Consider your future practice as a counselor, keeping in mind a population and/or setting for your work. The Final Course Project is a synthesis of literature and it is aimed to help you determine the

There were no interactions between time and treatment format. Contrary to expectation, parents in Group PCIT did not experience greater social support or treatment adherence.Conclusions:Group PCIT was not inferior to individual PCIT and may be a valuable format to reach more families in need of services. Future work should explore the efficiency and sustainability of Group PCIT in community settings.

What is the public health significance of this article?

Children and families who completed group parent– child interaction therapy (PCIT) demonstrated significant reductions in child conduct problems and increases in positive parenting skills that were not inferior to the treatment gains experienced by families in individual PCIT. Group PCIT may offer the potential to increase treatment availability for families in need of services.

Keywords:parent– child interaction therapy, PCIT, childhood conduct problems, group treatment, parent management training The past 40 years have seen meaningful advances in the treat- ment of childhood conduct problems. Parent management training (PMT), based primarily on behavioral principles, is the best-practice treatment for children with oppositional defiant disorder or conduct disorder, and several such interventions have been identified as well supported (Eyberg, Nelson, & Boggs, 2008).

However, now that efficacious treatments exist, the question has become how do we reach enough families to make a significant public health impact? Barriers at the system and family levels prevent many families from accessing services. A shortage of mental health professionals trained in effective interventions means that community agencies are overburdened (Satcher, 2000).

Even when families do enter treatment, many drop out prematurely or do not participate fully, limiting the impact of services. It has been estimated that as many as 67% of children in need of services do not receive them (Kazdin, 2008).

To address problems with access, attrition, and adherence, in- novative treatment delivery models are needed (Kazdin, 2008; Kazdin & Blase, 2011). We took a step to address these treatment barriers by evaluating an adaptation into a group format of the This article was published Online First March 28, 2016.

Larissa N. Niec, Center for Children, Families, and Communities, Cen- tral Michigan University; Miya L. Barnett, Department of Psychology, University of California, Los Angeles; Matthew S. Prewett, Department of Psychology, Central Michigan University; Jenelle R. Shanley Chatham, Institute of Public Health, Georgia State University.

This study was supported by a grant to the first author from the National Institute of Mental Health (MH 070483). The authors acknowledge the research and clinical staff of the Center for Children, Families, and Com- munities, Central Michigan University.

Correspondence concerning this article should be addressed to Larissa N.

Niec, Center for Children, Families, and Communities, Central Michigan University, Mount Pleasant, Michigan 48859. E-mail:[email protected] This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Journal of Consulting and Clinical Psychology© 2016 American Psychological Association 2016, Vol. 84, No. 8, 682– 6980022-006X/16/$12.00http://dx.doi.org/10.1037/a0040218 682 evidence-based intervention parent– child interaction therapy (PCIT). PCIT has innovative characteristics, such as the live coaching of actual parent– child interactions, which make it a powerful intervention, while at the same time creating particular challenges to implementing it in a group format (Niec, Hemme, Yopp, & Brestan, 2005). In this randomized control trial, we explored two primary questions: (a) Do families who complete Group PCIT demonstrate reductions of child conduct problems and increases in positive parenting skills that are not inferior to families who complete individual PCIT? (b) Does Group PCIT offer benefits—specifically, increased parental social support, treatment retention and adherence—relative to individual PCIT? Parent Management Training PMT programs are frequently provided in a group format and have demonstrated efficacy in preventing and reducing children’s conduct problems (e.g.,Pisterman et al., 1989;Ruma, Burke, & Thompson, 1996;Sheeber & Johnson, 1994;Webster-Stratton, Reid, & Beauchaine, 2011). Relative to the individual format of PMT, group parent training may have additional benefits. Because multiple families receive treatment with fewer therapist hours, group interventions offer a potential strategy to increase the avail- ability of services. It has been further proposed that support among group members can decrease parents’ feelings of isolation. When parents develop relationships with families who have similar prob- lems, it may reduce the stigma of having a “problem child” (Webster-Stratton & Herbert, 1993) and increase families’ percep- tions of social support. Group treatment also allows opportunities for therapists to create a culture of positive peer pressure among families that may play a role in engaging parents in treatment and increasing retention (Niec et al., 2005). Little effort has been made to test these hypotheses (Levac, McCay, Merka, & Reddon- D’Arcy, 2008;Webster-Stratton, 1997); however, preliminary ev- idence suggests that the social support provided by group treat- ment may increase attendance (McKay, Harrison, Gonzales, Kim, & Quintana, 2002).

Despite the potential benefits of group parent training, there are limitations to existing models. For example, many parent training groups teach new skills in a didactic or video modeling format, then require parents to implement the skills at home and report on the outcome at the following session. This approach relies on parent reports regarding the implementation of skills and chil- dren’s responses, with all the biases inherent in such reports (Feinberg, Neiderhiser, Howe, & Hetherington, 2001). If therapists do not observe parent– child interactions directly, they cannot assess and correct problems that are unreported, and parents lack the opportunity to practice the techniques with their children in a controlled setting (Herschell, Capage, Bahl, & McNeil, 2008).

PCIT has several unique and innovative characteristics that avoid these limitations. Parent–Child Interaction Therapy PCIT is a PMT program designed to be implemented with individual families to address the conduct problems of children 2 to 6 years 11 months of age (Eyberg & Funderburk, 2011). The first phase of PCIT, Child-Directed Interaction (CDI), teaches parents to use child-centered skills such as labeled praises anddescriptions of their children’s appropriate behaviors, reflections of appropriate verbalizations and imitation of appropriate play. At the same time, parents are taught to avoid leading verbalizations such as questions, commands, and criticisms. Therapists teach parents the use of child-centered skills and differential attention with the goal of improving the parent– child relationship and beginning to reduce children’s disruptive behaviors. In the second phase of PCIT, Parent-Directed Interaction (PDI), parents are taught how to use effective commands with consistent follow- through, including contingent praise for child compliance, a warn- ing for noncompliance, and the use of an effective, developmen- tally appropriate time-out procedure for noncompliance.

PCIT differs from other parent training programs in at least three critical ways that may make a group adaptation particularly valuable: First, the intervention gives equal focus to the promotion of the parent– child relationship and the development of parents’ behavior management skills (Eyberg & Funderburk, 2011). Be- cause parent– child interactions in families with children exhibiting conduct problems are frequently negative and coercive in nature (e.g.,Patterson, 1982;Stormont, 2002), a critical goal of PCIT is to increase positive, nurturing interactions. Second, in contrast to the traditional approaches to PMT (e.g., didactic and role play), parents rehearse new skills weekly in session through live inter- actions with their children. This provides the opportunity for direct coaching by the therapist, which is when the therapist gives im- mediate feedback on the parent’s skill development (e.g., from an observation room with a one-way mirror, while parents wear a radio-frequency earphone). Therapists use behavioral principles such as modeling, reinforcement, and differential attention in their coaching to shape parents’ behaviors as they occur (Barnett, Niec, & Acevedo-Polakovich, 2014). Immediate feedback using behav- ioral principles has been shown to increase positive parenting skills (Shanley & Niec, 2010). Further, in a meta-analysis of the components of parent management training associated with posi- tive changes in both children’s and parents’ behaviors, programs that included coaching of parent– child interactions had greater effect sizes than programs without this component (Kaminski et al., 2008). The active practice also allows the therapist to conduct ongoing behavioral assessments of each parent’s progress, which is the third way in which PCIT differs from most PMT models. In session, standardized coding of the parent– child interaction allows therapists to identify weaker parenting skills that need to be targeted in treatment and to recognize when a parent has reached mastery of the skills.

The robust outcomes of individual PCIT provide one rationale for investigating the adaptation of the intervention into a group format (e.g.,Chaffin, Funderburk, Bard, Valle, & Gurwitch, 2011; Nixon, Sweeney, Erickson, & Touyz, 2004;Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998). Families who complete PCIT demonstrate statistically and clinically significant reductions in children’s conduct problems and parents’ stress, as well as signif- icant increases in positive parenting behaviors and children’s com- pliance (Nixon et al., 2004;Schuhmann et al., 1998). PCIT has demonstrated efficacy in reducing the conduct problems of chil- dren from diverse ethnic backgrounds (McCabe & Yeh, 2009) and children with cognitive disabilities (Bagner & Eyberg, 2007).

Further, positive changes during treatment generalize from the clinic to school settings (McNeil, Eyberg, Eisenstadt, Newcomb, & Funderburk, 1991) and from the targeted child to untreated This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 683 GROUP PARENT–CHILD INTERACTION THERAPY siblings (Brestan, Eyberg, Boggs, & Algina, 1997). Maintenance of treatment gains is good, with families who complete PCIT showing positive gains maintained as long as 6 years posttreatment (Hood & Eyberg, 2003).

We developed a Group PCIT model that would maintain the core features of standard PCIT while adding the potential benefits of group treatment. As in the individual format, Group PCIT has a dual focus on parent– child relationship enhancement (CDI phase) and parent behavior management skills (PDI phase); children participate in all phases of the treatment; parents practice skills during live parent– child interactions; and therapists provide im- mediate feedback to parents. In addition to these core features, Group PCIT provides parents with opportunities to develop rela- tionships with one another. For example, parents actively observe and code one another during the live coaching of their interactions with their children. After each parent has been coded and coached, families provide one another with constructive feedback. This strategy has multiple benefits: it provides opportunities for vicar- ious learning; it allows parents to receive constructive feedback from their peers; and it provides opportunities for parents to support one another, fostering group cohesion.

Group PCIT may offer a valuable means of reaching more families in need of services. A small-scale (N 27) prepost evaluation of Group PCIT in a community setting found that parents who completed PCIT reported reductions in their chil- dren’s conduct problems and demonstrated gains in positive par- enting skills (Nieter, Thornberry, & Brestan-Knight, 2013). Al- though the study required only that parents express a need for parenting assistance, rather than children to meet criteria for a disruptive behavior disorder, the study supports the feasibility of conducting PCIT in a group format.

In the present study, we used a randomized control trial to investigate the outcomes for families with children diagnosed with oppositional defiant disorder or conduct disorder who completed either group or individual PCIT. We also evaluated the acceptabil- ity of both treatment formats, as acceptability of PMT models can predict adherence to the program (Reimers, Wacker, Cooper, & de Raad, 1992) and improvements in children’s behaviors (MacKen- zie, Fite, & Bates, 2004). To control for treatment dosage, families in the individual and Group PCIT conditions received the same number of PCIT sessions. PCIT was developed to be a mastery- based intervention, in that parents progress from the first phase of treatment (CDI) to the second phase (PDI) when they reach a specified level of competence with the child-centered parenting skills. Families successfully graduate from the program after mas- tery of the skills in both phases of treatment and their children’s conduct problems are within normal limits. However, session- limited PCIT has been identified as having similar long-term outcomes as standard PCIT (Nixon, Sweeney, Erickson, & Touyz, 2004), and time-limited PCIT has been evaluated in previous RCTs (e.g.,Chaffin et al., 2004,2011).

