Bayesian modeling has received little attention in quantitative methods courses in the social sciences, where the frequentist or null hypothesis significance testing (NHST) approach prevails. We believe our students need exposure to the basic concepts of Bayesian modeling as it (a) better corresponds to the manner in which scientific theories advance and (b) provides the fundamentals for many current state-of-the-art methods of data science including machine learning. This paper proposes a class activity whereby students can learn how Bayesian analysis works to produce information about predictions and how it differs from NHST. During the activity, students evaluate twenty pieces of evidence to assess the probability for a fictitious suspect to be innocent (H?) or guilty (H₁), first using NHST and then the Bayesian approach.