In the prior episode in this series, Dan described how to evaluate the overall predictive utility of the regression model. In this episode, Dan continues his discussion of inference but now focuses on the evaluation of individual parameter estimates...

He begins by reviewing the distinction between population parameters and their corresponding sample estimates. He then describes two ways of making inferences about the population parameters using sample estimates: conducting null hypothesis tests and computing confidence intervals. He illustrates the logic and interpretation of both approaches using real data on the relationship between alcohol consumption and psychomotor impairment (in which the population parameters are unknown) as well as simulated data (in which the population parameters are known).