This short lecture from Yale's online course Understanding Medical Research: Your Facebook Friend is Wrong walks through how researchers decide which statistical test fits a given study design. The instructor lays out a decision process built around the type of data collected, continuous, categorical, or ordinal, and whether the comparison involves two groups, more than two groups, or paired measurements. Common tests get matched to their use cases: t-tests for comparing two means, chi-square for categorical outcomes, ANOVA when there are multiple groups, and nonparametric alternatives when data don't meet normality assumptions. The goal is practical: giving a general audience enough vocabulary to recognize when a published study's statistics actually support its headline claim, tying back to the course's broader project of teaching people to read medical research critically rather than take a news summary at face value.