Design Matters

Key steps for choosing a proper study design are understanding the scientific aims of the study and what is required to achieve them. In this seminar we discuss some pitfalls in study designs and analysis choices, including regression to the mean or change score analysis. It is important to understand what sources of variation exist in your data and how your study design can address it. Threats to (internal) validity are measurement error, selection bias, imprecise estimation, confounding, missingness, and inappropriate analysis. Some design mistakes can be irremediable.