This document discusses evaluation of evidence and common problems in observational studies. It describes the role of chance and how statistical significance and p-values are used to evaluate chance. Bias such as selection bias and information bias are discussed. Selection bias can occur through inappropriate selection of study subjects. Information bias arises from mistakes made during data collection. Confounding occurs when a variable is associated with both an exposure and an outcome. Methods to control for confounding include randomization, restriction, matching, and adjustment in analysis. Criteria for evaluating causal relationships include strength of association, dose-response relationship, consistency, temporality, specificity, plausibility, and prevention.