Are Polls Useful?

The question has to be asked. I was awakened this AM by yet another pollster. I informed her that I don’t speak to pollsters and hung up. I hang up on pollsters and robocalls every day. My friends with Caller ID or Privacy Manager don’t even answer the phone.

So, just how reliable and useful can polls be?

Auf Stumbleupon zeigen
Auf tumblr zeigen

Author: HOLLY IN CINCINNATI, Copy Editor

Copy Editor

  • http://members.shaw.ca/csstrowbridge/MainPage.htm C.S.Strowbridge

    As long as people who don’t talk to pollsters represent a random selection of the people, then they are very useful.

  • Holly in Cincinnati

    I doubt they represent a random selection. I just got an email from a friend who read my post – he doesn’t have a land line, only a cell phone and therefore avoids them. The “no land line” crowd includes many young people.

  • Blue Neponset

    Polls are very useful, but as with any other statistical sample they can we wrong on occassion.

    You seem to believe polls are not useful, why don’t you tell us why you think that, if you do?

  • Holly in Cincinnati

    It’s not a matter of belief. Polls are getting skewed samples.

  • http://members.shaw.ca/csstrowbridge/MainPage.htm C.S.Strowbridge

    The no land line demographic is pretty small, and there are ways around it. (E-mail, for instance.)

  • Mike P.

    I think they are useful, though generally only as indicators of trends. At least until the last two weeks – at that point, polls always seem to converge, and get more accurate when compared against actual votes cast.

    That opinion is based purely my own observations, however.

  • Kim Ritter

    The other day I got a call from a push poll for the Md Senate race. It was obviously paid for by the RNC, because it asked at the beginning and end if I would be voting for Michael Steele. In between, I was treated to a distortion of both candidates stands. Example: Did you know that Ben Cardin favors using unborn babies for medical research?- I guess that means he would vote for embryonic stem cell research, which is favored by 60% of Americans. Md is a liberal state, so the number may be even higher here.

  • Daniel CAZ Greenberg

    If polls mean anything, the GOP is keeping the Senate.

    Ford’s falling behind in Tennessee and I don’t think the Old Dominion is going to elect Webb (call it a lack of confidence in my home state – and I’ll point to our Ballot Init 1 as proof of that,) and those are two of the important ones.

    Now, I’m no Sabato, and I haven’t been following this dead-on. But, I expect 51R-49 or 50R-50 with a light, heart-healthy Cheney crust topping, based on what I’ve seen.

    If Allen actually wins, I’ll make a point to come back here and link a picture of humble pie being eaten by yours truly (although since I have no stomach for deer organs, I’ll likely make mine from mixed berry.)

  • Idosyncrat

    I’ve done a good deal of work with a certain prominent polling organization. Like Holly, I too had my doubts and skpeticism — but I have to tell you, the methodology, at least in their case, is excellent. It’s counter-intuitive that small samples can be representative, but millions upon millions of data points over decades show that the results are statistically significant. There are ways to balance out the cell-phone effect, hangups, and just about every factor you can imagine. The trickier issue is how questions are worded and loaded as merely changing the way a question is asked yields different results (which is correct?), but even that is balanced with switching things around. For this skeptic, it’s really all quite impressive…

    NOW, whether or not the polls wind up being accurate because people say one thing and do another when their actually in the voting booth is another story.

  • Holly in Cincinnati

    The no land-line demographic includes my 40-yr-old brother, the 50-ish friend who emailed me earlier today, lots of young people and possibly soon my 72-yr-old mother.

    The no land line demographic is pretty small

  • Idosyncrat

    The cell-phone only demographic is estimated to be 6-8% of the general population, with a larger % (perhaps up to 12%) of 18-24 year olds. believe it or not, it’s estimated that up to 3% of the population has no phone whatsoever.

    For great analysis of polling issues, warts and all, I highly recommend Mark Blumenthal’s (formerly known as ‘Mystery Pollster’) http://www.pollster.com site. On a more academic level, Don Dillman, a fantastic survey methodology guy out of Washington State University, has a website with some of his papers and analyses.

  • Holly in Cincinnati

    Thanks Idosyncrat!

  • Mike P.

    Interesting – unless I’m missing it, Pollster.com doesn’t even list the latest Rassmussen for VA Sen, yet still has Webb over Allen in the 5-poll average (and tied in their 10 poll avg). Is the Ras still considered “premium” knowledge?