I don't know what's funnier, that you're citing Wiktionary as a source, or that you're suggesting "university level" philosophy courses to a statistician.;-)
If you're really that interested in pedantry, though, I might suggest Bartleby on the meaning of average as compared with median or mode. HTH, HAND, &c.
No, SidewaysPineTree was correct. The median is a measure of central tendency (as is the mean), but 'mean' is another word for average, 'median' isn't.
And by the definition of median (the data point in the middle when all the data are ordered), 50% of the population is below the median. 50% of the population is only below the mean when you have perfectly symmetrically distributed data, which happens not nearly as often as most analysts would like.;-)
I get the feeling that the congresscritters responsible for BS bills like this don't know much about PhD programs, much less those in the hard sciences or engineering. Nothing like a 5-to-7-year trial run to convince women* to run for the hills rather than shackle themselves to the rat race that is modern faculty life...
Quotas aren't even a quick fix for a system as broken as academe. I don't think there is one, personally.
(*The PhD experience should, in theory, demoralize men and women similarly, but I do think there's something to the idea that women are under far more societal pressure than men to find and maintain a suitable work-family balance -- erring on the side of family, always -- with or without children in the picture.)
It seems very fishy to me that anybody would be conducting studies, then deciding whether to publish them or not based on their conclusions. Welcome to science. There, fixed that for ya.
In ANY given field you'll find studies that disagree with most other studies. Of course. The solution to those is meta-analysis (using statistical techniques to combine results from many disparate studies), which...is exactly what these authors did. It's not "one small study."
And for all we know this study could've been funded by a company whose main competition is anti-depressants Oh come on. It says right there in the paper, "Funding: The authors received no specific funding for this study." I don't know about the fifth author, but everyone else on the paper is university-affiliated -- probably just your typical professors and grad students out to up their publication count, preferably with a nice sexy (controversial) paper that will get cited a lot.
Look, I'm troubled by these results too -- people very, very important to me have benefited greatly, I dare say in life-saving ways, from antidepressants, and having seen some of the very physical side effects I'm disinclined to think it's something as simple as a placebo effect. If you're going to go after something, consider that the study authors didn't do moderator analyses to test whether results differed for men and women, or based on the mean age of the samples, or (as one poster noted above) whether talk therapies were administered in addition to drugs. But the methodology in this study is sound, and impugning the authors for imagined conflicts of interest is just cheap.
So where is the Shakespeare or Bach of gaming? Try ringing up Alexander O. Smith. His work on Vagrant Story and the Occuria parts of FFXII are as literal a take on Shakespeare as you can get. If you're more after the Shakespearean concept than the language per se, I think the Phoenix Wright series comes pretty close in terms of applying clever twists and adaptations of modern language to tell an unexpectedly deep story.
Scurvy effectively doesn't exist in our society, and we are in absolutely no danger of forgetting why. Yeah, that was a pretty great episode of House...
I do too -- and appreciate your nod toward inclusiveness at the end there -- but it's a little late for the Jewish/. tech geeks: the last night of Hanukkah was just under a week ago.
(But I finally got my Roomba, so I can't complain.):D
Hey, don't blame all psychologists for Carol Dweck! *g*
You're absolutely right, though. 'Theories of Intelligence' as conceptualized by Dweck actually refer to theories of task performance -- one of the motivational properties that has long-term consequences is whether an individual believes that one's performance on a given task is fixed (decided by ability; increased effort won't yield results) or incremental (malleable; increased effort may yield results). I don't know what in the hell possessed her to call it "intelligence," because any psychologist worth his/her salt would conceptualize actual, IQ-measurable intelligence as fixed, but intelligence is of course not the sole predictor of performance on *any* task.
What a disappointment, there wasn't anything 'psychological' about this analysis -- contrary to popular belief, mention of angst does not psychology make.:D
More's the pity, because psychology is (as always) a few years behind the times, but some work is finally starting to be done on the real principles governing social networking behavior. Wendi Gardner and one of her graduate students at Northwestern, whose name I am chagrined to admit I cannot recall, have some work in online social perception (though I don't believe it's published yet), and a couple of folks at Berkeley are studying online dating, but I haven't yet seen any good empirical research on Facebook and its ilk.
TFA's conclusion is correct but their methods are wrong. For these kind of data, correlations aren't the appropriate test; they should have used a chi-square distribution test.
Using TFA's assumptions -- total sample size of 50, 4 yawners out of 16 not seeded, 10 yawners out of 34 seeded -- the chi-square value is.10, which pretty strongly misses the critical value of 3.84 for significance.
Not that it matters anyway, but it's pretty funny to read an article debunking statistics that employs inappropriate statistics itself...
I don't think *you're* familiar with pharma studies. Ever heard of IRBs? How about data monitoring boards? The process is structured exceptionally rigorously -- you have to specify your outcomes before you run the study, not after, or you can be accused of "fishing" for significant results. So if your primary outcome is, say, whether Prozac lowers incidence of depressive episodes, then that's what you investigate. You don't go digging through the data to see if it harms pregnant women (and usually they're excluded from studies anyway), or raises suicidality in teenagers, or... People do follow-up studies on subgroups and side effects, before and after the drug goes to market, but research takes time and the company is motivated to get their product out as fast as possible.
