UCLA

Home court advantage is an often confused term in sports literature that encompasses a variety of factors. Our goal is to identify and quantify home court advantage in the National Basketball Association within a causal framework, and to conduct an extended sensitivity analysis to measure the robustness of our estimated effect. The traditional omitted variable bias framework succeeds in providing an interpretable way of approaching sensitivity analysis, however it also introduces subjectivity in interpreting sensitivity plots and does not extend to multiple confounders very well. By reparameterizing the omitted variable bias framework in terms of R2 we can use expert knowledge to formally bound the strength of unobserved confounders and provide a more systematic way to conduct sensitivity analyses. We use the imbalance in number of rest days between home and away teams to gauge the robustness of our estimated home court advantage effect.