Quote:
Originally Posted by Luckbox Inc
I'm intrigued. What are you getting at? The discussion has taken enough turns that this is pretty vague.
Take any measurement you might like (ex: # of casual partners) about two populations (ex: gay guys and straight guys). A low level comparison is just to compare the means and make a claim like that gay guys have more casual partners on average than straight guys. But you can also look at the distributions, and the standard deviation is one way to do this which is a measure of how spread out the distribution is. A population with a large standard deviation would have large numbers of people far away from the mean. One common mistake people make when focusing on means and not standard deviations is if there is a tonne of variability within populations such that the difference between the means isn't really all that important.
Our whole example of gay people and casual partners is pretty cringe and getting tangential to long running attacks against gay communities, but you can make the same point when comparing something like Canadians vs Americans that are probably a bit different in their behaviours but the variance within the countries is a lot more of an important story than the difference between the two population means.