This is embarrassingly bad for Asimov.
1) Comparing scientists and "the general public" is quite misleading. A better comparison would be of creative scientists to intelligent people from well-to-do backgrounds who lack scientific creativity. I suspect that comparison would be much less dramatic.
2) He uses "sales of one of my books" as a proxy for "adults interested in science fiction". The hubris is staggering. Yes, Asimov was a popular science fiction writer. But there's still no support for the assertion that on average every serious fan of science fiction will own this particular unidentified book.
3) These two points exacerbate one another. Note the shifting of goal posts. "Half of scientists have some interest in science fiction" versus "only a small fraction of adults have bought my book."
1) Comparing scientists and "the general public" is quite misleading. A better comparison would be of creative scientists to intelligent people from well-to-do backgrounds who lack scientific creativity. I suspect that comparison would be much less dramatic.
Irrelevant for the proposed use.
2) He uses "sales of one of my books" as a proxy for "adults interested in science fiction". The hubris is staggering. Yes, Asimov was a popular science fiction writer. But there's still no support for the assertion that on average every serious fan of science fiction will own this particular unidentified book.
Considering only Asimov's sales makes the 1-in-450 or whatever too small, and exaggerates the power of the test. But it does not reduce it to an inverse correlation or even just uncorrelated.
Nor may it be too much of an exaggeration. Asimov was very popular - he was the iconic SF writer, bigger than Heinlein. My father was not an Asimov fan, yet he still had like a dozen of his books in his SF collection. I'm not a big fan of him either (his fiction writing is terrible as literature), yet I read them all because to be SF-literate, I had to know Asimov's ideas like the Laws of Robotics.
3) These two points exacerbate one another. Note the shifting of goal posts. "Half of scientists have some interest in science fiction" versus "only a small fraction of adults have bought my book."
A minimal goalshift per #2; Asimov does not have detailed surveys available about what SF, exactly, the creative scientists were reading. (I am surprised you object to the use of Asimov and not the list of prescribed authors he gives like Budrys or Heinlein.) As a matter of fact, I do not think such information is available now, 50 years later.
Marginal Revolution linked today an old 1963 essay by Isaac Asimov, who argues that a very cheap test for scientific capability in children & adolescents is to see whether they like science fiction and in particular, harder science fiction, "The Sword of Achilles".
I copied it out and made an HTML version of the essay: http://www.gwern.net/docs/1963-asimov-sword-of-achilles
I'd be interested if anyone knows of better tests for such scientific aptitude.
I think it'd also be interesting to see how well the SF test's predictive power has held up. Asimov's numbers seem reasonable for 1963, but may be very different these days: perhaps SF readers back then were <1% of the population and >50% of scientists, so it was a very informative, but these days? SF seems more popular, even discounting the comic books and Hollywood material as Asimov explicitly does, but the SF magazines are mostly dead and my understanding is that scientists are a vastly larger group in 2011 than 1963, both in absolute numbers and per capita.