We expected that families in both treatments would demonstrate significant reductions in children’s conduct problems and in- creases in positive parenting, and that Group PCIT would not be inferior to individual PCIT in either outcome. However, we ex- pected that parents who had the benefit of peer support through Group PCIT would report experiencing more social support, dem- onstrate greater treatment adherence, and manifest better retention than families in individual PCIT. Although we considered includ-ing a no-treatment control condition to investigate the efficacy of Group PCIT, a review of the research revealed (a) consistency in the positive outcomes for individual PCIT including session- limited PCIT (i.e., reduction of childhood conduct problems, in- creases in positive parent– child interactions, reductions in parent stress) and (b) robust effects for parent behavior training in the treatment of conduct problems (e.g.,Brestan & Eyberg, 1998; Chambless & Ollendick, 2001;Eyberg et al., 2008). Thus, we determined that sufficient evidence existed to suggest that Group PCIT would be more efficacious than no treatment, and we con- ducted a more rigorous test of efficacy by comparing Group PCIT to an individual PCIT condition. Method Participants Families seeking services at a university mental health clinic were referred to receive information about the study when the primary complaint was conduct problems in their 3- to 6-year-old children. One hundred-seven families were referred and scheduled to meet with a study clinician to receive a thorough explanation of the project (seeFigure 1). Of the 107 referred families, 10 did not follow through with the scheduled appointment and three families declined to participate. Ninety-four families provided written in- formed consent, as approved by the Central Michigan University human subjects review board, and completed measures to deter- mine study eligibility. To be included in the study, children were required to meet diagnostic criteria for aDiagnostic and Statistical Manual for Mental Disorders-Fourth Edition(DSM–IV) diagnosis of oppositional defiant disorder (ODD) or conduct disorder (CD; American Psychiatric Association, 2000) and to have conduct problems rated by a caregiver in the clinical range of severity (i.e., BASC-2 Externalizing Composite score of T 70 or ECBI Intensity score of T 60). Five families did not meet eligibility criteria and an additional eight families did not follow through to complete the initial assessment. Of the 81 children who were allocated to a treatment condition, 51 met criteria for ODD and 30 for CD. At least one caregiver was required to participate in treatment. Although participating caregivers included biological parents (78.8%) and other caregivers (e.g., grandparents, steppar- ents, 21.2%), for efficiency of communication, we refer here to male caregivers as “fathers” and female caregivers as “mothers.” Parents and children were excluded if they fell below a standard score of 70 on a cognitive screening measure and if there was a positive history of severe sensory or mental impairment (e.g., severe hearing impairments, pervasive developmental disorder).

Families with active involvement in the child protective services system were also excluded and offered services outside of the study, as families in which abuse has occurred present different treatment issues and may require somewhat different interventions than those in which the primary problem is childhood conduct problems (Chaffin et al., 2004,2011). Children were not excluded for concurrent treatment with psychoactive medication. Families whose children were taking psychotropic medication (n 9) were evaluated only after the children’s behaviors had stabilized on the medication. Stabilization was defined as (a) parent satisfaction with the current dosage, (b) consistent dosage for at least one This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 684 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM month, and (c) physicians’ verbalizations that further titration was not anticipated.

Procedure Following the recommendations ofKernan, Viscoli, Makuch, Brass, and Horwitz (1999), randomization was stratified by med- ication status into two strata: (a) children who were stabilized on medication and (b) children who were not on medication. That is, randomization was generated independently for families within each strata. Upon completion of the initial assessment, families were randomly assigned to one of two treatment conditions: (a) individual PCIT or (b) Group PCIT.

Treatment conditions.PCIT was originally developed to be conducted by co-therapy teams to individual families (McNeil & Hembree-Kigin, 2010). The therapists in this project were ad-vanced doctoral students in clinical psychology who had com- pleted the PCIT training workshop conducted by the first author (a clinical child psychologist with extensive expertise in PCIT and vetted by the developer of PCIT, Sheila Eyberg, PhD), observed a PCIT case, and participated in weekly PCIT supervision. In addi- tion, all lead therapists had experience treating PCIT cases for at least 1 year. All 13 therapists and 6 evaluators attended workshops in administering and scoring structured interviews and behavior observations. The same therapists ran the PCIT groups and indi- vidual cases to maintain equivalence of therapists across treatment conditions. Because the treatment conditions only differed in their format (group vs. individual), not their therapeutic protocols, con- tamination was not an issue. Core components of standard PCIT were maintained across both conditions; however, to control for dosage across conditions, the number of sessions was held con- Assessed for eligibility (n = 94) Not meeting inclusion criteria (n = 5) 2 = behaviors subclinical 2 = parent low cog funct 1 = CPS involvement Analyzed (n = 42) Completed intervention = 28 Discontinued intervention (n = 10) Lost to 6-mnth post-tx follw-up (n = 15) Allocated to Individual PCIT (n = 42) Received some intervention (n = 38) Received no intervention (n = 4) 1 = unstable medication, withdrew 2 = family moved 1 = child chan ged care givers Completed intervention = 29 Discontinued intervention (n = 10) Lost to 6-mnth post-tx follw-up (n = 12) Allocated to Group PCIT (n = 39) Received some intervention (n = 39) Received no intervention (n = 0 ) Analyzed (n = 39) Allocation Analysis Follow-Up Enrolled (n = 89) Randomization (n = 81) Referred for assessment (N = 107) Dropped prior to randomization (n = 8) 10 = failed to follow-through with assessment appointment.

3 = declined to participate Figure 1.Participant flow diagram. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 685 GROUP PARENT–CHILD INTERACTION THERAPY stant.Table 1summarizes the similarities and differences of the individual PCIT condition, Group PCIT condition, and standard PCIT protocol (Eyberg & Funderburk, 2011).

Individual PCIT.Individual PCIT sessions were held once a week for approximately 1 hr. The principles and skills of respon- sive and consistent parenting were presented during the CDI Teach Session using didactic, modeling, and role play, followed by four coaching sessions in which the therapists provided in vivo feed- back to parents who were actively practicing the skills with their children. Parents then learned about the use of effective commands and discipline (PDI Teach Session) and completed seven coaching sessions focused on these skills. Parents were asked to practice the skills at home during daily play sessions with their children (5 min each day during the child-directed phase of treatment; 10 –15 min a day during the parent-directed phase, which included 5 min of child-directed play and time to practice effective commands with the appropriate follow-through). For more information about the PCIT protocol seeEyberg & Funderburk, 2011).

Group PCIT.The aim of the group protocol was to maintain the unique aspects of PCIT, while including strategies to develop the potential benefits of group parent training. As in individual PCIT, Group PCIT included the same didactic and coaching ses- sions (Niec et al., 2005). Also similar to individual PCIT, two therapists led the groups. However, PCIT with multiple families required minor modifications that distinguished it from the original PCIT protocol and from other group models. For instance, suffi- cient time was required to coach each parent– child dyad. For this reason, group sessions were 2 hr long and groups contained three to seven parent– child dyads (i.e., two to five families with one to four caregivers in each family).

Initial group sessions focused on maximizing group cohesion by encouraging rapport among parents, establishing group guidelines, and setting a collaborative tone conducive to therapeutic gain.

Parents were encouraged to identify constructive similarities withother parents (e.g., that they were all attending treatment to learn new ways of managing their children’s behaviors). Therapists were trained to identify, prevent, and correct counterproductive discus- sion using a variety of therapeutic techniques (e.g., reframing, redirection, and differential attention). All group process concerns were discussed in weekly supervision.

Thirteen of the 14 sessions in the Group PCIT model (93%) were conducted with all the members of the group present. Only during one of the 14 sessions did parents meet with therapists individually: that was the coaching session in which parents im- plemented the new discipline procedure for the first time (PDI Coach 1). This PCIT session is sometimes longer because parents are just learning the procedure and children are learning and testing their parents’ new responses to noncompliance. As has been described elsewhere (Niec et al., 2005), the individual session provides parents and children the time and attention they need to negotiate the discipline process for the first time. However, as previously implemented within a small-scale community sample (Nieter et al., 2013), all the subsequent PDI coach sessions (six out of seven) were conducted in a group format, which provided families with the opportunity to (a) learn and become comfortable with the discipline process by observing other parents successfully completing it, and (b) support one another during difficult disci- pline scenarios. For example, parents observing other families during a time-out sequence frequently gave encouraging state- ments that the therapist communicated via the bug-in-the-ear (e.g., “You’re doing such an amazing job staying calm!”; “Wow! Stick with it!”).

Because make-up sessions were offered to families within the same week as the missed session, a few families received an additional individual session. Six of the 39 families allocated to Group PCIT received a total of nine make-up sessions. No family received more than two make-up sessions. No group sessions were held that did not include at least two families, although one group Table 1 Comparison of PCIT Protocols Protocol components Current PCIT protocol Individual PCIT Group PCIT Skills targeted in CDI Child-centered skills and differential attention; decrease leading and directive parent behaviorSame as current Same as current Skills targeted in PDI Effective commands; appropriate follow-through; timeout sequenceSame as current Same as current Orientation session None Overview of PCIT Overview; promote group cohesion Number of sessions Unlimited 14 orientation; CDI didactic and 4 coaching sessions; PDI didactic and 7 coaching sessions14 orientation; CDI didactic and 4 coaching sessions; PDI didactic and 7 coaching sessions Termination criteria Parent meets CDI and PDI mastery criteria child behavior WNLNone None Number of Families 1 1 3–7 dyads In-session behavior assessment (DPICS) Yes Yes Yes Daily homework Yes Yes Yes PDI coaching #1 Meet individually Meet individually Meet individually Group parents observe each other during P-C interactions No No Yes Group parents receive feedback from other parents No No Yes Note.FromEyberg and Funderburk (2011). This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 686 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM was discontinued when all three families ended treatment within the same two sessions. The mean number of families enrolled in each of the 11 groups was 3.5 and the mean number completing treatment in each group was 2.6 (74%; range 0 – 4 families).

As in individual PCIT, the final group session included specific discussion of what families could do to maintain treatment gains and how to deal with setbacks or new problems that might arise in the future. Specific to the group condition, families were encour- aged to seek positive support from one another.

Treatment Fidelity A number of procedures were implemented to help maintain treatment fidelity throughout the study. First, all assessment and treatment clinicians completed comprehensive training (described above). To increase fidelity during sessions, both the group and individual PCIT treatment manuals included outlines of the pri- mary components to be addressed during each session, which the therapists checked as they completed. All study clinicians received weekly supervision on every case. In addition, there was frequent live observation of sessions by the supervisor. Questions regarding protocol implementation that could not be clearly resolved with the PCIT manual were discussed with Dr. Sheila Eyberg.

Finally, all treatment sessions were recorded to assess fidelity upon completion. Evaluation of fidelity was rigorous, using ad- vanced PCIT therapists, who were not therapists in the study, to review a random selection of the recorded sessions from both treatment conditions. In total, 435 session components (e.g., re- viewing parents’ homework, defining skills, correctly implement- ing the precoaching behavior observation, providing adequate coaching time) were reviewed for fidelity in a total of 70 sessions.

