An analysis of covariance model was used to test for
differences between the intervention and nonintervention
groups on posttest outcome scores while controlling for
outcome scores at pretest. We also examined the potential
moderating effects of LOF (at baseline) by including
interaction terms in the model.