“…Preliminary analyses showed marginally positive (and mostly nonsignificant) correlations among these factors across the three informants (for the Diffusion sample, average r = .19; for the CDA sample, average r = .18). These results correspond with those obtained by other child process researchers who found poor agreement in therapy process ratings across sources (e.g., Eltz, Shirk, & Sarlin, 1995;Shirk & Saiz, 1992;Smith-Acuna, Durlak, & Kaspar, 1991) and suggest that attempts to find common variance across informants might result in an adherence construct with little practical meaning. For this reason, we decided to evaluate the two core models (i.e., family functioning/cohesion and parent monitoring) using alternative measures of adherence (i.e., caregiver, youth, and therapist).…”