Ratings demonstrated high treatment fidelity (88% for group and individual treatment formats).

Measures Wonderlic Personnel Test (WPT).The WPT is a 50-item test designed as a screening scale of adult intelligence (Dodrill, 1981). In a sample of 120 normal adults, the WPT estimate of intelligence correlated 0.93 with the WAIS Full Scale IQ score, and the WPT score was within 10 points of the WAIS IQ score for 90% of the subjects (Dodrill, 1981;Dodrill & Warner, 1988). We used the WPT Timed Score as a cognitive screening measure for parents.

Peabody Picture Vocabulary Test-Third Edition (PPVT-III).

The PPVT-III is a standardized test that measures receptive lan- guage in individuals aged 2.6 years through adulthood (Dunn & Dunn, 1997). The PPVT-III correlates 0.90 with the Wechsler Intelligence Test for Children—Third Edition Full Scale IQ (Dunn & Dunn, 1997), and was used in this study as a cognitive screening measure for child participants.

Diagnostic Interview Schedule for Children (NIMH DISC IV-P).The NIMH-DISC-IV-P is a highly structured diagnostic interview for administration to parents (Shaffer, Fisher, Lucas, Dulcan, & Schwab-Stone, 2000). It includes all common mental disorders of children included in theDSM–IVthat are not depen- dent on specialized test procedures. One-week test–retest reliabil- ity on administration to parents of 9- to 17-year-old children has been reported at 0.54 for ODD and 0.54 for CD (Grills & Ollen-dick, 2002). We consulted with Dr. Sheila Eyberg (trained by C.

Lucas at Columbia University) on the modifications to certain questions needed for developmentally appropriate administration to parents of children as young as 3 years. We used the ODD and CD modules of the DISC to assess whether children met criteria for study inclusion.

Behavioral Assessment System for Children—2; Parent Rat- ing Scale 2–5 year olds and Parent Rating Scale 5–11 year olds (BASC-2).The BASC-2 is a broad-band rating scale developed to assess the behaviors of children 2 through 18 years of age (Reynolds & Kamphaus, 2004). The measure yields a variety of scales including scales of Externalizing Problems, Internalizing Problems, a Behavior Symptoms Index, and an Adaptive Skills Scale. Reliability and validity of the scale have been found to be good. The BASC has shown sensitivity in discriminating various groups of clinic-referred children including children with conduct problems (Whitcomb & Merrell, 2013). Test–retest reliability is good (0.81– 0.92 for the Behavior Symptoms Index for children 2–11 years), suggesting that without intervention, scores remain stable (Reynolds & Kamphaus, 2004). The BASC Externalizing Problems Composite Score was used as a criterion for study inclusion. The Externalizing, Internalizing, and Adaptive Skills Composite Scores provided measures of treatment outcome.

Eyberg Child Behavior Inventory (ECBI).The ECBI is a 36-item inventory developed to measure parents’ perceptions of the conduct problems of children 2 through 16 (Eyberg & Pincus, 1999). The Intensity Scale (IS) measures the frequency of conduct problems on a 7-point scale from 1 (never)to7(always). The IS has yielded an internal consistency coefficient of 0.95; interrater (mother–father) reliability coefficients of 0.69 (Eyberg & Pincus, 1999), and without treatment, has shown good long-term (10 month) test–retest reliability (Funderburk, Eyberg, Rich, & Behar, 2003). The ECBI has demonstrated sensitivity to treatment effects (Eisenstadt et al., 1993;Schuhmann et al., 1998;Webster-Stratton & Hammond, 1997) and has good sensitivity and specificity re- lated to the identification of children with oppositional defiant and conduct disorder (Rich & Eyberg, 2001). We used the ECBI IS as a criterion for study inclusion and as a primary measure of treat- ment outcome.

Dyadic Parent–Child Interaction Coding System-III (DPICS-III).The DPICS-III is a behavioral observation mea- sure developed to assess the quality of parent– child interactions in a standardized format (Eyberg et al., 2010;Robinson & Eyberg, 1981). Parent– child dyads referred for the treatment of childhood conduct problems differ from nonreferred parent– child dyads on a number of DPICS variables (Forster, Eyberg, & Burns, 1990; Webster-Stratton, 1985). The DPICS also demonstrates sensitivity to treatment effects (Schuhmann et al., 1998). In the present study, during treatment sessions in which parents were coached (i.e., weeks 2–5, 7–13), parents were first observed in the Child Led Play (CLP) situation of the DPICS. The CLP situation is meant to assess parents’ acquisition of responsive parenting skills. Frequen- cies of parent “Do Skills” (e.g., praises, behavior descriptions, and reflections) and “Don’t Behaviors” (questions, commands, and negative talk) were coded from the video-recorded measure to examine skill acquisition across treatment conditions over time.

To assess interrater reliability in the current sample, 825 five minute video-recorded segments of the DPICS were coded by a primary coder blind to the study hypotheses. Before coding for this This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 687 GROUP PARENT–CHILD INTERACTION THERAPY study, the primary coder was trained intensively in the DPICS-III coding system, had met criteria ( 0.80 for all categories) with an expert-rated tape, and has been coding DPICS for 4 years when this project began. Interrater coders at an outside institution inde- pendently coded 211 (25%) randomly selected segments. Interrater coders were blind to study hypotheses, participants’ treatment condition, and phase of treatment. Interclass correlation coeffi- cients were calculated on the child-centered interaction skills (“Do Skills”), as well as those behaviors targeted to be reduced (“Don’t Behaviors”;Table 2). Of the seven codes, all demonstrated at least adequate reliability (r .65) and five demonstrated good to excellent reliability (r .80).

Treatment adherence.We measured three aspects of partic- ipants’ treatment adherence: (a) session attendance, (b) self- reported completion of weekly homework assignments, and (c) therapist-rated in-session participation.

Number of sessions attended.Therapists maintained atten- dance records for each participant. Therapists were available for families to schedule make-up sessions within 1 week of a missed appointment.

Homework completion.The number of homework assign- ments that caregivers completed was also used to assess treatment adherence. Following the initial CDI coaching session, caregivers were asked to complete a daily five-minute rehearsal of skills with their children at home for the 6 days of the week they did not attend therapy sessions. During PDI coaching sessions, practice of the discipline procedure was added to the rehearsal of child- centered skills and later generalized to other times of the day.

Caregivers recorded their homework completion including the dates, the type of practice, and any problems they encountered.

Therapist-rated participation.Immediately subsequent to each session, both therapists rated each caregiver’s participation. Level of participation was rated for each caregiver on a 3-point scale (1 seldom, 2 sometimes, or 3 frequent), and ratings were averaged across therapists. A similar scale has been used to assess treatment adherence in other parenting programs (e.g.,Webster- Stratton, 1990). Because the current study used the mean partici- pation scores across raters, reliability for this measure was com- puted using the average-measures intraclass correlation coefficient (ICC[2]), which indicates the reliability of a mean score. The ICC(2) for parent participation was 0.64, indicating acceptable reliability.

Multidimensional Scale of Perceived Social Support-Child Behavior (MSPSS-C).The MSPSS-C is a 16-item self-report questionnaire adapted from the original MSPSS (Niec, Cochran, & Stayer, 2003;Zimet et al., 1988) to assess parents’ perceived social support from family, friends, significant others, and other parents,specifically in relation to their children with conduct problems.

Items are scored on a seven-point scale that ranges from “very strongly disagree”to“very strongly agree.” Similar to the original scale, internal consistency (Cronbach’s ) for the Total Score was high (0.90). Concurrent validity has been supported by demon- strating associations between the MSPSS-C Total Score and mea- sures of parent stress, parent psychopathology, children’s behavior problems, and parent– child attachment in a sample of families referred for treatment of child behavior problems (Niec et al., 2003).

Parenting Stress Index-Short Form (PSI-SF).The PSI-SF is a 36-item parent self-report instrument designed to measure the relative degree of stress in a parent– child relationship and to identify the sources of distress (Abidin, 1995). The Total Stress score of the PSI-SF has correlated 0.94 with the Total Stress score of the full PSI. Test–retest reliability was 0.84 over a 6-month interval. On the long-form PSI, it has been found that higher scores are associated with increased severity of conduct-disordered be- havior (Eyberg, Boggs, & Rodriguez, 1992;Ross, Blanc, McNeil, Eyberg, & Hembree-Kigin, 1998). PSI scores are sensitive to treatment changes with young children (Eisenstadt et al., 1993). In this study, the PSI-SF Total Score was a measure of treatment outcome and demonstrated good internal consistency ( .92).

Therapy Attitude Inventory (TAI).The TAI was designed to assess parental satisfaction with the process and outcome of therapy (Eyberg, 1993). It consists of 10 multiple-choice questions addressing the impact of parent training on areas such as confi- dence in discipline skills, quality of the parent– child interaction, the child’s behavior, and overall family adjustment. Internal con- sistency (Cronbach’s ) of the TAI was 0.92 in the current study.

Previous research has demonstrated discriminative validity be- tween outcomes of alternative treatments (Eisenstadt et al., 1993; Eyberg & Matarazzo, 1980). We used the TAI total score to compare parent satisfaction at posttreatment between group and individual conditions. Results Randomization Check and Assessment of Nesting We conducted conservative intention-to-treat analyses, main- taining all participants within the dataset after their randomized assignment to treatment condition and imputing missing outcome data forward based on the last observation (Higgins & Green, 2011). Hypothesis testing was conducted with significance set atp values of less than 0.05. Before testing the primary hypotheses, treatment groups were compared for equivalence on all major demographic variables. Chi-square analyses revealed no signifi- cant differences in the distribution of children’s medication status, gender, ethnicity, and race or caregivers’ ethnicity and race across groups. Similarly, no significant differences were found on chil- dren’s ages and receptive vocabulary scores or caregivers’ ages, cognitive functioning, and education levels (seeTable 3).

The data collected from families who participated together within a group cannot be assumed to be independent; they present a level of nesting that is often neglected in analyses of the efficacy of group interventions. Therefore, before testing the primary hy- potheses, we used Hierarchical Linear Modeling (HLM) analyses to estimate the degree of nesting exhibited by study variables Table 2 Interrater Reliability of the Weekly DPICS-III Parent Variables “Do Skills” “Don’t Behaviors” n 211 UP LP RF BD QU CO NTA ICC .76 .80 .81 .84 .91 .65 .67 Note.UP unlabeled praise; LP labeled praise; RF reflection; BD behavior descriptions; QU auestions; CO commands; NTA negative talk (criticisms). This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 688 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM within specific groups in the Group PCIT condition (HLM 7 software;Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2011).

This nesting was examined with a three—level HLM model using only the participants in Group PCIT: the time of observation was treated as the first level of analysis (e.g., intake, posttreatment, or follow-up), the caregiver was treated as the second level, and the specific therapy group was treated as the third level. These anal- yses were conducted separately for mothers and fathers. Signifi- cant variation between Level 3 units (therapy groups) was used to determine the effect of group membership while controlling for other sources of variation. As an exception, the TAI (that was only collected at posttreatment) was tested using a two-level model, with variance between therapy groups reflecting the variance be- tween Level 2 units.

