(This article was originally published at Statistical Modeling, Causal Inference, and Social Science, and syndicated at StatsBlogs.)
David Hogg points to this excellent science-based news article by Martha Gill. It’s not about astrology at all; rather, Gill alludes to scientific findings correlating different attributes with time-of-year of birth. At least, I assume this is science-based. It’s possible that Gill is just making it all up (or, for that matter, that the author of the article is not named “Martha Gill” at all) but something about the article’s tone makes me think it’s legit.
Astrologers say that people born in different seasons have different capabilities and traits. So does statistical scientific research. The only fundamental differences are (a) statistical researchers have better methods, and (b) scientists use different causal models (the models rarely involve Jupiter, for example).
I’d add another difference. Scientists, because of their more reasonable models, have a more reasonable sense of the precision of their claims. Some astrologers will claim the actual date of your birth is relevant, and even the newspaper columns do it by month, but I doubt you’d get that sort of sharp pattern coming from any real effects.
Because I come from a profession with its roots in astrology, this is an issue I have been following for a few years. . . . If you think back to when astrology was born, these effects would have been much larger, because the availability of light, heat, and (most importantly) a balanced diet would be a strong function of month, and have a big effect on pregnancies and infants. So the evidence for astrology might have been enormous in the past.
Maybe . . . but people hardly need evidence to believe in things! It’s enough for the stories to be pleasant, I think.
P.S. I have only two complaints about Gill’s article. First, she doesn’t point to any of her sources. That just sucks. Link, please. My second problem—and, yes, I know I’m being picky—is that she gives a different “horoscope” for each month. Really, though, there’s no way these findings are as precise as to be assigned to particular months. I think it would’ve been more reasonable to have just four categories, one for each season. That must be roughly the precision of these results.
Please comment on the article here: Statistical Modeling, Causal Inference, and Social Science