Two colleagues of mine who have both served extensively on data monitoring committees for clinical trials have seen fewer than 5% of the drugs being tested get eventual approval by the FDA, invariably because the committee has to shut down the trial for one reason or another. Doesn't sound to me like a failure of "independence"; what are your numbers?
Slashdot Mirror
I don't know what's funnier, that you're citing Wiktionary as a source, or that you're suggesting "university level" philosophy courses to a statistician. ;-)
If you're really that interested in pedantry, though, I might suggest Bartleby on the meaning of average as compared with median or mode. HTH, HAND, &c.
No, SidewaysPineTree was correct. The median is a measure of central tendency (as is the mean), but 'mean' is another word for average, 'median' isn't.
;-)
And by the definition of median (the data point in the middle when all the data are ordered), 50% of the population is below the median. 50% of the population is only below the mean when you have perfectly symmetrically distributed data, which happens not nearly as often as most analysts would like.
Hey, I saw that episode of House too!
I get the feeling that the congresscritters responsible for BS bills like this don't know much about PhD programs, much less those in the hard sciences or engineering. Nothing like a 5-to-7-year trial run to convince women* to run for the hills rather than shackle themselves to the rat race that is modern faculty life...
Quotas aren't even a quick fix for a system as broken as academe. I don't think there is one, personally.
(*The PhD experience should, in theory, demoralize men and women similarly, but I do think there's something to the idea that women are under far more societal pressure than men to find and maintain a suitable work-family balance -- erring on the side of family, always -- with or without children in the picture.)
I'd even suggest 'mondegreen' is anti-geek at this point, when one can simply Google lyrics that one is unsure of or may have heard incorrectly.
Look, I'm troubled by these results too -- people very, very important to me have benefited greatly, I dare say in life-saving ways, from antidepressants, and having seen some of the very physical side effects I'm disinclined to think it's something as simple as a placebo effect. If you're going to go after something, consider that the study authors didn't do moderator analyses to test whether results differed for men and women, or based on the mean age of the samples, or (as one poster noted above) whether talk therapies were administered in addition to drugs. But the methodology in this study is sound, and impugning the authors for imagined conflicts of interest is just cheap.
I do too -- and appreciate your nod toward inclusiveness at the end there -- but it's a little late for the Jewish /. tech geeks: the last night of Hanukkah was just under a week ago.
:D
(But I finally got my Roomba, so I can't complain.)
Guess they can afford to after selling LiveJournal to the tune of about $30 mil.
Hey, don't blame all psychologists for Carol Dweck! *g*
You're absolutely right, though. 'Theories of Intelligence' as conceptualized by Dweck actually refer to theories of task performance -- one of the motivational properties that has long-term consequences is whether an individual believes that one's performance on a given task is fixed (decided by ability; increased effort won't yield results) or incremental (malleable; increased effort may yield results). I don't know what in the hell possessed her to call it "intelligence," because any psychologist worth his/her salt would conceptualize actual, IQ-measurable intelligence as fixed, but intelligence is of course not the sole predictor of performance on *any* task.
What a disappointment, there wasn't anything 'psychological' about this analysis -- contrary to popular belief, mention of angst does not psychology make. :D
More's the pity, because psychology is (as always) a few years behind the times, but some work is finally starting to be done on the real principles governing social networking behavior. Wendi Gardner and one of her graduate students at Northwestern, whose name I am chagrined to admit I cannot recall, have some work in online social perception (though I don't believe it's published yet), and a couple of folks at Berkeley are studying online dating, but I haven't yet seen any good empirical research on Facebook and its ilk.
TFA's conclusion is correct but their methods are wrong. For these kind of data, correlations aren't the appropriate test; they should have used a chi-square distribution test. Using TFA's assumptions -- total sample size of 50, 4 yawners out of 16 not seeded, 10 yawners out of 34 seeded -- the chi-square value is .10, which pretty strongly misses the critical value of 3.84 for significance.
Not that it matters anyway, but it's pretty funny to read an article debunking statistics that employs inappropriate statistics itself...
I don't think *you're* familiar with pharma studies. Ever heard of IRBs? How about data monitoring boards? The process is structured exceptionally rigorously -- you have to specify your outcomes before you run the study, not after, or you can be accused of "fishing" for significant results. So if your primary outcome is, say, whether Prozac lowers incidence of depressive episodes, then that's what you investigate. You don't go digging through the data to see if it harms pregnant women (and usually they're excluded from studies anyway), or raises suicidality in teenagers, or... People do follow-up studies on subgroups and side effects, before and after the drug goes to market, but research takes time and the company is motivated to get their product out as fast as possible. Two colleagues of mine who have both served extensively on data monitoring committees for clinical trials have seen fewer than 5% of the drugs being tested get eventual approval by the FDA, invariably because the committee has to shut down the trial for one reason or another. Doesn't sound to me like a failure of "independence"; what are your numbers?