Using Wald 2tests of significance, most variables did not display a significant amount of variation between therapy groups, with a few exceptions explained next. Results suggested that therapy group membership accounted for 23% of the variance in mother TAI scores ( 2 16.92,p .050) and 38% of the variance in father TAI scores ( 2 13.27,p .039). This effect was not surprising, as one’s specific therapy group is likely to influence one’s satisfaction with the therapy overall. Group membership also accounted for 33% of the variance for father ratings on the Adapt- ability Scale of the BASC-2 ( 2 24.36,p .004). For mothers, therapy group membership accounted for significant variance on the DPICS-III behavior observation measure: 25% of the variance for the parenting “Do Skills” ( 2 35.24,p .001) and 30% of the variance for the parenting “Don’t Behaviors” ( 2 22.99,p .001). Statistical tests that include a nested variable tend to under- estimate the standard errors and increase the likelihood of Type I error (Hox, 2010). Although these few outcome variables exhib- ited a nesting effect from specific therapy groups, the resulting bias in our statistical tests is muted by the fact that the nesting only exists in half of the sample and frequently for only one set of caregivers. Regardless, significant effects regarding nested vari- ables warrant some caution in interpretation. Child Behavioral Functioning We examined changes in children’s behavioral and emotional functioning across time (intake, posttreatment, or 6-month follow- up) and treatment condition (group or individual PCIT) using repeated measures multivariate analysis of variances (MANO- VAS). The ECBI Intensity Scale and the BASC-2 Externalizing Problems Composite, Internalizing Behaviors Composite, and Adaptive Skills Composite were compared for mothers’ and fa- thers’ reports separately (seeTable 4for descriptive statistics and Table 5for MANOVA statistics). Parent reports of child function- ing demonstrated a significant main effect for time (mothers:F(8, 70) 16.78,p .000; Pillai’s Trace 0.66, p2 .66; fathers:

F(8, 33) 4.99,p .000; Pillai’s Trace 0.55, p2 .55), but not for treatment (mothers:F(4, 74) .86,p .495; Pillai’s Trace 0.04, p2 .04; fathers:F(4, 37) 1.62,p .190; Pillai’s Trace 0.15, p2 .15). Across time, children not only experi- enced a reduction of conduct problems, but also a reduction of internalizing symptoms and an increase in adaptive behaviors. The results of planned comparisons with a Least Significant Differ- ences test indicated that fathers and mothers rated their children’s externalizing and internalizing symptoms on the BASC-2 as sig- nificantly less severe from intake to posttreatment and from intake to follow-up. Fathers and mothers also rated their children as having significantly more adaptive behaviors from intake to post- treatment, and intake to follow-up. Neither mothers nor fathers had a significant interaction effect between time and treatment condi- tion. That is, for both mothers and fathers, children’s externalizing behaviors and internalizing symptoms decreased over time and adaptive skills increased regardless of treatment format.

On the ECBI Intensity Scale, a narrowband measure of the frequency of children’s conduct problems, mothers’ reports showed a significant main effect for time, but not for treatment condition. Fathers’ reports on the ECBI Intensity Scale demon- strated significant differences across time and treatment condition, with fathers in the Group PCIT condition rating their children’s Table 3 Demographics of the Group and Individual PCIT Conditions DemographicGroup Individual %orMnorSD%orMnorSD tor 2 Child age 4.12 1.01 4.33 1.35t(78) .82 Child PPVT-III 98.21 13.38 96.72 15.78t(78) .45 Child gender 2(1) .65 Boys 74.4% 29 69% 29 Girls 25.6% 10 31% 13 Child ethnicity (Hispanic) 8.57% 3 0% 0 2(1) 3.31 Child race 2(2) 2.84 White 84.6% 33 85.7% 36 Native American 5.1% 2 0% 0 Multi-racial 7.7% 3 11.9% 5 Not reported 2.6% 1 2.4% 1 Child medication status (yes) 7.89% 3 14.63% 6 2(1) .89 Caregiver cog functioning 99.46 15.29 99.22 15.19t(122) .09 Primary caregiver age 32.57 8.69 31.07 7.01t(78) .85 Other caregiver age 36.42 10.96 33.97 10.07t(77) .93 Primary caregiver education 13.99 2.70 13.63 2.34t(77) .62 Other caregiver education 14.06 2.47 13.09 2.63t(77) 1.54 Note.All means are reported in years, except the child PPVT-III, which is the standard score of the Peabody Picture Vocabulary Test-III, and the caregiver cog functioning, which is the timed score of the Wonderlic Personnel Test (transformed onto the standard IQ scale). This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 689 GROUP PARENT–CHILD INTERACTION THERAPY conduct problems as less severe than those in the individual PCIT condition. Pairwise comparisons revealed that according to both mothers and fathers, children’s conduct problems decreased sig- nificantly from intake to posttreatment and intake to follow-up.

Neither mothers’ nor fathers’ reports of children’s conduct prob- lems on the ECBI Intensity Scale demonstrated significant inter- actions between time and treatment condition.

A primary goal of the study was to test the efficacy of Group PCIT relative to individual PCIT; thus, we conducted analyses to determine whether the group condition yielded noninferior post- treatment scores relative to the individual condition. Noninferiority tests were conducted using a one-sided equivalencettest, as outlined byWalker and Nowacki (2010). Although a standard equivalence range ( ) has not been established in prior PCIT research, research in other domains has generally used a set per- centage of the control group mean to determine . The current study used 10% of the individual PCIT condition mean as the acceptable range of noninferiority, which was the same threshold used between control and treatment groups for the Minnesota Multiphasic Personality Inventory (MMPI) byRogers, Howard, and Vessey (1993). The results for the noninferiority tests are noted inTable 4. The Group PCIT condition was not inferior to the individual PCIT condition for 8 out of the 10 comparisons, includ- ing the primary outcome variable, child conduct problems, as rated by both mothers and fathers on the BASC Externalizing ProblemsComposite and the ECBI Intensity Scale. Mothers’ ratings of children’s internalizing problems and parent stress served as the two exceptions to the findings of noninferiority. Parent Functioning Parent stress.We examined changes in mothers’ and fathers’ stress across time and treatment condition using repeated measures ANOVAS (Tables 4and6). As expected, mothers and fathers expe- rienced a significant main effect in their parenting stress levels for time but not treatment condition. Pairwise comparisons indicated that mothers and fathers demonstrated a significant decrease in parenting stress from intake to posttreatment, and intake to follow-up. Neither mothers nor fathers had a significant interaction effect across time and treatment condition.

Parenting skills acquisition.Analyses on the development of positive parenting skills over time were performed using HLM with restricted maximum likelihood (RML) estimation, which improves estimation accuracy for studies with smaller Level 2 sample sizes (n 100). The significance of the fixed regression coefficients ( ) were tested as means upon the t-distribution, and the significance of the random coefficients (variance components) were tested using the Wald’s 2test (Raudenbush & Bryk, 2002).

The effect size for the growth function was assessed by standard- izing the regression coefficients as . A two-level HLM model was Table 4 Parent Report of Child and Parent Functioning Measure of FunctioningMothers Fathers Group PCIT n 39Individual PCIT n 42Group PCIT n 22Individual PCIT n 24 M SD M SD M SD M SD BASC-2 externalizing composite Intake 72.84 13.34 74.81 12.15 65.50 10.13 69.58 11.10 Posttreatment 65.69 a 13.76 68.33 14.51 58.09 a 9.48 65.42 11.15 Follow-up 62.44 12.85 67.36 14.14 59.73 9.43 64.13 12.10 BASC-2 internalizing composite Intake 64.13 16.58 60.71 12.81 56.00 9.02 57.92 11.75 Posttreatment 58.23 17.28 57.00 13.37 53.50 a 9.30 55.42 11.58 Follow-up 54.10 15.62 55.88 13.87 52.55 9.52 53.96 11.69 BASC-2 adaptability scale Intake 41.82 10.35 39.36 8.79 42.50 7.18 37.33 8.83 Posttreatment 42.64 a 9.64 42.12 9.09 44.82 a 8.27 39.92 9.43 Follow-up 42.77 10.41 42.21 7.93 44.59 8.86 41.33 10.42 ECBI intensity scale Intake 163.42 24.84 170.74 27.23 147.53 28.07 166.58 22.98 Posttreatment 129.03 a 40.00 134.55 41.93 112.71 a 35.72 138.58 36.26 Follow-up 123.90 38.12 137.36 36.68 116.19 35.01 135.52 36.19 PSI-SF total stress score Intake 106.74 19.48 100.50 17.94 92.63 11.34 93.15 15.48 Posttreatment 97.89 19.45 90.40 17.34 84.83 a 14.82 90.23 16.06 Follow-up 93.69 21.92 90.00 17.61 86.87 15.13 85.55 19.92 MSPSS-C Intake 79.80 16.03 79.74 17.64 72.95 17.62 78.25 13.62 Posttreatment 82.13 18.08 82.18 17.95 80.14 22.61 81.79 14.90 Follow-up 81.15 17.33 83.10 17.46 79.55 21.82 82.17 15.05 Note.Ns for analyses with mother scores ranged from 73 (PSI) to 80 (BASC scales).Ns for analyses with fathers ranged from 43 (ECBI and PSI) to 44 (BASC scales) BASC-2 Behavioral Assessment System for Children-2; ECBI Eyberg Child Behavior Inventory; BASC-2 composites are T-scores; all other scores are raw scores. MSPSS-C Multidimensional Scale of Perceived Social Support-Child Behavior. aOne-sidedt-test indicated statistically noninferior results in the Group PCIT compared to the individual PCIT ( 10% of the individual PCIT condition mean). This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 690 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM tested for fathers and mothers separately for each set of primary outcomes (“Do Skills” and “Don’t Behaviors”). Analyses for mothers included 544 observations nested within 65 caregivers, whereas analyses for fathers included 247 observations nested within 35 caregivers (seeTable 7).

The time variable was centered such that “0” reflected the behavioral scores for the first week, with each unit reflecting one week’s time and increasing up to 11. Examination of scatter and mean plots revealed a pattern of change in which scores tended to increase or decrease (depending on the behavioral outcome) for the first few weeks and then level off, indicating parents’ tendency to gain skills rapidly early in treatment. To statistically model the growth patterns illustrated in the scatterplot, the time variable was transformed by taking its square root. This transformation would predict notable gains in parenting skills for the first weeks and smaller gains in the last few weeks. Consistent with the nonlinear patterns observed in the scatterplot, preliminary analyses indicated that the transformed time values (square root function) outper- formed the untransformed values (linear function) in predicting skill acquisition scores over the course of treatment. Thus, all skillacquisition models used a square root function to examine the growth in parental skills over time. Patterns of change for both the “Do Skills” and “Don’t Behaviors” for mothers and fathers were consistent with predictions. That is, parents’ child-centered skills significantly increased and their negative behaviors significantly decreased over time (seeTables 8and9for descriptive statistics).

