Comments on: Dear News Media…

By: How Many Ways To Interpret “No”? Just One | The Intersection | Discover Magazine

Mon, 15 Feb 2010 17:28:43 +0000

[...] said, I’m skeptical the results of this so-called survey reflect [...]

By: Brian Too

Brian Too — Thu, 28 Jan 2010 00:23:52 +0000

Is the lady news anchor wearing a ‘stache? How European!

By: twisminion

twisminion — Tue, 26 Jan 2010 19:17:38 +0000

“It seems to me that many of the posters here are more interested in feeling like a member of a “smarter” minority than creating a political majority that might result in concrete accomplishment.” -moptop

If being part of the majority didn’t require a science lobotomy i’d be all for it…
What we need is a science-y philosopher king state in which the science educated elite rule over the ignorant masses…

national physics care plan to include pre-emptive as well as emergency and in class round the clock coverage of all critical thinking needs…

By: Georg

Georg — Mon, 25 Jan 2010 11:50:50 +0000

The necessity of this plea is obvious,
as is its vanity.
Georg

By: moptop

moptop — Mon, 25 Jan 2010 11:36:10 +0000

P.S. It seems to me that many of the posters here are more interested in feeling like a member of a “smarter” minority than creating a political majority that might result in concrete accomplishment.

By: moptop

moptop — Mon, 25 Jan 2010 11:33:18 +0000

The importance of polls on global warming is not to show which side is “right”, and which side is “wrong”; in that regard they are dispositive of nothing. Polls tell you whether your rhetoric is working or, as in the case of global warming alarmism, no so much.

What they tell you is that mockery of your opponents, instead of answering their objections, eventually devolves into mockery of the majority. You can say that you are tired of answering objections, then people conclude that you must think the issues in not nearly as important as you claim.

If you are losing in the polls, the message is not do more of the same only louder.

By: Ian

Ian — Mon, 25 Jan 2010 08:21:58 +0000

I would also day that in respect of point no 2, the example given is historically inaccurate. Very few people in Galileo’s time believed the earth was flat.

http://www.bede.org.uk/flatearth.htm

By: Ian

Ian — Mon, 25 Jan 2010 08:16:12 +0000

Panel no 2 is historically inaccurate.

By: Adeist

Adeist — Sun, 24 Jan 2010 10:48:47 +0000

Not bad. But there are lots more.

5. Find out how the sample was selected. If the subset chosen is linked to the quantity under study, the results will be biased.
6. The quoted errors are usually calculated assuming an independent sample selection, the true errors may be bigger.
7. How many people refused to take part? Could the balance of results be different among them?
8. Ask what post-survey adjustments were made. Professionals sometimes knowingly take selection-biased samples and then try to correct for them mathematically.
9. Did everyone tell the truth? Some things people are embarrassed to admit. Did everyone understand the question?
10. People will often give the answer they think the surveyor wants or expects. Are the questions worded to hide that? Are any of the questions loaded? Many issues are in shades of grey – do the tick-box categories cover all cases?

I could go on.

Incidentally, your point number 1 might not be true if the errors are positively correlated.

And on point number 2, scientific facts are not established by *any* sort of opinion poll, (unless you’re studying opinions).

By: nathan

nathan — Sun, 24 Jan 2010 08:52:52 +0000

Thank you thank you thank you. I support this message wholeheartedly. This is a major pet peeve of mine.