97% of world Climate Scientists are 75 people

What BS. Even as propaganda spin, Solomon is an idiot. I do notice an interesting movement of goalposts. Now the story is that denialists always thought AGW was present, just not as important as the excuse-of-the-day. Will this be enough to save denialists from going down in history as the morons of the 21st century ?

This is a more objective analysis of the survey:
New Poll Shows Most Earth Scientists Agree on Global Warming « Johnny Rook’s Climaticide Chronicles

 

No, genius.

It says that 30% of those surveyed responded, typical of survey response rates. Of those responding, 82% thought AGW was significant, and when the respondents were sub-grouped by profession the lion's share of the denialists were associated with the petroleum industry or meteorologists.

The sub-groups actively publishing in peer journals overwhelmingly accept AGW, and lastly, 97% of respondents answered 'yes' to question #2 if they were part of the group actively publishing in climate research.

Who wudda thunk ?

For a prime example of denialist spin, contrast and compare:
Solomon:

When in fact

Denialists are so full of sh*t

 

mojo - why are you so tied up with the "75 climatologists" number?

As SageBrush has tried to tell you - ignoring them - it seems pretty clear that the vast, vast majority of scientists believe that global warming is largely influenced by human activity.

 

Maybe it's because real scientists won't stake a "belief" in something. They will posit a theory and then test it. New theories use results of old tests, and that's how a consensus is formed. I don't think climate sceintists are at all worried about some people "believing" AGW and some people "not believing" in it. the topic is so complex that funding is required for decades, and the results of better understanding are so useful in the long and short term that they are not going to be out of a job. The people digging black slime out of the ground to set it on fire and pump it into people lungs, however, are worried about their future.

 

The Canadian survey was a survey primarily of petroleum geologists and related professions. (The professional organization is APEGGA (The Association of Professional Engineers, Geologists, and Geophysicists of Alberta) -- Google them up if you want to verify that.) You know, Alberta -- as in Alberta Tar Sands? Given that, the 50% figure is darned impressive, I'd say. Of course, you'll never see that stated on any of the denialist websites -- it's always "a study of Canadian scientists".

Japan survey, earth science, 90%? That's so obscure I can't find a trace of it via Google. Surveys of the population of Japan show about 70% think its a significant problem, as here. Not hugely different from US data.
http://www.angus-reid.com/polls/2186/global_warming_a_major_concern_in_japan/


Most people do not understand sampling. The source of the original article clearly is one of them. The publishers of that posting (WattsUp) are clearly another.

Let's apply the same logic to any poll, sample, or survey obtained anywhere in the US. Here's the first one that shows on Google:

National Survey on Global Warming - Fight Global Warming - Environmental Defense Fund

Of 1200 respondents to the survey, 300 did not believe that global warming was happening.

So there, we've proved it. Using the same logic as above, there are only 300 people in the US who don't believe in global warming.

But wait. Exactly how stupid and innumerate are we allowed to be here, before somebody says, hey, that's not even correct arithmetic? So, using the exact same logic, couldn't you just as well argue that there are only two climate scientists in the entire US who don't believe in global warming?

If the initial post is logically correct, then the statement above is logically correct.

Sure, poll questions can be misleading, polls maybe subject to select bias and other types of non-random response bias. And polls of the general population are measures of popular sentiment and so have (essentially) nothing to to with facts. But basing a poll on a sample is not sufficient cause to disbelieve the results.

Doing arithmetic off the counts in the sample is just ... well .. wrong. No other word for it. Any point you try to make, that the literal count of YES answers is small, you can make the exact opposite point, because the number of NO answers is also small. Hey, guess what -- in a sample, the number of answers is small. That's why it's a sample.

EDIT: I'll also point out another astonishing aspect of the original citation. Well, astonishing to people who a) are interested in facts and b) are willing to take 60 seconds to do a little homework. The poll in question did, in fact, include all those other types of scientists that the original article mentions. As I recall, all your earth science types were, in fact, included. The guy writing this up didn't even bother to go look up the original poll. The analysts of that poll tabulated the "yes" responses by scientific field. The point of the analysis was that the more you knew about climate science, the more likely you were to understand that global warming was a serious problem. The "yes" responses for all those other scientific fields were still way above 50%. The original research is here:
http://tigger.uic.edu/~pdoran/012009_Doran_final.pdf

The key result was this: The more you know, the more you think this is a real issue. The exact question was: "Do you think human activity is a significant contributing factor in changing mean global temperatures?"

 

I guess when you don't work with environmental scientists on a regular basis it is easy to make assumptions based on terribly researched articles.

Thanks, chogan.

 

You may want to tone down your nastiness. At least with regards to chogan2. He responds to you in a civil manner and supplies well thought out and detailed facts instead of just linking to hyperbole.

 

so we should include microbiologists, sociologists, psychologists, constitutional scholars in survey too? do they have to be licensed scientists? are they licensed by state?

is this a referendum of some sorts?

 

^^ no referendum, just tea-bagger stupidity. I presume the intended message is that the survey cherry-picked a group that gives the wanted answer, but review of the source article makes it crystal clear nothing of the sort took place.

Mojo's understanding of stat analysis and polling is about on par with my neighbor's dog. I give him benefit of the doubt, and suggest his problem is ignorance rather than stupidity or simply trolling.

 

well they obviously did it with 17 Nobel prize winners! why didn't they get at least 1,000??? otherwise the margin of error is too great???

 

op2:

 

Yep, that about sums it up. Anyway, tea-baggers know that Nobel Prize winners are librel scientists, and should be ignored. Ms. Palin and Mr. Inhofe say so.

 

Let me finish my part of this by calculating the confidence intervals for that 97% "yes" response.

I should probably explain what a "finite sample correction" is, when you are estimating the statistical power you are going to get as you draw a sample. Run-of-the-mill sampling statistics are calculated assuming that the population you are drawing from is quite large -- for all intents and purposes, infinite. From that, if you ask simple yes-no question, and have some idea of what fraction of the population is likely to answer yes, then you can readily estimate the variance or the standard error of the mean response that you'll get out of the sample -- the statistical "power", meaning your ability to distinguish differences in responses from just random variation caused by sampling.

But if you actually sample an appreciable fraction of the universe of relevant persons (US climatologists in this case), you have to make a finite sample correction. The math works out very simply. You reduce your infinite-sample variance estimate by the proportion of the total universe you have sampled. So that, e.g., if you sample 100% of the universe, your sampling variance is zero. Ain't no sample, it's a census. If you sample half the universe, the variance of your mean response is half as large as you would calculate, based on standard (infinite universe) statistics. In effect, you pin down the answer tighter than you would otherwise think.

In this case, assuming the reported 30% response rate is the same for all disciplines, and assuming no bias in who responded and who didn't, it's easy enough to show the standard errors around the mean response rate.

Undoubtedly there are websites that'll do that for you. Here's one:
National Statistical Service § Sample Size Calculator

The 95% confidence interval in this case is (94%, 99%). (Actually, the upper bound is listed at 100%, but they may be imposing symmetric confidence intervals on this.) If the 75 respondents were randomly sampled, there's no more than a 5% chance that the true "yes" rate (for the universe of climatologists who publish in this area) would lie outside that interval.

Again, assuming that respondents were a random draw from the field, and that about 30% of the relevant population responded, we can with high probability say that the true estimate, for all US climatologists, would yield a "yes" answer in the high 90 percents.