I notice there's a whole lot of confusion in the mainstream media when scientific results are published; people don't seem to know how to interpret findings, or what a lot of the terms mean (standard deviation, causation versus correlation) or how a GOOD study should be set up (placebos, double-blinds, possibility of influence by other factors such as corporate sponsorship, how many subjects should be used to get a reasonably accurate result, how data is collected) or what to look for when something is probably sketchy (sample size, lacks citations from other research, data comes from surveys, ...?) and common flaws found in many studies (same as "sketchiness indicators"?).
As you can tell, I know some of these things, or at least have vague memories of them from school, but not enough to really write the entry!
This prompt might benefit from multiple answers, in fact -- I'm pretty sure that a reasonable number of subjects could be "10,000" when studying mice or corn plants, but "50" when studying people with heart murmurs... likewise it's probably silly to give a placebo to a cat, but sensible to a human. (At least if it's a drug study. If it's a study of, say, allele frequencies in population X, it's not going to be at issue at all.)
![[community profile]](https://www.dreamwidth.org/img/silk/identity/community.png){kind=small}
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
zvi and
PROMPT: interpreting study results (for beginners)
Date: 2011-04-29 02:55 am (UTC)As you can tell, I know some of these things, or at least have vague memories of them from school, but not enough to really write the entry!
This prompt might benefit from multiple answers, in fact -- I'm pretty sure that a reasonable number of subjects could be "10,000" when studying mice or corn plants, but "50" when studying people with heart murmurs... likewise it's probably silly to give a placebo to a cat, but sensible to a human. (At least if it's a drug study. If it's a study of, say, allele frequencies in population X, it's not going to be at issue at all.)