This paper analyzes various empirical methodologies on policy evaluations in the labor market. It focuses especially on the methods which are easily understandable and conveniently implemented by those non-specialists in economics or statistics. First, the paper explains the selection bias, the intrinsic problem in evaluating the labor market policies. It also reviews theoretical approaches to solve the selection problem. Second, it provides detailed explanations on propensity score and matching model, double-difference model, selection model and instrumental variable model. Finally, it recommends the best and/or easiest method in a given situation and dataset.