The HDL data set contains the results of an experiment investigating the effect of exercise activity and weight on HDL cholesterol. The study involved 26 subjects randomly allocated to three treatment groups. Eight subjects are assigned to a control group (variable GROUP = 'control'), eight to a group that participates in a relatively rigorous running program (GROUP = 'run'), and ten to a group that both runs and lifts weights (GROUP = 'run/lift'). The weight and HDL cholesterol of each subject are recorded after ten weeks on the program.
This data set is particularly suited for analysis of covariance, with HDL as the response variable, GROUP as the classification variable, and WEIGHT as the covariate.