I have to wonder how pollsters, who largely rely on phones for polling, contend with the increasing number of call screaners out there. I never answer my phone unless I know who is calling, and I have quite a few friends now who don’t even have a land line anymore. Is this an acknowledge problem, and are polls getting less accurate as a result? I can’t imagine how pollsters could get a moderately well off person on the phone these days, and even low income people often have caller ID.
I have no idea, but I know that while I have caller ID on both my cell and home phone (SBC included it free, so hey, why not?) I still answer if it’s a phone number I don’t know because it could be someone I know calling from somewhere important - like the hospital. So the pollsters do get me, and I usually either refuse to talk to them, hang up if they’re especially rude, or if I have the time, I sit down and do the poll.
I know a lot of people who have the same problem about answering the phone…just in case.
But pollsters usually have weird area codes that are dead give aways. Anyway, I think there is the potential for significant parts of the demographic to simply be unavailable to phone surveys. Not that they won’t take them if they answer the phone, but they won’t answer the phone in the first place, or won’t even have a phone (land line).
The fact that some people refuse to participate could skew the poll, but I think it’s generally considered to not be a factor. That is, there’s no demographic that’s more likely than others to refuse. The pollsters just have to plan on calling enough random people to get the number of responses desired. If half of all people usually refuse and they need 3,000 responses, they plan on calling 6,000 people.
On the other hand, people having just a cell phone and no land line is skewing polls. The reason is that young people are much more likely than others to have this arrangement. So they are less likely to be included in polls. I’m not sure what, if anything, they can do about this.
Having worked in Market Research before (NOT a pollster) we acutally get MORE than the desired results, as a properly managed project will have representatives from specific demographics, useage patterns, sex etc - so we will keep going till all bases are covered.
So in response to the OP - caller ID may make it harder, and may introduce bias, but for a robust project it won’t make any difference - beyond just extending the timeline a little
Also just a simple request, if the researcher (excluding pollsters) is polite, and you feel the survey is genuine and not just a telemarketing ploy - please do give them your time. Market Research is genuine, very important and can be very hard work for those conducting the survery (I could really tell some fun stories)
Here’s one way we used: You take a (small) sample of the non-responders, and you call them back over and over again. Eventually you get a handful of them to respond, and you give up on the rest because you don’t have enough resources to keep after them and the study is due to be released soon. Then you run your preferred statistical significance test to show that there’s no evidence of a difference between the sample of non-responders and the responders overall.
Of course, since you don’t want to show a difference, you have an incentive to take a small sample of non-responders and thus decrease the power of your test. And you try as best you can to forget that the majority of the people you called were non-responers.
Yes, there are methodological issues here that I’m not even getting into. I’ve gotten out of the phone survey business, and I’m very glad I don’t have to keep pretending that the results have a “margin of error of plus or minus 3%” :rolleyes:
Young people are more likely than others not to vote, too. I’m not sure of this, but don’t pollsters still use random digit dialing? Wouldn’t that theoretically get through to cell phone users, assuming they weren’t checking it against a landline database (which would only have listed numbers, anyhow)?
I suspect many telemarketers would make exactly the same statement. I’m not convinced the holy grail of Market Research (particularly by means of incessant telephone calls) is important enough for me to treat you any differently from a low-life telemarketer. And, while we’re at it, why did you exclude pollsters? That’s just a different form of Market Research.
I’m totally at a loss for words that would be acceptable outside of The Pit (or even inside it). You call them over and over again?! You definitely are in the same category as low-life telemarketers!
At this point, if you’ve irritated me by pulling this stunt on my phone, I inform you that I am voting for Cthulhu because he promised that when he rises from R’yleh he will eat the souls of telephone pests first.
Pollsters take that into account. They also take into account that some people in this country vote but do not have phones. Yes it is a fact. Mostly in poorer neighborhoods, especially with cell phones, the older folks who only use phones in emergencies all just use one neighbors phones. This isn’t as unheard of as people think. Most older Americans over age 40 grew up with one phone and don’t use it for every little thing.
There’s no ethnic group or political bloc that’s more likely to refuse to talk to pollsters, but the results could still be skewed.
If enough ordinary folks refuse to talk to pollsters, the pollsters will eventually be stuck with a group of respondents who are MUCH more passionate about politics and MUCH more eager to talk about politics than the average person. And that can give an erroneous idea of how worked up the populace is about a given issue.
I used to administer surveys for market research as well. I always wondered whether any of the higher-ups had considered whether there are any important implications to be drawn from the fact that their sample, however demographically diverse it might be, is still by nature skewed in at least one way: All the respondants are such that they are willing to answer surveys over the telephone.
Anyway, while I do encourage people to answer marketing surveys if they can and if they have time, and not to be rude to the poor people calling them, still, I wouldn’t call market research exactly important. Not for any respectable purpose I can think of, anyway. It’s just a way for the corporations to learn how better to make money selling their product.
a) I don’t get a lot of phone calls from unfamiliar area codes, so any one I can’t immediately identify is suspect
b) This week I’ve gotten calls daily from 703. Northern Virginia.
Not just poor people and seniors. In my experience, the demographic that doesn’t have a land-line phone are mostly younger, middle-class people. They’re used to using their cell phone as their primary phone, and they don’t have a land-line phone because they see it as an unnecessary expense.
Yeah, I empethize. We had only a few callers (the best ones) put on this duty to call back non-responders, and always felt pity for them that they’d have to do it. Many people we called weren’t so friendly.
In my case, we weren’t selling anything - these were polls of public opinion to help decide where to spend gov’t money for road surfaces, health issues, etc.
I recognize that this could easily go to the pit, so I’m trying to stay factual and answer the questions. This is how we delt with the issue of nonresponse. To date, I haven’t seen a better proposal.
Nope. It’s what we did. For public opinion measurements, the phone survey still beat all other alternatives. Using the mail generated an even higher nonresponse rate, and you still had the same issues.
In the end, as I said, I’m glad I quit and don’t have to justify these issues on a daily basis, or go home and ponder whether the results have much resemblance to reality.