Degree to which a sample is not characteristic of the populace from which it was subsequently taken. Variance in the approximation of a parameter that happens from sample to sample. Expected margin of error that happens in research utilizing sampling