ABSTRACT
This chapter presents common statistical tests that a practitioner can use for determining the statistical significance of an experiment. Statistical tests are designed to work for specific problems and with specific variables. The chapter explores some of the basic tests for comparing the means of two populations, evaluating the association of two category variables, comparing the proportions of two populations, and comparing the means across many categories. Tests for continuous outcomes apply when there is a single numeric variable. Two-sample t tests are appropriate for testing hypotheses about the average numeric values across two different groups. Analysis of variance is a generalization of the two-sample t test to handle situations where there are more than two groups. The tests are useful in determining whether improvements in key metrics like shrink, return rate, and employee turnover from various loss prevention initiatives are really different or whether the difference is simply due to random noise.
