Incredible Years – Parent

Study 1

Evaluation Methodology

Design: Interested mothers were continuously assigned at random to the BASIC parent training program (n=19) (four weeks, eight hours) or to a waiting-list control group (n=16, received program two weeks after post assessment). Assessments were conducted at baseline, immediate posttest, six week follow-up, and one year follow-up.

Sample: 35 non-clinic, middle-class, Caucasian families and their 3-5 year old children were recruited with a flyer announcing a parent program designed to help with common behavior problems. One year follow-up measures were completed by 32 of the original 35 mothers.

Measures: The Parent Attitude Survey (PAS) and Eyberg Child Behavior Inventory (ECBI) were used both pre- and post-treatment to assess changes in behavior. The PAS measured parents' perceptions of the parent-child interaction: confidence, causation, acceptance, understanding, and trust. The 36 item ECBI measured parental perceptions of their children's behavior problems. In addition to these two measures, observations were made of mother-child interactions in a playroom via a one-way mirror by observers who were blind to the intervention condition.

Outcomes

Baseline Equivalence: There were no significant differences between groups at baseline on demographic, socioeconomic, or dependent measures.

Posttest: At immediate posttest, there were significant differences on four of five variables that favored the intervention group, with mothers showing fewer lead-taking, dominance, and nonacceptance behaviors, and more positive affect (but no change in mother watch). Intervention children also showed improvement, with significant differences in observed negative affect behavior (defined as negative verbal behavior, teasing, direct hostility including hits, threats, snatches and other aggressive acts), submissive behaviors, and positive affect, all in the expected directions. There were no differences in the groups in terms of mother report variables on the PAS. Early treatment mothers rated significantly less intense behavior problems in their children than did the delayed treatment group (control group). At six-week follow-up, the initial changes in observed behaviors were maintained in the treatment group, and nearly identical changes were replicated with the waiting list control group after they received the program.

Long-term: One year after treatment ended, long-term effects were assessed by retesting 32 of the 35 original subjects on all measures. Comparisons were performed by (1) pretreatment vs. one-year posttreatment, and (2) immediate one-week post-treatment vs. one-year posttreatment. There continued to be no significant differences in mother reports on the PAS. However, there was a significant reduction in number of behavior problems reported by mothers and intensity of problems at one year compared to baseline; however, there was a non-significant increase from post-treatment to one-year post-treatment. According to independent observations, there were continued significant improvements in mother nonacceptance, dominance, and lead-taking behaviors and positive affect, as well as a significant increase in mother watch behaviors. Intervention children also showed a significant decrease in nonacceptance/oppositional behavior, dominance and submissiveness, compared to their baseline scores. When compared to the post-treatment scores, there was a significant change in three of the parent and three of the child variables. However, there was a significant drop in positive affect for mothers and children.

Follow-up analyses also compared mothers' reports of behavior problems on the ECBI and behavioral interactions to normative data on similar populations. At baseline, participating children's ECBI Problem and Intensity Scores were significantly higher than "normal" four year-olds, with 38 percent of the sample more than one standard deviation above the mean of the normal population. One year later, there was no longer a significant difference between the two populations. At baseline, intervention children also differed from the normal population in positive affect behaviors, but this difference was not demonstrated at follow-up. Similar results were found for mothers. Intervention mothers were significantly more submissive and less positive than normal mothers at baseline, but there was no difference between groups at follow-up.

Study 2

Evaluation Methodology

Design: A second study included 35 clinic-referred, low-income mothers with children aged 3-8 (25 boys and 10 girls) who had sought help at a children's hospital for their children's behavior problems. Mothers were admitted to the study based on their reports of children's oppositional behavior (e.g., refusing to follow commands, tantrums, and aggression) and were considered at high risk due to the single-parent status of many participants (54%), low socioeconomic status, low mean education level, high prevalence of child abuse, and the large number of negative behaviors of their children. Participants were continuously and randomly assigned to one of three groups:

• One-on-one personalized parent therapy (n=11)
• Videotape-based group therapy (BASIC) (n=13)
• Waiting-list control group (n=11)

Both treatment groups attended nine weekly sessions (18-20 hours) of therapy. The one-on-one treatment consisted of individual therapy in which facilitators "modeled" parent skills, and parents role-played interactions with their children while facilitators gave direct feedback via a microphone in the parents' ear (i.e., "bug-in-the-ear" therapy). These sessions focused on improving general parenting skills and correcting targeted child behavior problems. The BASIC program consisted of groups of parents discussing over 180 videotape vignettes of parent-child interactions. Because children did not attend these sessions, parents did not model behaviors or receive feedback on their performance. Those in the waiting-list control group were assigned to one of the two treatment groups three to four months after baseline measures.

Measures: Assessments were conducted at baseline, post treatment and one-year follow-up. Assessments included parent reports of behavior problems (using the Child Behavior Checklist, CBCL; Eyberg Child Behavior Inventory, ECBI for Intensity and Total number of problem behaviors; and Parent Daily Reports, PDR, to rate positive and negative child behaviors and use of spanking), blinded observations of parent-child interactions at home (assessing mothers' praise, critical statements, total commands, ineffective commands, and no opportunities; and children's deviancy and noncompliance); and teacher reports of children's problem behaviors in school (using the Behar Preschool Behavior Questionnaire, PBQ, which was only completed at the one-year follow-up).