To test the interaction between time and treatment format, variance in the growth functions between caregivers was tested in the models for changes over time to determine if this variance should be modeled in subsequent analyses. Results indicated sta- tistically significant variation in the growth function for all behav- ioral outcomes and samples. As a result, analyses with these outcomes included an additional random coefficient to reflect the variance in growth functions between caregivers ( 11): that is, differences in the rate of change between caregivers. Estimates of the caregiver variance between intercepts ( 00) along with esti- mates of growth functions between caregivers ( 11), were used for comparison with the model that includes the primary predictor— that is, PCIT format—to determine if the relative variance was accounted for by treatment condition. The analysis of the impact of Table 5 Repeated-Measures Multivariate Analysis of Variance of Parent Report of Child Behavioral Functioning Measure of Child FunctioningMothers Fathers df F p 2 pdf F p 2 p BASC-2 externalizing composite Condition 1,77 1.02 .01 .315 1,40 3.44 .08 .071 Time 1.64, 126.10 39.07 .34 .000 1.54, 61.75 20.68 .34 .000 Time Condition 1.64, 126.10 1.74 .02 .305 1.54, 61.75 1.39 .03 .255 BASC-2 internalizing composite Condition 1,77 .28 .00 .596 1,40 .56 .01 .458 Time 1.73, 133.51 26.41 .26 .000 1.35, 53.95 7.92 .17 .003 Time Condition 1.73, 133.51 3.08 .06 .056 1.35, 53.95 .03 .00 .921 BASC-2 adaptability scale Condition 1,77 .36 .01 .548 1,40 3.81 .09 .058 Time 1.89, 145.31 3.51 .04 .035 1.81, 72.51 5.21 .12 .010 Time Condition 1.89, 145.31 .44 .01 .632 1.81, 72.51 .49 .01 .596 ECBI intensity scale Condition 1,77 1.50 .02 .224 1,40 5.85 .13 .020 Time 1.86, 143.18 78.13 .50 .000 1.19, 47.73 33.96 .46 .000 Time Condition 1.64, 143.18 .66 .01 .509 1.19, 47.73 .47 .01 .532 Note.Greenhouse-Geisser correction for violations of sphericity. BASC-2 Behavioral Assessment System for Children-2; ECBI Eyberg Child Behavior Inventory. Table 6 Repeated Measures Analysis of Variance of Parent Reports of Stress and Social Support Measure of Parent FunctioningMothers Fathers df F p 2 pdf F p 2 p PSI-SF Condition 1,70 2.31 .03 .133 1,42 2.77 .06 .631 Time 1.63, 114.14 23.18 .25 .000 1.75, 71.74 12.09 .23 .000 Time Condition 1.63, 114.14 .49 .01 .573 1.75, 71.74 .23 .03 .268 MSPSS-C Condition 1,79 .27 .04 .851 1,42 .32 .01 .459 Time 1.46, 115.40 1.47 .02 .234 1.45, 60.95 9.63 .17 .001 Time Condition 1.46, 115.40 .25 .00 .709 1.45, 60.95 .47 .00 .565 Note.PSI-SF Parent Stress Index-Short Form; MSPSS-C Multidimensional Scale of Perceived Social Support-Child Behavior. Greenhouse-Geisser correction for violations of sphericity. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 691 GROUP PARENT–CHILD INTERACTION THERAPY PCIT format took the model for change over time and entered treatment condition at the second level. The model examined the ability of PCIT format to explain intercept differences among caregivers (reflecting a main effect) and to explain differences inthe growth functions among caregivers (reflecting an interaction between the predictor and the growth function). There were no main effects for treatment condition and no interaction effects between treatment condition and time (seeTable 7). That is to say, Table 7 HLM Results for Parent Skills Acquisition Model 1: “Do Skills” Model 2: “Avoid Skills” Effects for mothers (SE) (SE) Fixed effects Intercept ( 00)17.49 (1.07) — 18.37 (1.21) — Growth function ( 01)2.50 (.35).21 2.48 (.41) .27 PCIT Format a( 10) .03 (2.02) .00 .04 (1.81) .00 Interaction ( 11).04 (.71) — 1.03 (.91) — Coefficient PseudoR 2 Coefficient PseudoR 2 Random effects Level 1 variance ( 2)60.58 .00 21.94 .00 Intercept variance ( 00)40.45 .00 83.67 .00 Growth curve variance ( 11)2.21 .00 8.69 .00 Effects for fathers (SE) (SE) Fixed effects Intercept ( 00)17.79 (1.94) — 17.85 (1.49) — Growth function ( 01)2.01 (.63).17 1.88 (.33) .18 PCIT Format a( 10) 1.88 (3.18) .07 .85 (2.97) .04 Interaction ( 11).87 (1.25) — .13 (.66) — Coefficient PseudoR 2 Coefficient PseudoR 2 Random effects Level 1 variance ( 2)44.71 .00 27.08 .00 Intercept variance ( 00)106.33 .00 68.79 .00 Growth curve variance ( 11)9.79 .00 1.12 .02 Note.In calculating PseudoR 2, baseline variance estimates for mothers were 2 59.77, 00 39.84, 11 2.17 for Model 1 and 2 21.95, 00 82.52, 11 8.56 for Model 2. Baseline estimates for fathers were 2 44.70, 00 105.96, 11 9.67 for Model 1 and 2 27.08, 00 66.72, 11 1.14 for Model 2. aGives the relationship between participating in Group PCIT (vs. Individual) and the outcome. Estimated before entering the cross-level interaction. p .05. p .01. Table 8 Parents’ “Do Skills” Across Treatment Sessions Mothers Fathers Group PCITIndividual PCIT Group PCITIndividual PCIT SessionM SD M SD M SD M SD CDI 1 14.97 1.68 15.70 1.46 16.44 2.56 15.00 2.99 CDI 2 21.38 3.49 19.82 1.99 20.33 2.91 25.40 4.45 CDI 3 23.15 2.43 20.85 2.14 19.42 3.83 21.07 3.07 CDI 4 26.26 2.89 23.58 2.61 19.75 3.06 27.85 3.92 PDI 1 25.57 3.10 24.29 1.86 22.11 4.02 25.18 4.23 PDI 2 23.36 2.82 23.16 1.41 25.3 4.07 24.58 3.15 PDI 3 27.05 2.64 24.71 1.68 27.00 6.24 21.83 2.60 PDI 4 27.41 3.18 26.15 1.89 28.46 4.24 22.92 2.69 PDI 5 25.52 2.42 26.76 1.74 23.50 4.19 24.17 3.32 PDI 6 22.93 2.24 23.29 1.92 22.00 4.47 26.10 2.54 PDI 7 29.20 2.89 24.88 2.16 29.00 4.44 22.18 2.80 Note. “Do Skills” sum of praises, reflections, and behavior descrip- tions by parent toward child during 5-min DPICS Child-led Play interac- tion. CDI Child-directed Interaction coaching session; PDI Parent- directed Interaction coaching session. Table 9 Parents’ “Don’t Behaviors” Across Treatment Sessions Mothers Fathers Group PCITIndividual PCIT Group PCITIndividual PCIT SessionM SD M SD M SD M SD CDI 1 19.94 1.86 22.778 2.33 17.94 2.19 20.07 2.77 CDI 2 13.21 1.83 15.32 1.58 16.67 3.41 17.00 3.56 CDI 3 13.74 2.00 14.26 1.19 10.17 2.31 14.21 3.47 CDI 4 12.39 1.76 11.85 1.67 14.42 2.49 10.54 2.63 PDI 1 11.43 2.20 11.82 1.27 13.33 3.40 11.27 3.29 PDI 2 9.36 1.79 10.72 1.63 8.20 1.70 12.25 4.21 PDI 3 8.80 1.57 10.04 1.33 8.80 2.92 12.33 3.51 PDI 4 9.82 1.91 9.85 1.17 9.85 1.91 10.62 2.02 PDI 5 11.57 2.02 11.92 1.67 11.80 2.64 12.67 3.94 PDI 6 11.85 2.21 10.04 2.06 15.17 4.48 12.80 4.54 PDI 7 9.80 1.55 10.00 1.54 9.25 1.70 11.81 3.23 Note. “Don’t Behaviors” sum of questions, commands, and negative talk by parent toward child during 5-min DPICS Child-led Play interaction.

CDI Child-directed Interaction coaching session; PDI Parent-directed Interaction coaching session. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 692 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM variance in parents’ skill acquisition for both the “Do Skills” and “Don’t Behaviors” for both mothers and fathers occurred over time and often reflected individual differences, but skill acquisition was not affected by whether parents were in group or individual PCIT.

Clinical Significance To explore the clinical significance of changes in child and parent functioning, we computed the percentage of cases demon- strating reliable change between the intake and posttreatment measurements for primary outcome measures. The reliable change index (RC) was computed using the procedure proposed byJacob- son and Follette (1984), and theSEMwas computed using the normative test–retest reliability of the measure. RC indices more extreme than the 1.96 indicated either a reliable increase or decrease in functioning, with the direction depending on the spe- cific measure. The percentage of cases showing reliable improve- ment according to mothers’ reports was 60% for child conduct problems as assessed by the ECBI; 36.3% for externalizing be- haviors, 58.8% for internalizing behaviors, and 18.5% for adapt- ability on the BASC, and 31.5% for parental stress. For fathers, the percentage of cases showing reliable improvement was 53.5% for ECBI-rated child conduct problems, 27.3% for externalizing prob- lems, 20.5% for internalizing problems, and 18.2% for adaptability on the BASC, and 20.9% for parental stress. Further, of the 28 families who completed individual PCIT, 20 children (71%) moved from the clinical range at intake to within normal limits at 6-month follow-up on the ECBI Intensity Scale. Similarly, of the 29 families who completed Group PCIT, 20 children (69%) moved from the clinical range to within normal limits.

Coaching Time One challenge of implementing PCIT in a group format is ensuring that parents receive sufficient in vivo coaching (Niec et al., 2005). To test whether the amount of coaching that parents received varied across treatment conditions, we conducted inde- pendent samplesttests for mothers and fathers. The mean amount of coaching time per session for mothers did not significantly differ for participants in group and individual PCIT (Group PCIT coaching in minutesM(SD) 18.98(9.75), individual PCIT coach- ing in minutesM(SD) 19.91(9.00);t(603) 1.22,p .05; Cohen’sd 0.10). Fathers, however, received significantly less coaching if they participated in individual PCIT,M(SD) 15.52(5.81), rather than Group PCIT (M(SD) 18.69(8.24); t(268) 3.69,p .001; Cohen’sd 0.45).

Attrition, Adherence, and Social Support Attrition.Of the 81 families allocated to an intervention, 77 began treatment (seeFigure 1). Fifty-seven of the 77 families completed treatment, for an overall intervention attrition rate of 26%. Group PCIT had an attrition rate of 25.6% (n 10), while individual PCIT had an attrition rate of 26.3% (n 10). The phi coefficient testing the association between treatment condition and attrition was not significant ( .01,p .05).

