We expected that each treatment would reduce child symptoms
and improve prosocial functioning and that the combined
treatment would show more marked and consistent changes
across child, parent, and family measures. To examine changes
over time, within-group t tests were computed separately for
each group. To examine group differences, analyses of covariance
(ANCOVAs) were used to obtain the posttreatment and
follow-up means. Newman-Keuls comparisons (p < .05) were
used to control familywise error rates to test the primary prediction,
namely, that the combined treatment was different
from that of each of the other two treatments.