My wife is studying psychology, and sometimes they make her do statistical analysis on data from surveys, etc. She sucks at math so she makes me do the math parts. Often she's told to calculate "the standard deviation" of some set of data. Almost always, this data will be the results of some experiment/survey that's supposed to make a statement about the general population, for example, "we had 50 blacks and 50 whites do an IQ test, here are the results, calculate the average IQ and standard deviation for blacks and whites." In this case, the 50 people are samples representative of their entire races. We're interested in the standard deviation of IQ among all blacks, not just the 50 we happened to select for the study. So obviously the formula for "sample" standard deviation should be used. But it seems like her teachers are always using the formula for "population" standard deviation, which afaik is incorrect, or at least, it doesn't give as accurate an estimate of the true value as the "sample" version does. Are her teachers just ignorant, or is it common/reasonable to use the "population" standard deviation formula against a sample of a population?