Assessing the Benefits of a Parenting Skills Training Program: A Theoretical Approach to Predicting Direct and Moderating Effects* Using a theoretical model to ground this investigation, specific hypotheses about factors that moderate the benefits of attending the Preparing for the Drug-Free Years (PDFY) program were tested. PDFY is a skills-training program designed to teach parents and children skills that reduce a child's risk for drug and alcohol use. We hypothesized that high levels of family stress (i.e., marital difficulties or financial concerns) reduce the benefits of program attendance, and that strong pre-program skills (i.e., parental communication, parental negativity, or parent-child relationship quality) increase the benefits of program attendance. These hypotheses were experimentally tested on a sample of families that each included a sixth or seventh grade child. The results for fathers (N = 144) supported the study hypotheses, while mothers (N = 150) who benefited most from program participation showed the weakest pre-program communication skills and reported the greatest marital difficulties. Family interactions, and parental behavior in particular, are important factors in determining a child's risk for initiation of substance use (c.f., Barnes, Farrell, & Cairns, 1986; Coie et al., 1993; Conger, Rueter, & Conger, 1994; Coombs & Landsverk, 1988; Hawkins, Catalano, & Miller, 1992; Patterson & Dishion, 1985; Yoshikawa, 1994). In an effort to reduce children's risk for substance use or abuse, prevention specialists have developed numerous family-oriented education programs. For the most part, these programs involve the parents and sometimes children in learning parent-child interaction skills expected to strengthen family relationships and thus enhance a family's ability to protect youth from involvement with alcohol, tobacco, or other drugs. A handful of the available programs have been scientifically evaluated. Reports of these evaluations usually suggest that, for parents as a whole, attendance at a skills-training program can bring about the desired improvements in parental skills and family interactions (c.f., Andrews, Soberman, & Dishion, 1995; DeMarsh & Kumpfer, 1985; Friedman, 1989; Grady, Gersick, & Boratynski, 1985; Kosterman, Hawkins, Haggerty, Spoth, & Redmond, 1996; Kosterman, Hawkins, Spoth, Haggerty, & Zhu, 1997; Serna, Schumaker, Hazel, & Sheldon, 1986; Spoth, Redmond, Haggerty, & Ward, 1995; Spoth, & Redmond, 1996). Educators and evaluators alike, however, have long recognized that no two individuals benefit from a learning experience in the same way (c.f., Beutler, 1991; Cronbach & Snow, 1977; Dance & Neufeld, 1988; Finney & Moos, 1989; Shoham-Salomon & Hannah, 1991; Smith & Sechrest, 1991; Snow, 1991; Spoth & Redmond, 1996; Webster-Stratton, 1997). Various factors, ranging from a person's individual characteristics to his or her life circumstances, can moderate one's ability to learn in a given setting. Unfortunately, the moderating role of personal characteristics and life contexts is too often ignored in the evaluation of parent training programs. Yet, through identification of conditions that facilitate or interfere with a parent's ability to learn, training programs can be optimally tailored to meet the needs of participants. Of course, the list of potential moderating characteristics or conditions is almost endless. For this reason, researchers have long emphasized the need to rely on a theoretical model to guide both the selection of moderating factors and the interpretation of research results (Beutler, 1991; Cronbach & Snow, 1977; Finney & Moos, 1989; Shoham-Salomon & Hannah, 1991; Smith & Sechrest, 1991). The model used to guide this investigation (see Figure 1) draws upon two previously defined theoretical frameworks. The basic form of the model closely resembles Finney and Moos' (1989) Framework for Evaluation Research. …