Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
Politics

Journal pudge's Journal: What Are You? 16

Journal by pudge

The reason why Time and Newsweek have Bush 11 points up and others have him only a few points up is because of polling methodology differences.

You see, some people would have you believe polling is an essentially exact science. You get a good sample, ask good questions, and boom! You have your answer.

But real life is not that simple. Political pollsters weight their results based on different demographic factors. If 6% of respondents are black, but 12% of voters are black, then they increase the black respondents vote in the results accordingly. That mostly makes sense to most of us, I imagine.

They do the same thing with party affiliation. The thing is, we don't know party affiliation numbers. We have to guess. So what Zogby did is guess 39% Democrat, 35% Republican, 26% Independent (last three elections were, in order, 34-34-33, 39-34-27, 39-35-26). So Zogby is doing a reasonable scientific thing, and he comes out with results far closer than Time and Newsweek did. Newsweek had 31-38-31.

But just because it is a reasonable scientific thing, does that mean it is correct? No, because he's only guessing. Zogby says he came closest to outcome in 1996 and 2000, but that doesn't mean he is closest now. Maybe Time and Newsweek really did discover a party shift in the electorate. But the point is that party is not a constant like the other demographic figures are, and we can't assume it does not change, and in the end, they are all only guessing.

Only time will tell who was right. Less than 60 days' time, actually.

This discussion was created by pudge (3605) for no Foes, but now has been archived. No new comments can be posted.

What Are You? More

What Are You?

Comments Filter:
  • Well, when I filled out my information for my absentee ballot(since I'm military) I declared my party affiliation to be Republican, so now if every other voter in the country would be so kind as to list their affiliation below, we can finally clear thjis whole thing up.
    • this kind of party registration is not very accurate. for example, this primary season, i said my party affiliation was democrat so that i could vote for a presidential candidate. in the last presidential election, i voted for g.w. bush. but this year i plan on voting for michael peroutka of the constitution party. actually, i almost voted libertarian in 2000, but chickened out.
    • This is strange to me. It seems that if you have a significantly large enough random sample of the US population you won't be doing twiddling with the percentages.

      And if your sample is 6% black and the US is 12% black then I argue that your sample, well, isn't very random and kinda sucks.

      P.S. I am NRA. :)
      • This is strange to me. It seems that if you have a significantly large enough random sample of the US population you won't be doing twiddling with the percentages.

        That's exactly the point: they don't get good samples, so they have to use more-or-less established means to fudge them into usability, and sometimes those means are not as clear cut as at other times.
        • Ah, so like many people, they're lazy at their job.

          Ah, well. Dewey defeats Truman.
          • It is somewhat more complicated than that.

            The problem is polling is expensive and the more people you poll the more expensive your poll is to run. Since most of the polls you see are paid for by somebody, either subscribers to regular polls or polls comissioned for a specific organization, there is a need to ballance effectiveness with cost.

            There is also the problem that no matter how many people you try to get in your sample to some extent it is going to be off from actual demographics.

            In any case I'm w
            • In any case I'm willing to bet on Zogby being among the most accurate polls out there if for no other reason than his polls have proven to be among the most accurate since 1996 (a better record than Gallup has in the same time period).

              If I had to bet on a poll it might be Zogby, but I'd want to find out more his reasons for assuming the weights of the parties should remain the same in 2004. But luckily, I don't have to bet, and I don't really trust any of them. :-)
              • As I'm sure you are aware the polls right now won't necessarily reflect the state of the race the week before the election.

                Zogby is probably the best snapshot of the current state of the race and overall trends but I'm not going to make any bets that the numbers today will at all reflect the final outcome.
                • As I'm sure you are aware the polls right now won't necessarily reflect the state of the race the week before the election.

                  Oh, sure. No polling on governor up to this point -- except between Sims and Gregoire -- is meaningful. The race begins this week.

                  The one really positive thing for the Republicans, especially Rossi and Bush, is that each will pick up votes from the other. A lot of people in WA dislike Bush and would not be too keen to vote for him, but also would not vote for Kerry, but would come
                  • Agree we can't know what the Governor's race will look like until after the winner in the Dem primary is declared (I'm guessing it will be Gregiore)

                    To some extent you are right that support for one canidate can translate into votes for another but that is most typically seen in a Presidential canidate having "coattails".

                    One thing about Washington voters is we are habitual ticket splitters. Witness Slade Gorton or Jennifer Dunn winning their contests at the same time the voters went for Democratic Presiden
                    • Agree we can't know what the Governor's race will look like until after the winner in the Dem primary is declared (I'm guessing it will be Gregiore)

                      It is.

                      One thing about Washington voters is we are habitual ticket splitters. Witness Slade Gorton or Jennifer Dunn winning their contests at the same time the voters went for Democratic Presidential and Governor's canidates.

                      Sure. I am used to that, as it happened quite a bit in CA, and also happened in MA in governor's races (all statewide races are won b
                    • Sigh ... oh well. As I said I'm going to give Rossi a good long look.

                      My predictions for WA for November:
                      President: Kerry
                      Senate: Murray
                      Governor: Gregiore (that soccer mom thing again)
                      AG: Senn

                      WA-08 primaries:
                      Both parties had extremely well-qualified canidates running for this office unfortunately the boobs with the soundbites won both party's primaries. As you can probably guess I'm not a big fan of either Ross or Reichart.

                      Sure. I am used to that, as it happened quite a bit in CA, and also happened in MA
                    • Rossi will win, mark my words.

                      Supprised, I thought Gregiore was trying to duck participating in any debates if she could help it.

                      She can't. She is not the incumbent, and has little choice. Maybe if Rossi were a nobody with no money, but he isn't. Murray is the one who surprised me a bit; not that Nethercutt is a nobody with no money, but she is the incumbent.
  • Democratic (as I think you know).

    But often people don't know or forget their party affilation, so I would guess that this is not that reliable a way to adjust polling results. For example, if more Kerry voters self-identify as Democratic than are, would the use of the ratio of Democrats to others still be correct?

  • Who cares about the national polls? The electoral votes is what matters.
    • There's a few reasons to care about national polls, but the basic reason is trends. State polls are not all done the same way with the same methodology at the same time, so national polls can give you a way to see the bigger picture. But it is all about trends and seeing the bigger picture.

      Also, the Commission on Presidential Debates uses national polls to determine eligibility ...

Slashdot Top Deals

Contemptuous lights flashed flashed across the computer's console. -- Hitchhiker's Guide to the Galaxy

Close