Outcomes
Baseline Equivalence: There were no significant differences at baseline between groups on demographic or outcome variables.

Posttest: At post treatment, the BASIC and one-on-one groups both showed improvement on mother and child behaviors, with no differences between the two treatment groups. Thus, the BASIC treatment was as effective as the high-cost, one-on-one therapy, and both treatments were superior to the control group in regard to attitudinal and behavioral changes. More specifically, BASIC treatment children were significantly different from the control group on five of the six parent reports of children's problem behavior, with fewer total and less intense behavior problems according to the ECBI, fewer negative and more prosocial behaviors, and less use of spanking by mothers (but no changes according to the CBCL). According to independent observations, BASIC mothers improved on four of five measures of parenting skills (compared to two of five for the one-on-one treatment approach), and intervention children had less noncompliance (but not deviance) than control children.

Long-Term: At the one-year follow-up, 32 of the 35 families were reassessed (88 percent), and, again, no differences were noted between the two treatment groups in terms of demographic characteristics. Approximately 70 percent of both groups maintained significant positive behavioral changes at the one-year follow-up. Compared to baseline scores, BASIC children demonstrated continued improvements in parent reports of children's problem behaviors, as measured by both the CBCL and ECBI, and blinded observations indicated improvement in BASIC mothers' critical behaviors and total praise, as well as in children's noncompliance and total deviance. It is noteworthy that child deviance, which had shown borderline decreases at post treatment, showed significant changes one year later. There were no significant differences according to teacher reports of children's problem behaviors. The BASIC program was five times more cost-effective than one-on-one therapy, using 48 hours of facilitator time compared to 251 hours of facilitator time. Families who had little or no social support were most likely to relapse following treatment.

Study 3

Evaluation Methodology

Design: A third study was conducted to ascertain the most useful and cost-effective component of the BASIC program (facilitators, videotapes, or group discussions). The sample included 114 families of varying socioeconomic background who were either self-referred (43 percent) or professionally referred (57 percent) to the study based on their children's clinically significant problem behaviors. Study children ranged in age from 3 to 8 years and included 79 boys and 35 girls. Parents were continuously and randomly assigned to one of four groups:

• Videotape-based group therapy (BASIC) (n= 28)
• Individually self-administered videotape modeling therapy (IVM) (n= 29)
• Group therapy alone (GD) (n= 28)
• Waiting-list control group (n= 29)

The 28 mothers and 20 fathers in the BASIC condition met weekly at the clinic for twelve, two-hour sessions in which groups of 10 to 15 parents and a trained facilitator discussed over 200 videotaped vignettes. The 29 mothers and 20 fathers in the IVM condition met weekly at the clinic for 10 to 12 self-administered, hourly sessions in which they watched videotapes with no facilitator present. The 28 mothers and 19 fathers in the GD condition met in groups of 10 to 15 persons with trained facilitators to discuss parenting practices but did not watch any videotaped vignettes. Finally, the 29 mothers and 21 fathers in the control condition had no contact with facilitators, although they did receive the bi-weekly calls concerning their children's behaviors. After 12 weeks and repeat assessments, these individuals were assigned to one of the three treatment conditions.

Measures: Assessments were conducted at baseline, one month after treatment, and at one and three years after treatment. Assessments consisted of previously used measures: parent reports of behavior problems (CBCL, ECBI Intensity and Total Problems); mother bi-weekly reports of children's negative and prosocial behaviors (PDR); mother reports of discipline (PDR, including spanking, time out, and low rate events such as fire-starting and running away); parenting stress level (using the Parenting Stress Index, PSI, to assess parents' depression, competence, isolation, spouse support, health, attachment, and restricted role); blinded observations of parent-child interactions in the home (assessing parent praise, critical statements, no-opportunity commands, and affect, as well as children's total deviance and noncompliance); and teachers' reports of child behavior problems at school (using the PBQ).

Outcomes
Baseline Equivalence: At baseline, there were no significant differences between groups in demographic or family background characteristics.