Adherence.Families’ adherence to treatment was assessed in three ways: (a) number of sessions attended, (b) parent-reported completion of weekly homework assignments, and (c) therapist-rated in-session participation. To test whether the format of PCIT affected parents’ adherence, we conducted independent samplest tests on each measure. Neither mothers nor fathers demonstrated significant differences in attendance, homework completion, or therapist ratings of participation across treatment conditions. In other words, treatment adherence as assessed from three different perspectives was not statistically different for group and individual PCIT (seeTable 10).

Social support.Repeated measures ANOVAs were con- ducted to evaluate the effects of treatment condition and time on mothers’ and fathers’ perceptions of social support as rated on the MSPSS-C (Tables 4and6). Fathers, but not mothers, showed a significant main effect for time, with fathers reporting experienc- ing significantly more social support from intake to posttreatment and intake to follow-up. Neither mothers nor fathers demonstrated a significant effect for treatment condition. Contrary to the hy- pothesis that parents in group treatment would report more social support after treatment, there were no significant interaction ef- fects across time and treatment condition for either parents. Treatment Satisfaction Independent samplesttests revealed that parents reported high levels of satisfaction with PCIT (above 40 on a 50-point scale) regardless of whether it was delivered in a group or individual format (Mothers Individual PCITM(SD) 44.60(4.27); Mothers Group PCITM(SD) 45.06(4.20),t(100) 0.54,p .05; Cohen’s d 0.11; Fathers Individual PCITM(SD) 43.87(3.76); Fathers Group PCITM(SD) 45.52(3.60),t(36) 1.36,p .05; Cohen’s d 0.45). Discussion Young children with conduct problems are at high risk for serious negative consequences as they reach school-age and ado- lescence, including for example, poor academic functioning, peer rejection, antisocial activity, incarceration, and substance abuse (Moffitt, Caspi, Harrington, & Milne, 2002;Van Lier & Koot, 2010). Research indicates that early interventions are more effec- tive in comparison to interventions for older children (Frick, 2012). However, system factors such as the cost of services and shortages of child mental health staff, and family factors such as premature termination and poor treatment adherence prevent many families from benefiting from services (Kazdin, 2008;Satcher, 2000).

Table 10 Parent Adherence to Treatment Mothers Fathers Group PCITIndividual PCITGroup PCITIndividual PCIT VariableM SD M SD M SD M SD Participation 2.78 .31 2.77 .31 2.48 .60 2.46 .58 Homework 4.07 1.11 3.92 1.19 3.17 1.35 3.32 1.31 Attendance 12.28 4.08 10.72 4.99 9.81 5.35 9.32 4.87 Note. Independent samplest-tests for all comparisons between Group and Individual PCIT conditions were not statistically significant. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 693 GROUP PARENT–CHILD INTERACTION THERAPY To begin addressing the problems of availability and adherence, we evaluated the relative efficacy of an innovative adaptation of parent– child interaction therapy, Group PCIT. The study design was rigorous in that the innovative model was tested against an individual format of PCIT. Results suggest that Group PCIT is an efficacious intervention with outcomes that are not inferior to individual PCIT. Families participating in Group PCIT experi- enced improvements in multiple domains of functioning as as- sessed by differing methods and sources. That is, parents reported reductions in their children’s conduct problems and internalizing symptoms, as well asincreases in their children’s adaptive skills.

Parents also reported significant reductions in their parenting stress. Effect sizes for reductions in child conduct problems and parenting stress were large, which is consistent with an earlier trial of PCIT (McCabe & Yeh, 2009; child conduct problems Cohen’sd 1.41; parenting stress Cohen’sd 1.04). The improvement in child functioning experienced by families in Group PCIT were not inferior to those experienced by families in the individual condition. Further, observations of parenting behaviors demonstrated that parents’ child-centered skills in- creased significantly during treatment, while negative parent behaviors decreased. Similarly to the parent-reported improve- ments, the observed changes in parents’ behaviors over time were not influenced by the format of PCIT that families re- ceived. Treatment gains were not only statistically significant, but also demonstrated clinical significance. The findings are promising as children whose conduct is within the typical range of functioning are less likely to experience negative conse- quences in later childhood and adolescence than those whose conduct remains disturbed (Eyberg et al., 2014).

Challenges and Benefits of Group PCIT One of the challenges identified when implementing PCIT in a group format is ensuring that every parent receives a sufficient amount of coaching during each session (Niec et al., 2005). Our comparison of coaching time across treatment conditions revealed that families did not receive less coaching in the group format.

This finding helps to allay concerns that a group format of PCIT must sacrifice coaching time for other elements of the intervention or that it is impossible to give parents in a group sufficient coaching to alter their skills. While we did find that fathers who participated in Group PCIT received more coaching than fathers in individual PCIT, we encourage caution in interpreting this result.

Within the individual format, when two parents participate in treatment, the second parent coached receives less therapist coach- ing. This disparity is meant to be eliminated by alternating the order of parents coached across treatment sessions. If, counter to the protocol, therapists do not routinely rotate parents’ order of coaching, one parent will receive less coaching overall. Alternat- ing coaching order may be clinically valuable, as it allows parents to interact with their children in different conditions (e.g., at the beginning or the end of the play) and to observe their partner using the parenting skills before their own coaching. Although parents in the group format were coached in alternating order, the potential to inadvertently receive less coaching did not arise as each parent– child dyad was coached for approximately the same amount of time in every session.One of the hypothesized benefits of group parenting interven- tions is the opportunity for parents to interact with other parents of children with difficult behaviors. When caregivers develop rela- tionships with others who have similar experiences, it may reduce their perceptions of stigma around having a child with conduct problems (Webster-Stratton & Herbert, 1993). Although some research has proposed that the reduction of stigma and existence of group support may help to engage families in treatment (Webster- Stratton, 1997;Webster-Stratton & Herbert, 1993), our findings did not support the hypothesis that parents participating in Group PCIT would experience greater feelings of social support as relates to their children. It is worth noting that despite the lack of statis- tically significant differences across treatment conditions on our measure of social support (MSPSS-C), anecdotal evidence sug- gests that Group PCIT provided opportunities to receive support that the individual format did not. For instance, families in group treatment routinely reported calling one another with reminders to complete homework or questions about skill use. In some groups, families carpooled to sessions. Families reported planning post- group activities and presented therapists with group pictures.

While we did not systematically assess this qualitative information and, therefore, cannot interpret it, the possibility exists that the lack of differences on the MSPSS-C may be due in part to a failure to tap the types of support experienced in group treatment. The qualitative evidence suggests that additional study of the potential differential benefits of Group PCIT is warranted.

Similar to the findings for social support, families in Group PCIT demonstrated no significant differences in their treatment adherence relative to families in individual PCIT. While Group PCIT cannot therefore be considered superior to individual PCIT in this domain, our finding does counter the concern that the group format does not provide families with enough tailored attention to maintain their participation. Furthermore, while as many as 40 to 60% of children who begin outpatient mental health treatment drop out prematurely (Kazdin, 2008), the overall intervention attrition rate for this study was 26%, with no differences across conditions.

Finally, not only did families in Group PCIT demonstrate treat- ment adherence that did not differ from families in individual PCIT, they also expressed as much satisfaction with the interven- tion. In both conditions, treatment satisfaction was high. Limitations and Future Directions While treatment attrition for families in Group and individual PCIT was (Eyberg et al., 2014) relatively low, study attrition was somewhat higher, with 33% of families who were allocated to an intervention failing to complete the follow-up assessment 6 months after treatment. No significant differences in study attrition occurred across treatment conditions (36% Individual PCIT, 31% Group PCIT; .05,p .05), but attrition at follow-up limits our ability to draw conclusions regarding the maintenance of treatment effects. Maintaining study participation in clinical re- search has long been a challenge (Foster & Bickman, 1996).

Research on participant follow-up in longitudinal designs suggests that one key factor to reducing attrition is maintaining contact with participants during the follow-up period (David, Alati, Ware, & Kinner, 2013). A number of other strategies have been described to improve retention, such as providing rewards, establishing a rec- ognizable study “brand” with logos on all contacts, maintaining This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 694 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM regular tracking, and designing brief follow-up assessments (Ma- son, 1999). Although we implemented retention strategies consis- tent with the literature (e.g., obtaining multiple contacts for fam- ilies, keeping families informed of the follow-up schedule, providing reasonable rewards for completing assessments, provid- ing interim phone calls midway between posttreatment and follow- up), they were insufficient to obtain good retention rates ( 20%; Mason, 1999). Future research should continue to explore study, therapist, and family factors in parent management training re- search that contribute to retention.

The underlying goal of this study was to determine whether Group PCIT may be a valuable option to help address shortages in the availability of evidence-based interventions for young children with conduct problems. Our findings suggest that Group PCIT is an efficacious intervention, with important par- ent and child outcomes comparable to individual PCIT. Further, the findings here support the feasibility of using the live coach- ing model, which is central to PCIT, in a group format. Future work remains necessary to determine whether Group PCIT reduces barriers to treatment. Specifically, we recommend con- sideration of the following issues.

First, the PCIT protocol in the present study differed from the standard PCIT protocol (Eyberg & Funderburk, 2011)inan important aspect: families did not progress in treatment based on their mastery of parenting skills; rather, families in both group and individual PCIT moved through treatment at the same pace. This design allowed us to control for the impact of dosage on treatment outcome across formats, and was consis- tent with past research that used session-limited PCIT (e.g., Chaffin et al., 2011;Nixon et al., 2004). However, limiting the number of treatment sessions may have implications for fami- lies’ success and maintenance of treatment gains. Some find- ings suggest that families who remain in treatment until their children’s behaviors are within normal limits may have more time to establish their acquisition of parenting skills and main- tain gains longer (Eyberg et al., 2014). While treatment gains in the present study remained significant from intake to 6-month follow-up, given the limitations already expressed, it would be valuable to evaluate the outcome of time-limited Group PCIT with a longer follow-up period before broadly disseminating the model.

Second, although parents’ CDI skills (“Do Skills”) were evaluated throughout treatment with a behavioral observation measure (DPICS), PDI skills were not regularly assessed. It is important to note that in both treatment conditions parents were taught the same sequence of child discipline techniques and all parents were required to complete daily homework to practice the new techniques. Subsequent to this instruction, children’s conduct problems were significantly reduced. However, future evaluation of the group format should collect additional data regarding parents’ use of the discipline strategies.

Third, the cost of implementing Group PCIT must be con- sidered. Assumptions have sometimes been made about the costs of group interventions. A few early studies focused on the reduced therapist time that is required to implement groups as an indication of its “cost-effectiveness,” without computing actual cost-effectiveness analyses (Barkley, 1987;Brightman, Baker, Clark, & Ambrose, 1982). Group PCIT may be a suc- cessful way to reduce the costs of treatment for families. Forinstance, when multiple families are treated simultaneously, fewer sessions may be cancelled for families who do not show, thus improving therapists’ productivity. Especially within com- munity agencies, greater productivity may make an intervention more sustainable. However, a need remains to investigate the actual costs of Group PCIT before decisions are made regarding the most economical format of the intervention.

