How Did It Work? Who Did It Work for? Mediation in the Context of aModerated Prevention Effect for Children of DivorceJenn-Yun Tein, Irwin N. Sandler, David P. MacKinnon, and Sharlene A. WolchikArizona State UniversityThis study presents a reanalysis of data from an effective preventive intervention for children fromdivorced families (S. A. Wolchik et al., 2000) to test mediation of program effects. The study involved157 children, age 9–12 years, who were randomly assigned to a parenting program or a literature controlcondition. Program effects to reduce posttest internalizing problems were mediated through improvementin mother–child relationship quality. Program effects to reduce externalizing problems at posttest and 6months were mediated through improvement in posttest parental methods of discipline and mother–childrelationship quality. The study also describes a new methodology to test mediation of Program ⫻Baseline Status interactions. Analyses demonstrate mediation effects primarily for children who beganthe program with poorer scores on discipline, mother–child relationship quality, and externalizingproblems.Many prevention programs are based on a theory of the mech-anisms that mediate their effects on problem outcomes (Lipsey,1990; Sandler, Wolchik, MacKinnon, Ayers, & Roosa, 1997). Thetheory typically proposes that the intervention program affects aproximal variable (e.g., discipline) and that change in that variableleads to change in the problem outcomes (e.g., externalizing prob-lems). Variables that represent the mechanisms by which a pro-gram affects a problem outcome are referred to as mediators of theprogram effects. Methods for analyzing mediators of programeffects have advanced rapidly over the past decade (e.g., Holm-beck, 1997; Kraemer, Wilson, Fairburn, & Agras, 2002; MacKin-non, Lockwood, Hoffman, West, & Sheets, 2002), and preventionprograms have identified mediators of program effects on prob-lem outcomes (e.g., Komro et al., 2001; Spoth, Redmond, &Shin, 1998). However, mediation analysis is complicated whenthe effects of a program are moderated by another variable (i.e.,differ as a function of individuals’ scores on another variable).For example, evaluations commonly report that prevention pro-grams are more beneficial for those with more problems at thestart of the program (Brown & Liao, 1999; see also Pacifici,Stoolmiller, & Nelson, 2001; Stoolmiller, Eddy, & Reid, 2000).In this article, we have two objectives. First, we examine me-diators of the effects of the New Beginnings Program (NBP), apreventive intervention for divorced families that found both maineffects and Program ⫻ Baseline Status (moderated) effects toreduce children’s mental health problems (Wolchik et al., 2000).Second, because to our knowledge this is the first article in theliterature to report on the analysis of mediation of preventiveeffects in which the program effects were moderated by baselinestatus rather than just by main effects (Baron & Kenny, 1986), wediscuss the methodology in length as well as the conceptual issuesin testing these effects.NBP: A Theoretically Based Preventive InterventionThe NBP (Wolchik et al., 2000) was designed to change poten-tially modifiable correlates that, demonstrated by research, wereassociated with mental health problems for children of divorce.The small theory (Lipsey, 1990) of the intervention is thatprogram-induced change in these variables would lead to (i.e.,mediate) program-induced change in children’s mental healthproblems. We assessed whether changes in four variables targetedby a preventive intervention program for custodial mothers(mother program [MP]) mediate the previously reported effects ofthe MP relative to a self-study literature control (LC) condition1(Wolchik et al., 2000) to reduce internalizing and externalizingproblems: (a) mother–child relationship quality (e.g., Hethering-ton et al., 1992), (b) effective discipline (e.g., Forgatch & De-1The third condition in the trial was a multicomponent MPCP. Thiscondition was compared with the MP to test additive effects; no significantdifferences were found between the effects of the two programs on out-come measures.Jenn-Yun Tein, Program for Prevention Research, Arizona State Uni-versity; Irwin N. Sandler and Sharlene A. Wolchik, Program for PreventionResearch and Department of Psychology, Arizona State University; DavidP. MacKinnon, Department of Psychology, Arizona State University.Support for this research was provided by National Institute of MentalHealth Grant P30 MH39346 to establish the Preventive Intervention Re-search Center at Arizona State University, National Institute of MentalHealth Grant R01 MH057013–01A1 to evaluate a preventive interventionfor children of divorce, and Public Health Service Grant DA09757 fordevelopment and application of methods to assess mediation. We thank themothers and children for their participation as well as the group leaders fortheir careful and thoughtful delivery of the program.Correspondence concerning this article should be addressed to Jenn-YunTein, Program for Prevention Research, Arizona State University, P.O.Box 876005, Tempe, AZ 85287-6005. E-mail: [email protected] of Consulting and Clinical Psychology Copyright 2004 by the American Psychological Association2004, Vol. 72, No. 4, 617–624 0022-006X/04/$12.00 DOI: 10.1037/0022-006X.72.4.617617Garmo, 1999), (c) father access to child2(e.g., Braver et al., 1993),and (d) interparental conflict (e.g., Grych, Seid, & Fincham, 1992).As described by Wolchik et al. (2000), at posttest and at 6-monthfollow-up, children with higher baseline externalizing problemsbenefited more from the MP than did those with lower baselineexternalizing problems (i.e., a significant Program ⫻ BaselineStatus interaction effect). There was also a significant main effectof the program to reduce internalizing problems at posttest. Sig-nificant Program ⫻ Baseline effects were also found for thefollowing theoretical mediators: mother–child relationship quality,mother’s attitude toward father’s visitation and father–child rela-tionship, and interparental conflict. The MP had a more beneficialeffect for families who had worse baseline scores on each variable.In addition, there were significant main effects for the MP toimprove effective parental discipline and behavioral observationmeasures of attending to and validating child conversation contentas compared with the LC.Theory