This seminar focuses not on a specific statistical method, but on the selection of methods. Bushway refers to this process as modeling, the exercise by which the researcher specifies a statistical model that matches or corresponds with the theory, process or phenomenon under study. Two examples of textbook modeling are presented, the first where the model matches a policy question, and the second where the model matches a theoretical approach. Non “modeling” based approaches to the choice of statistical methods are also discussed, including “control freak”, “method tyrant” and the ever popular, “cook book statistics”. Participants will also be encouraged to think about putting theoretical “structure” on their statistical models, essentially creating estimates of what would happen if people acted in a certain ways.