Finally, it is important to note that within the Group PCIT condition, some measures of outcome were influenced by the specific therapy group in which a family participated. Therapy group influenced both mothers’ and fathers’ treatment satisfaction, and also influenced mothers’ acquisition of child-centered parent- ing skills. Although the sample size of this study did not allow for exploration of group characteristics that might make a PCIT group more or less effective, future research should examine these fac- tors. Furthermore, clinicians considering implementing group treatment should be deliberate in their selection of families, as our findings indicate that the composition of the group may matter to family outcome. Clinical Implications The present study suggests that Group PCIT is a promising format to reach more families with young children with conduct problems. Our model of Group PCIT included the core compo- nents of individual PCIT (e.g., weekly behavioral assessment, live coaching). Families in Group PCIT experienced reductions in child conduct problems that were not inferior to individual PCIT. Fam- ilies in both treatment conditions demonstrated significant im- provements in their child-centered parenting skills and in both treatment conditions families expressed high levels of satisfaction with the treatment. To make a significant public health impact, we must increase the availability of evidence-based practices for young children with conduct problems. The findings of the current study warrant the continued development and evaluation of the group format of PCIT. References Abidin, R. (1995).Parenting Stress Index: Professional manual. San Antonio, TX: Psychological Assessment Resources, Inc.

American Psychiatric Association. (2000).Diagnostic and statistical man- ual of mental disorders: DSM–IV–TR(4th ed. Text Revision). Wash- ington, DC: Author.

Bagner, D. M., & Eyberg, S. M. (2007). Parent-child interaction therapy for disruptive behavior in children with mental retardation: A random- ized controlled trial.Journal of Clinical Child and Adolescent Psychol- ogy, 36,418 – 429.http://dx.doi.org/10.1080/15374410701448448 Barkley, R. A. (1987). What is the role of group parent training in the treatment of ADD children?Journal of Children in Contemporary Society, 19,143–151.http://dx.doi.org/10.1300/J274v19n01_12 Barnett, M. L., Niec, L. N., & Acevedo-Polakovich, I. D. (2014). Assessing the Key to Effective coaching in parent– child interaction therapy: The Therapist-Parent Interaction Coding System.Journal of Psychopathol- ogy and Behavioral Assessment, 36,211–223. Advance online publica- tion.

Brestan, E. V., & Eyberg, S. M. (1998). Effective psychosocial treatments for children and adolescents with disruptive behavior disorders: 29 years, 82 studies, and 5272 kids.Journal of Clinical Child Psychology, 27,179 –188.

Brestan, E. V., Eyberg, S. M., Boggs, S. R., & Algina, J. (1997). Parent- child interaction therapy: Parent perceptions of untreated siblings.Child This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 695 GROUP PARENT–CHILD INTERACTION THERAPY & Family Behavior Therapy, 19,13–28.http://dx.doi.org/10.1300/ J019v19n03_02 Brightman, R. P., Baker, B. L., Clark, D. B., & Ambrose, S. A. (1982).

Effectiveness of alternative parent training formats.Journal of Behavior Therapy and Experimental Psychology, 13,113–117.http://dx.doi.org/ 10.1016/0005-7916(82)90051-9 Chaffin, M., Funderburk, B., Bard, D., Valle, L. A., & Gurwitch, R. (2011).

A combined motivation and parent-child interaction therapy package reduces child welfare recidivism in a randomized dismantling field trial.

Journal of Consulting and Clinical Psychology, 79,84 –95. Advance online publication.http://dx.doi.org/10.1037/a0021227 Chaffin, M., Silovsky, J. F., Funderburk, B., Valle, L. A., Brestan, E. V., Balachova, T.,...Bonner, B. L. (2004). Parent-child interaction therapy with physically abusive parents: Efficacy for reducing future abuse reports.Journal of Consulting and Clinical Psychology, 72,500 –510.

http://dx.doi.org/10.1037/0022-006X.72.3.500 Chambless, D. L., & Ollendick, T. H. (2001). Empirically supported psychological interventions: Controversies and evidence.Annual Review of Psychology, 52,685–716.http://dx.doi.org/10.1146/annurev.psych.52 .1.685 David, M. C., Alati, R., Ware, R. S., & Kinner, S. A. (2013). Attrition in a longitudinal study with hard-to-reach participants was reduced by ongoing contact.Journal of Clinical Epidemiology, 66,575–581.http:// dx.doi.org/10.1016/j.jclinepi.2012.12.002 Dodrill, C. B. (1981). An economical method for the evaluation of general intelligence in adults.Journal of Consulting and Clinical Psychology, 49,668 – 673.http://dx.doi.org/10.1037/0022-006X.49.5.668 Dodrill, C. B., & Warner, M. H. (1988). Further studies of the Wonderlic Personnel Test as a brief measure of intelligence.Journal of Consulting and Clinical Psychology, 56,145–147.http://dx.doi.org/10.1037/0022- 006X.56.1.145 Dunn, L. M., & Dunn, L. M. (1997).Examiner’s Manual for the Peabody Picture Vocabulary Test(3rd ed.). Circle Pines, MN: American Guid- ance Service.

Eisenstadt, T. H., Eyberg, S., McNeil, C. B., Newcomb, K., & Funderburk, B. (1993). Parent-child interaction therapy with behavior problem chil- dren: Relative effectiveness of two stages and overall treatment out- come.Journal of Clinical Child Psychology, 22,42–51.http://dx.doi .org/10.1207/s15374424jccp2201_4 Eyberg, S. M. (1993). Consumer satisfaction measures for assessing parent training programs. In L. Vandecreek, S. Knapp, & T. Jackson (Eds.), Innovations in clinical practices: A source book(Vol. 12, pp. 261–271).

Sarasota, FL: Professional Resource Press.

Eyberg, S., Boggs, S., & Jaccard, J. (2014). Does maintenance treatment matter?Journal of Abnormal Child Psychology, 42,355–366.http://dx .doi.org/10.1007/s10802-013-9842-9 Eyberg, S. M., Boggs, S., & Rodriguez, C. (1993). Relationships between maternal parenting stress and child disruptive behavior.Child & Family Behavior Therapy, 14,1–9.http://dx.doi.org/10.1300/J019v14n04_01 Eyberg, S. M., & Funderburk, B. (2011).Parent-child interaction therapy protocol. Gainesville, FL: Parent-Child Interaction Therapy Interna- tional, Inc.

Eyberg, S. M., & Matarazzo, R. G. (1980). Training parents as therapists:

A comparison between individual parent-child interaction training and parent group didactic training.Journal of Clinical Psychology, 36, 492– 499.http://dx.doi.org/10.1002/jclp.6120360218 Eyberg, S. M., Nelson, M. M., & Boggs, S. R. (2008). Evidence-based psychosocial treatments for children and adolescents with disruptive behavior.Journal of Clinical Child and Adolescent Psychology, 37, 213–235.http://dx.doi.org/10.1080/15374410701820117 Eyberg, S. M., Nelson, M. M., Duke, M., & Boggs, S. R. (2010).Manual for the Dyadic Parent-Child Interaction Coding System(3rd ed.). Un- published manual.Eyberg, S. M., & Pincus, D. (1999).Eyberg Child Behavior Inventory/ Sutter-Eyberg Student Behavior Inventory-Revised professional manual. Odessa, FL: Psychological Assessment Resources, Inc.

Feinberg, M., Neiderhiser, J., Howe, G., & Hetherington, E. M. (2001).

Adolescent, parent, and observer perceptions of parenting: Genetic and environmental influences on shared and distinct perceptions.Child De- velopment, 72,1266 –1284.http://dx.doi.org/10.1111/1467-8624.00346 Forster, A., Eyberg, S., & Burns, L. (1990). Assessing the verbal behavior of conduct problem children during mother-child interactions: A pre- liminary investigation.Child & Family Behavior Therapy, 12,13–22.

http://dx.doi.org/10.1300/J019v12n01_02 Foster, E. M., & Bickman, L. (1996). An evaluator’s guide to detecting attrition problems.Evaluation Review, 20,695–723.http://dx.doi.org/10 .1177/0193841X9602000603 Frick, P. J. (2012). Developmental pathways to conduct disorder: Impli- cations for future directions in research, assessment, and treatment.

Journal of Clinical Child and Adolescent Psychology, 41,378 –389.

http://dx.doi.org/10.1080/15374416.2012.664815 Funderburk, B., Eyberg, S., Rich, B., & Behar, L. (2003). Further psycho- metric evaluation of the Eyberg and Behar rating scales for parents and teachers of preschoolers.Early Education and Development, 14,67– 82.

http://dx.doi.org/10.1207/s15566935eed1401_5 Grills, A. E., & Ollendick, T. H. (2002). Issues in parent-child agreement:

The case of structured diagnostic interviews.Clinical Child and Family Psychology Review, 5,57– 83.http://dx.doi.org/10.1023/A:1014 573708569 Herschell, A. D., Capage, L. C., Bahl, A. B., & McNeil, C. B. (2008). The role of therapist communication style in parent-child interaction therapy.

Child & Family Behavior Therapy, 30,13–35.http://dx.doi.org/10.1300/ J019v30n01_02 Higgins, J., & Green, S. (Eds.; 2011).Cochrane Handbook for Systematic Reviews of InterventionsVersion 5.1.0 [updated March 2011]. The Cochrane Collaboration. Retrieved fromwww.cochrane-handbook.org Hood, K. K., & Eyberg, S. M. (2003). Outcomes of parent-child interaction therapy: Mothers’ reports of maintenance three to six years after treat- ment.Journal of Clinical Child and Adolescent Psychology, 32,419 – 429.http://dx.doi.org/10.1207/S15374424JCCP3203_10 Hox, J. J. (2010).Multilevel analysis: Techniques and applications(2nd ed.). New York, NY: Routledge.

Jacobson, N. S., & Follette, W. C. (1984). Psychotherapy outcome re- search: Methods for reporting variability and evaluating clinical signif- icance.Behavior Therapy, 15,336 –352.http://dx.doi.org/10.1016/ S0005-7894(84)80002-7 Kaminski, J. W., Valle, L. A., Filene, J. H., & Boyle, C. L. (2008). A meta-analytic review of components associated with parent training program effectiveness.Journal of Abnormal Child Psychology, 36, 567–589.http://dx.doi.org/10.1007/s10802-007-9201-9 Kazdin, A. E. (2008). Evidence-based treatment and delivery of psycho- logical services: Shifting our emphases to increase impact.Psychologi- cal Services, 5,201–215.http://dx.doi.org/10.1037/a0012573 Kazdin, A. E., & Blase, S. L. (2011). Rebooting psychotherapy research and practice to reduce the burden of mental illness.Perspectives on Psychological Science, 6,21–37.http://dx.doi.org/10.1177/1745 691610393527 Kernan, W. N., Viscoli, C. M., Makuch, R. W., Brass, L. M., & Horwitz, R. I. (1999). Stratified randomization for clinical trials.Journal of Clinical Epidemiology, 52,19 –26.http://dx.doi.org/10.1016/S0895- 4356(98)00138-3 Levac, A. M., McCay, E., Merka, P., & Reddon-D’Arcy, M. L. (2008).

