Five of our school's math teachers decided not to attend the last 3 hours of the unconscious bias training a few weeks ago, so this past Thursday we met for an hour to review the statistics standards that are imbedded in the Common Core Integrated Math 1, 2, and 3 courses that our district is switching to next year. Believe it or not, standard deviation is mentioned in the Integrated 1 standards--freshman math! To be honest, we're not sure exactly what level of detail we're supposed to address in these standards, but it's best anyway if teachers know significantly more about a subject than merely what they're supposed to teach, so we spent our first hour (two more to go!) reviewing introductory statistics.
One of the foundational concepts in inferential statistics is the difference between a population and a sample. If you want to learn something about a population, you take a sample from it and infer about the population from that sample. The formulas for calculating the standard deviations of a population and a sample are slightly different; if we were to use the "population" formula for a "sample", that sample would always underestimate the true standard deviation. Therefore the formula has to be tweaked a little bit, the value inflated, so that the sample standard deviation becomes a good predictor of the population standard deviation.
This tweaking makes the sample standard deviation what is known as an unbiased estimator for the population standard deviation.