Many developmentalists are interested in causal questions, including those concerned with the dosage and timing of exposures experienced repeatedly over time. However, causal inferences are challenging with observational data, and common statistical tools (e.g., regression adjustment) break down in the context of time-varying confounding. This paper introduces one powerful causal inference tool for addressing dosage and timing effects, marginal structural models (MSMs). It provides a conceptual overview, describing the potential outcomes framework, "exposure histories," and inverse-probability-of-treatment weighting. To illustrate, dosage and timing effects of economic strain across infancy, toddlerhood, and early childhood on behavior problems are examined, using the longitudinal Family Life Project (N = 1,292; 49% Female; 58% White). Step-by-step guidance to a novel R package, devMSMs, is provided.