Exploring parent participation in a parent training program for children’s aggression: Understanding and illuminating mechanisms of change.

Journal of Child and Adolescent Psychiatric Nursing, 21,78 – 88.http:// dx.doi.org/10.1111/j.1744-6171.2008.00135.x This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 696 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM MacKenzie, E. P., Fite, P. J., & Bates, J. E. (2004). Predicting outcome in behavioral parent training: Expected and unexpected results.Child & Family Behavior Therapy, 26,37–53.http://dx.doi.org/10.1300/ J019v26n02_03 Mason, M. (1999). Methods, plainly speaking: A review of procedural and statistical methods for handling attrition and missing data in clinical research.Measurement and Evaluation in Counseling and Development, 32,111–118.

McCabe, K., & Yeh, M. (2009). Parent-child interaction therapy for Mexican Americans: A randomized clinical trial.Journal of Clinical Child and Adolescent Psychology, 38,753–759.http://dx.doi.org/10 .1080/15374410903103544 McKay, M. M., Harrison, M. E., Gonzales, J., Kim, L., & Quintana, E.

(2002). Multiple-family groups for urban children with conduct diffi- culties and their families.Psychiatric Services, 53,1467–1468.http:// dx.doi.org/10.1176/appi.ps.53.11.1467 McNeil, C., Eyberg, S., Eisenstadt, T., Newcomb, K., & Funderburk, B.

(1991). Parent-child interaction therapy with behavior problem children:

Generalization of treatment effects to the school setting.Journal of Clinical Child Psychology, 20,140 –151.http://dx.doi.org/10.1207/ s15374424jccp2002_5 McNeil, C. B., & Hembree-Kigin, T. (2010).Parent-child interaction therapy(2nd ed.). New York, NY: Springer Science & Business Media.

http://dx.doi.org/10.1007/978-0-387-88639-8 Moffitt, T. E., Caspi, A., Harrington, H., & Milne, B. J. (2002). Males on the life-course-persistent and adolescence-limited antisocial pathways:

Follow-up at age 26 years.Development and Psychopathology, 14, 179 –207.http://dx.doi.org/10.1017/S0954579402001104 Niec, L. N., Cochran, H. M., & Stayer, C. A. (2003, July).Social support and parental functioning in parents with behavior-disordered children.

Poster presented at the Fourth Annual parent-child interaction therapy Conference, Sacramento, CA.

Niec, L. N., Hemme, J., Yopp, J., & Brestan, E. (2005). Parent-child interaction therapy: The rewards and challenges of a group format.

Cognitive and Behavioral Practice, 12,113–125.http://dx.doi.org/10 .1016/S1077-7229(05)80046-X Nieter, L., Thornberry, T., & Brestan-Knight, E. (2013). The effectiveness of group parent-child interaction therapy with community families.

Journal of Child and Family Studies, 22,490 –501.http://dx.doi.org/10 .1007/s10826-012-9601-5 Nixon, R. D. V., Sweeney, L., Erickson, D. B., & Touyz, S. W. (2004).

Parent-child interaction therapy: One- and two-year follow-up of stan- dard and abbreviated treatments for oppositional preschoolers.Journal of Abnormal Child Psychology, 32,263–271.http://dx.doi.org/10.1023/ B:JACP.0000026140.60558.05 Patterson, G. (1982).A social learning approach: Vol. 3.Coercive family process. Eugene, OR: Castalia.

Pisterman, S., McGrath, P., Firestone, P., Goodman, J. T., Webster, I., & Mallory, R. (1989). Outcome of parent-mediated treatment of preschool- ers with attention deficit disorder with hyperactivity.Journal of Con- sulting and Clinical Psychology, 57,628 – 635.http://dx.doi.org/10 .1037/0022-006X.57.5.628 Raudenbush, S. W., & Bryk, A. S. (2002).Hierarchical linear models:

Applications and data analysis methods(2nd ed.). Newbury Park, CA:

Sage.

Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon, R. T., & du Toit, M. (2011).HLM 7: Hierarchical linear and nonlinear modeling. Lin- colnwood, IL: SSI, Inc.

Reimers, T. M., Wacker, D. P., Cooper, L. J., & de Raad, A. O. (1992).

Clinical evaluation of the variables associated with treatment acceptabil- ity and their relation to compliance.Behavioral Disorders, 18,67–76.

Reynolds, C., & Kamphaus, R. (2004).Behavior assessment system for children(2nd ed.). Circle Pines, MN: American Guidance Service.Rich, B., & Eyberg, S. (2001). Accuracy of assessment: The discriminative and predictive power of the Eyberg Child Behavior Inventory.Ambula- tory Child Health, 7,249 –257.http://dx.doi.org/10.1046/j.1467-0658 .2001.00141.x Robinson, E. A., & Eyberg, S. M. (1981). The dyadic parent-child inter- action coding system: Standardization and validation.Journal of Con- sulting and Clinical Psychology, 49,245–250.http://dx.doi.org/10.1037/ 0022-006X.49.2.245 Rogers, J. L., Howard, K. I., & Vessey, J. T. (1993). Using significance tests to evaluate equivalence between two experimental groups.Psycho- logical Bulletin, 113,553–565.http://dx.doi.org/10.1037/0033-2909.113 .3.553 Ross, C., Blanc, H., McNeil, C., Eyberg, S. M., & Hembree-Kigin, T.

(1998). Parenting stress in mothers of young children with oppositional defiant disorder and other severe behavior problems.Child Study Jour- nal, 28,93–110.

Ruma, P., Burke, R., & Thompson, R. (1996). Group parent training: Is it effective for children of all ages?Behavior Therapy, 27,159 –169.

http://dx.doi.org/10.1016/S0005-7894(96)80012-8 Satcher, D. (2000). Mental health: A report of the Surgeon General— Executive summary.Professional Psychology: Research and Science, 31,5–13.http://dx.doi.org/10.1037/0735-7028.31.1.5 Schuhmann, E. M., Foote, R. C., Eyberg, S. M., Boggs, S. R., & Algina, J. (1998). Efficacy of parent-child interaction therapy: Interim report of a randomized trial with short-term maintenance.Journal of Clinical Child Psychology, 27,34 – 45.http://dx.doi.org/10.1207/s15374 424jccp2701_4 Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, M. E.

(2000). NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses.Journal of the American Acad- emy of Child & Adolescent Psychiatry, 39,28 –38.http://dx.doi.org/10 .1097/00004583-200001000-00014 Shanley, J. R., & Niec, L. N. (2010). Coaching parents to change: The impact of in vivo feedback on parents’ acquisition of skills.Journal of Clinical Child and Adolescent Psychology, 39,282–287.http://dx.doi .org/10.1080/15374410903532627 Sheeber, L., & Johnson, J. (1994). Evaluation of a temperament-focused, parent-training program.Journal of Clinical Child Psychology, 23,249 – 259.http://dx.doi.org/10.1207/s15374424jccp2303_3 Stormont, M. (2002). Externalizing behavior problems in young children:

Contributing factors and early intervention.Psychology in the Schools, 39,127–138.http://dx.doi.org/10.1002/pits.10025 van Lier, P. A., & Koot, H. M. (2010). Developmental cascades of peer relations and symptoms of externalizing and internalizing problems from kindergarten to fourth-grade elementary school.Development and Psy- chopathology, 22,569 –582.http://dx.doi.org/10.1017/S095457 9410000283 Walker, E., & Nowacki, A. (2010). Understanding equivalence and non- inferiority testing.Journal of General Internal Medicine, 26,192–196.

http://dx.doi.org/10.1007/s11606-010-1513-8 Webster-Stratton, C. (1985). Mother perceptions and mother-child inter- actions: Comparison of a clinic-referred and a nonclinic group.Journal of Clinical Child Psychology, 14,334 –339.http://dx.doi.org/10.1207/ s15374424jccp1404_11 Webster-Stratton, C. (1990). Enhancing the effectiveness of self- administered videotape parent training for families with conduct- problem children.Journal of Abnormal Child Psychology, 18,479 – 492.

http://dx.doi.org/10.1007/BF00911103 Webster-Stratton, C. (1997). From parent training to community building.

Families in Society, 78,156 –171.http://dx.doi.org/10.1606/1044-3894 .755 Webster-Stratton, C., & Hammond, M. (1997). Treating children with early-onset conduct problems: A comparison of child and parent training This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 697 GROUP PARENT–CHILD INTERACTION THERAPY interventions.Journal of Consulting and Clinical Psychology, 65,93– 109.http://dx.doi.org/10.1037/0022-006X.65.1.93 Webster-Stratton, C., & Herbert, M. (1993). What really happens in parent training?Behavior Modification, 17,407– 456.http://dx.doi.org/10 .1177/01454455930174002 Webster-Stratton, C. H., Reid, M. J., & Beauchaine, T. (2011). Combining parent and child training for young children with ADHD.Journal of Clinical Child and Adolescent Psychology, 40,191–203.http://dx.doi .org/10.1080/15374416.2011.546044 Whitcomb, S., & Merrell, K. W. (2013).Behavioral, social, and emotional assessment of children and adolescents(4th ed.). New York, NY:

Routledge.Zimet, G., Dahlem, N., Zimet, S., & Farley, G. (1988). The multidimen- sional scale of perceived social support.Journal of Personality Assess- ment, 52,30 – 41.http://dx.doi.org/10.1207/s15327752jpa5201_2 Received May 2, 2014 Revision received November 18, 2015 Accepted February 8, 2016 Members of Underrepresented Groups:

Reviewers for Journal Manuscripts Wanted If you are interested in reviewing manuscripts for APA journals, the APA Publications and Communications Board would like to invite your participation. Manuscript reviewers are vital to the publications process. As a reviewer, you would gain valuable experience in publishing. The P&C Board is particularly interested in encouraging members of underrepresented groups to participate more in this process.

If you are interested in reviewing manuscripts, please write APA Journals at [email protected].

Please note the following important points:

•To be selected as a reviewer, you must have published articles in peer-reviewed journals. The experience of publishing provides a reviewer with the basis for preparing a thorough, objective review.

•To be selected, it is critical to be a regular reader of the five to six empirical journals that are most central to the area or journal for which you would like to review. Current knowledge of recently published research provides a reviewer with the knowledge base to evaluate a new submission within the context of existing research.

•To select the appropriate reviewers for each manuscript, the editor needs detailed information.

Please include with your letter your vita. In the letter, please identify which APA journal(s) you are interested in, and describe your area of expertise. Be as specific as possible. For example, “social psychology” is not sufficient—you would need to specify “social cognition” or “attitude change” as well.

•Reviewing a manuscript takes time (1– 4 hours per manuscript reviewed). If you are selected to review a manuscript, be prepared to invest the necessary time to evaluate the manuscript thoroughly.

APA now has an online video course that provides guidance in reviewing manuscripts. To learn more about the course and to access the video, visit http://www.apa.org/pubs/authors/review- manuscript-ce-video.aspx. This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 698 NIEC, BARNETT, PREWETT, AND SHANLEY CHATHAM