To control the internal validity of the study, as
well as decrease the problem of alternative hypotheses,
a degree of control was imposed on the extraneous
variables that might confound the data of relapse. The
key extraneous variables of concern were: medication nonadherence;
partial remission; previous depressive
episodes; psychiatric comorbidity; and, symptom
severity. The researchers screened all potential subjects
for these factors and eliminated those from the study
who had any of these factors present.