http://www.mediaite.com/online/yale-...comprehension/
So a Yale professor came forward with a study that says Tea Party members are statistically significantly more scientifically literate than other Americans. (!!!) I listen to conservative talk radio all day and naturally this result was trumpeted by hosts who are, ironically, scientifically illiterate and don't know what statistical significance actually means. In particular, in medical and behavioral research, one—or at least the more competent researchers—often speaks of "practical significance" as opposed to "statistical significance". "Statistical significance" is finding an effect, something that is incidentally a lot easier—even trivially easy—to do with large samples, which the author had. "Practical significance" is having an effect that actually matters. For instance, if I am a researcher developing a penis extension pill and find a statistically significant effect, but of only less than a millimeter in added length, then it is not practically significant. And indeed upon visual inspection, the test results on scientific literacy from Tea Partiers do not differ markedly from the results for the wider population. So practical significance is nowhere to be found here.
As mentioned in the article, the researcher said that the effect was "absolutely trivial" and suggests that he would not be able to replicate it. Of course, we're talking about people who think that mean reversion in summer Arctic thaw means climate change is a hoax. Conservative talking heads will never, and I mean never understand the difference between statistical and practical significance, especially insofar as their bias makes them believe otherwise.
So a Yale professor came forward with a study that says Tea Party members are statistically significantly more scientifically literate than other Americans. (!!!) I listen to conservative talk radio all day and naturally this result was trumpeted by hosts who are, ironically, scientifically illiterate and don't know what statistical significance actually means. In particular, in medical and behavioral research, one—or at least the more competent researchers—often speaks of "practical significance" as opposed to "statistical significance". "Statistical significance" is finding an effect, something that is incidentally a lot easier—even trivially easy—to do with large samples, which the author had. "Practical significance" is having an effect that actually matters. For instance, if I am a researcher developing a penis extension pill and find a statistically significant effect, but of only less than a millimeter in added length, then it is not practically significant. And indeed upon visual inspection, the test results on scientific literacy from Tea Partiers do not differ markedly from the results for the wider population. So practical significance is nowhere to be found here.
As mentioned in the article, the researcher said that the effect was "absolutely trivial" and suggests that he would not be able to replicate it. Of course, we're talking about people who think that mean reversion in summer Arctic thaw means climate change is a hoax. Conservative talking heads will never, and I mean never understand the difference between statistical and practical significance, especially insofar as their bias makes them believe otherwise.
via JREF Forum http://forums.randi.org/showthread.php?t=267099&goto=newpost
Aucun commentaire:
Enregistrer un commentaire