This chapter describes the methods of assessing the exposure-outcome relationship in a bivariate fashion, including modeling assumptions and regression techniques. It reviews key assumptions in statistical analysis within the framework of descriptive epidemiology, and then transitions to analytic epidemiology. The foundation of the majority of epidemiological analyses is regression techniques. Regression is a predictive modeling method, where given the value of one or more independent variables, the change in the dependent variable is estimated. The chapter depicts the relationship between two continuous variables in the birthwt dataset: maternal weight at the last menstrual period ("lwt") and infant's birth weight ("bwt"). The data in birthwt represent a cross-sectional study of infants at a medical center and can be inferred to represent the secondary data derived from an electronic health record (EHR). More sophisticated sensitivity techniques include mitigating bias, imputing missing data, and adjusting for unaccounted confounding.