$27 million anti-evolution museum to open soon

[elSicomoro]

Originally Posted by TheMercenary

Ok pick a poll that has original data which we can inspect and we can pick it apart. The point here is that you rarely if ever can see how or where the data was gathered. I spend part of my job reading original source research. You have to know how to find the weaknesses before you accept the data. And the validity would increase as multiple researchers are able to replicate the data you gathered in exactly the same manner. You rarely have access to how polling data is gathered, therefore the research cannot be properly evaluated.

I have experience in both social and physical science research and am familiar with poll development. I understand what you are saying. While I don't think that polls get it right all the time, the major ones (Gallup, Zogby, etc.) seem to have taken great pains to be more transparent in their polling. For example...from this Gallup poll:

Results are based on telephone interviews with 1,007 national adults, aged 18 and older, conducted March 23-25, 2007. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±3 percentage points. In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.

If you see more into it than I do, and would prefer raw data, that's all well and good. From what I'm seeing, this is a pretty solid poll...99% confidence would be nice, but 95% is usually a fair standard for statistical significance. And I don't have a reason to suspect that Gallup is trying to manipulate numbers for some sort of advantage or benefit.

[TheMercenary]

Originally Posted by sycamore

I have experience in both social and physical science research and am familiar with poll development. I understand what you are saying. While I don't think that polls get it right all the time, the major ones (Gallup, Zogby, etc.) seem to have taken great pains to be more transparent in their polling. For example...from this Gallup poll:

Results are based on telephone interviews with 1,007 national adults, aged 18 and older, conducted March 23-25, 2007. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±3 percentage points. In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.

If you see more into it than I do, and would prefer raw data, that's all well and good. From what I'm seeing, this is a pretty solid poll...99% confidence would be nice, but 95% is usually a fair standard for statistical significance. And I don't have a reason to suspect that Gallup is trying to manipulate numbers for some sort of advantage or benefit.

Good stuff. The only thing that I wonder, and I don't know this, are not poll organizations hired by groups to study data? Is there no potential for bias? The greatest weakness is not only in the regularly small sample size, for example I would hardly say that 1000 people with telephones represent and extrapolation to "most people think...", but the weakness in buried in the questions and how they are constructed.

[elSicomoro]

Originally Posted by TheMercenary

Good stuff. The only thing that I wonder, and I don't know this, are not poll organizations hired by groups to study data? Is there no potential for bias? The greatest weakness is not only in the regularly small sample size, for example I would hardly say that 1000 people with telephones represent and extrapolation to "most people think...", but the weakness in buried in the questions and how they are constructed.

I don't know either...I would suspect so. The 5% should cover most bias, though they have the disclaimers. As far as extrapolation, I guess it would depend on how they figure out who they're going to call...I don't know how they're doing that.

[TheMercenary]

Originally Posted by sycamore

I don't know either...I would suspect so. The 5% should cover most bias, though they have the disclaimers. As far as extrapolation, I guess it would depend on how they figure out who they're going to call...I don't know how they're doing that.

Agreed. And then the press picks up a result and runs with it. And then it is exploited by one group or another.