, is there any strategic advantage to accepting polls from some groups and not others, or to accepting all polls but lying to some of them? No. Again, the small fraction of untrue or biased individual responses wont be enough to truly skew data and is obvious on a scatter plot anyway.
posted by ananci at 7:48 AM on October 19, 2020


Best answer: Based on some long-ago experience in the industry: They are very unlikely to know who the sponsor is, and if they know they are unlikely to be allowed to tell you.

"Push poll" generally refers to a supposed poll where the goal is to convince you, not to gather information. The only possible benefit to continuing a push poll would be to waste their time; 10 minutes taking to you is 10 minutes they can't be talking to someone else.

However, some polls are designed to not only gather information about your current opinions but also to test messaging: "If you knew X, would you be more willing to vote for candidate Z? If you knew Y, would you be more likely to vote for her?" This isn't really a push poll. Respond if you can figure out who's sponsoring it and want to help that candidate or group.
posted by Mr.Know-it-some at 7:48 AM on October 19, 2020


Best answer: I worked briefly for a reasonably big polling company in the US twenty years ago. We did some political polls, mostly apparently hired by regional organizations, but the real income came from surveys of for-profit hospital patients and broadcasters. We never knew who commissioned a poll and were instructed to give out the name of our particular subcontracting company and a public phone number if pressed. We never challenged a person's responses or made any comments if what they said seemed unlikely. (Even in the case of people obviously who were unable to truly understand the questions or who said very strange things.) We were also instructed to lie, constantly. For example, we often said, "in order to achieve a random sample, we'd like to interview the youngest adult man in the house today," when there was obviously no actual attempt at randomness and a sports radio network had paid us to interview only young men. I wouldn't believe anything an interviewer tells you about their sponsors or methods.
posted by eotvos at 8:32 AM on October 19, 2020 [2 favorites]


Best answer: Some interesting stuff on recent moves away from Random Digit Dialling since 2016: https://fivethirtyeight.com/features/what-pollsters-have-changed-since-2016-and-what-still-worries-them-about-2020/
posted by Chairboy at 6:09 PM on October 19, 2020


« Older tell-tale heart, but it's coming from the roof?!   |   Request: Memoir About Moving to a New City Newer »

You are not logged in, either login or create an account to post comments