This thesis has two main foci. First, it seeks to outline how recent findings from the genetics and neuroscience literature can enrich both the empirical and theoretical understanding of economic behavior. Second, it presents models of health/education investment in teens as well as a special case of the Becker-Murphy model of smoking that treats adolescence separately. To address the hypotheses generated by these models, a unique and rich longitudinal dataset of 1000 high school students that includes detailed information on smoking and genetic markers is analyzed. Two staged least squares models that use genetic markers as instruments are used to infer causality from health to academic achievement. The main findings from this thesis are: (1) Genetic markers possess good statistical properties as instrumental variables that can yield rich insights when included in estimation. (2) The impact of health status on educational outcomes varies greatly by gender. This difference becomes more striking once endogeneity of health outcomes is accounted for. (3) Co-morbid health conditions complicate results and present an additional hurdle for empirical researchers. (4) Among boys with ADHD and girls who are diagnosed as obese, cigarette smoking is correlated with higher GPA, although low sample size may bias those results. (5) Models of addiction do not take into account changes in the adolescent brain that can impact addictive behavior. (6) Though data limitations make it impossible to empirically test the predictions of the smoking model, future data sets should allow for such analysis.