Scary Psych Study

[judith_s]

Apparently women who succeed in male dominated fields are seen as unlikeable and pushy. Even worse, if the performace indicators were ambiguous the male employee was rated as competent and the female employee as floundering.

I know it's a very small study, but it seems to have been properly designed. I hope it's repeated with much larger and more diverse groups. I think it's very interesting. Also, quite depressing. We can't win for losing.

Comments

it's not the first or second study to show this finding...

[mamawrites.livejournal.com]

Madeline Heilman is a very famous and distinguished psychologist with a long list of publications, many showing similar biases.

http://www.psych.nyu.edu/heilman/

Re: it's not the first or second study to show this finding...

[gizbot.livejournal.com]

Cool. Now we need a couple of repeats by psychologists trying to show an opposite bias. :)

Re: it's not the first or second study to show this finding...

[xthread.livejournal.com]

Oh, hey, looking for truth by testing the opposing hypothesis is just cheating. I mean, what are you going to do next, propose actual use of the scientific method?

[xthread.livejournal.com]

Why do you say it seems to have been properly designed? Their sample set was 48 grad students for Pete's sake.

[judith-s.livejournal.com]

Proper design, IMHO, means it's a double blind study, where the subjects don't know what is being tested, and there is no official bias introduced. And yes, it was too small a study. But it was a properly designed study, as far as I can tell. If I'm wrong, tell me why.

[xthread.livejournal.com]

My complaint is rooted only sample size - fifty people is so few that there's no way to be derive statistically valid results off of that small a population, no matter how good the rest of the study. If they were doing this for commercial use, they'd be using a sample size of thousands. This is sociology, we need statistically relevant numbers.

[judith-s.livejournal.com]

Unfortunately, psych studies that use significant numbers of subjects are few and far between. The average study size seems to be between 20 and 200. I haven't seen a large size study in quite some time. I expect this is because of weird "must protect the subjects" combined with lack of funding.

[xthread.livejournal.com]

That makes sense and seems entirely reasonable, but still doesn't change the basic statistical problem - if your sample set is too small to predict the population, your sample set is still too small to predict the population. Even if the results are overwhelming, you still don't become able to predict the large population with a very small size.

It is worth noting that if the results of the study are overwhelming, in which case you still can't make statements about the overall population, but you can make clearer statements about the portion of the population that your sample reflects.