The results are in, and even though Joe Romm suggested:
“Please click here and freep this poll until the magazine has the decency to take it down.”
He apparently had little effect. I do agree with Joe though, the poll was poorly designed. For example look at this result:
Only one problem, the math for percentages doesn’t add up:
That’s because the poll allowed for more than one answer to some questions. When you do that, percentages then don’t reference properly to 100%.
As I and many others said, it was poorly designed and poorly presented. From my vantage point it looked like very little thought went into it.
That said, the results, while interesting, should be taken with a grain of salt.
This reminds me of a humourous sketch by the Chasers,
(well the bit 50 seconds in), and I think their statistics bureaus are very similar. Still, we should recognise this poll for what it is, an overwhelming proof that the consensus of the public is on our side no matter how they skew the presentation or results.
In England, football (soccer) players are routinely encouraged by their managers to give 110%; so we are quite used to having more percentages than a paltry 100.
Ditto what Lucy said.
But what’s the meta-message here? Is SciAm engaged in exploratory self-correction? Is this a prelude to a CAGW flip-flop? If you were publisher, and saw your revenues decreasing and profits disappearing and sensed that your readership was deserting you, what would you do? We know they have eschewed real science for quite some time. This effort was political in every sense, not a return to science. Because that’s a touchy subject: if SciAm admits returning to science, then where have they been?
From my vantage point it looked like very little thought went into it.
It is an unScientific American poll. Why would you expect any thought to go into it. Si-Am has proven itself to be a garbage magazine with no credibility.
Aynsley Kellow — and Anthony’s reply, concerning multiple answers to poll questions:
In my view the reporting of outcomes to such a question is deceptive in any case unless answers are reported in grouped form. Nominally, since there are 4 binary choices there are 2^4 = 16 potential answers, though one would think any combining the fourth option with any of the first three are nonsense. Still I would be interested in the number selecting the last two. Persons who did this would be quite distinct from those selecting the first three, although both groups would contribute to option 3 in the statistics they reported here.
My article on accuracy of surveys