Posttest: At post-treatment, all measures of parent and children's behaviors were significant when the combined treatment groups were compared to the control group. For mother reports of children's behavior problems, BASIC mothers reported significant changes according to the CBCL and ECBI, and BASIC fathers reported significant changes according to the CBCL and ECBI Intensity. IVM mothers reported improvement on the ECBI and CBCL measures, while IVM fathers reported improvement on only the CBCL score. Last, GD mothers reported improvement on the ECBI Intensity and Total Problem Behaviors, while fathers reported no changes. Only BASIC mothers reported improvement on the PSI measure of parenting stress. According to mothers' observations of children's problem behaviors, both BASIC and IVM mothers reported improvement on 4 of 5 behaviors (negative behaviors, low-rate events, spanking, and time out for BASIC children, and positive and negative behaviors, spanking and time out for IVM children), while GD mothers noted improvement in negative behaviors, spanking, and low-rate events. Teachers of children whose parents were in the BASIC and GD groups reported improvements in children's behavior at school, with no changes reported for IVM children. Similarly, observers found that only the children whose parents were in the GD and BASIC groups improved on measures of child deviance displayed during parent-child interactions. Last, observer reports of mother behaviors during parent-child interactions revealed significant effects for 4 of 4 mother behaviors for the BASIC group, 2 of 4 for the IVM group (mother criticisms and positive affect), and 3 of 4 for the GD group (positive affect, commands, and criticisms). Similarly, observer reports of father behaviors demonstrated significant improvements in 2 of 4 BASIC father behaviors (criticisms and praise), 2 of 4 IVM father behaviors (commands and criticisms), and 1 of 4 GD father behaviors (positive affect). In summary, each treatment appeared to be effective in improving parent and child behaviors, and relatively few differences were noted between treatment groups on most outcome measures. Where changes were found, they generally favored the BASIC treatment. For example, the BASIC condition was the only group to show a significant reduction in mothers' reports of parenting stress, a reduction in the fathers' reports of the intensity of child behavior problems (on the ECBI), and an increase in both parents' praise statements. Surprisingly, given their lack of direct facilitator contact and lack of group support, the IVM treatment also showed many significant differences when compared to the control group, and was at least as effective as the GD treatment; however, cost-effectiveness was the major advantage of the IVM treatment.

Long-term: At the one-year follow-up, 94 of 114 parents (93.1 percent) were assessed (94 mothers and 60 fathers). All significant behavioral changes reported at post-treatment were maintained. When analyses compared baseline and one-year follow-up scores, all treatment groups reported fewer child behavior problems according to the CBCL and ECBI, and there was a trend for BASIC mothers and IVM fathers to report further reductions in these measures since post assessment. Comparing baseline and one-year scores, all treatment groups also showed significant improvement in child negative and positive behaviors, less spanking, and reduced parent stress levels. In addition, observer ratings indicated improvements in parent-child interactions, with all three groups of mothers demonstrating increased praise and positive affect and those in the BASIC group displaying fewer critical statements since baseline. Changes in father behaviors were similar, with reductions for all three treatment groups in commands and criticisms, improvement in praise for BASIC and IVM fathers, and improvement in positive affect for BASIC and GD fathers. While there were no differences between any of the treatment and control groups according to teacher reports of children's problem behaviors, observer ratings of children's behaviors demonstrated improvement in all three groups for noncompliance and deviance when interacting with fathers, and noncompliance and deviance for BASIC and IVM children when interacting with mothers. As in the post assessment analyses, there were few overall differences between groups, but consumer satisfaction ratings indicated that the BASIC treatment was superior. With each of the treatment programs, 70 percent of the sample showed clinically significant improvements to within normal ranges.

Three-year follow-up information was obtained from 83 mothers (82 percent) and 51 fathers (73 percent), and attrition analyses indicated no differences between groups in demographic characteristics, reports of child adjustment, parent psychological status, or life stressors. Comparison of the three-year follow-up and baseline scores indicated that parents in all three treatment groups reported significantly reduced total behavior problems and increased child social competence on the CBCL. Overall, BASIC was somewhat superior in producing long-term results, with BASIC fathers reporting significantly lower CBCL total behavior problems and externalizing scores than either GD or IVM fathers. Similarly, analyses comparing one-year and three-year follow-up data revealed a significant deterioration in GD parents' reports of children's externalizing scores and in IVM mothers' reports of total problem behaviors, whereas BASIC scores remained stable. No other three-year results are available.

Assessment of the clinical significance of the program revealed that 44 mothers (53.7 percent) and 35 fathers (74.5 percent) reported their children as having CBCL scores in the normal range at the three-year follow-up. There were also improvements according to teacher ratings on the PBQ. At baseline, teachers reported that 47 (61 percent) children were behaving in the normal range and 30 children (39 percent) were deviant. At the three-year follow-up, teachers rated 61 children (73.5 percent) as normal and 22 (26.5 percent) as deviant.

Of those children who did not reach clinically normal scores ("non-responders"), significantly more had mothers who reported lower incomes, depression, and alcoholism in their immediate families. In addition, significantly more nonresponders (81.8 percent) than responders had mothers who were single or divorced. Thus, it appears that marital distress may be largely influential in children's response to treatment.

Study 4

Evaluation Methodology

Design: The sample included clinic-referred parents representing a wide socioeconomic range, and 43 children (34 boys and 9 girls) aged three to eight who had clinically significant levels of behavioral problems according to the ECBI. Parents were assigned to one of three groups:

• Individually self-administered videotape modeling program (IVM) (n=17 mothers, 9
  fathers)
• IVM plus facilitator consultation (IVMC) (n=14 mothers, 7 fathers)
• Waiting-list control group (n=12 mothers, 7 fathers)

Parents in the IVM condition received the self-administered program described in Study Three, with parents attending an average of 10, weekly, one-hour sessions. The parents in the IVMC condition received the same intervention and were also offered two, two-hour consultations, one half-way through the program and one at the end of the program, which they could use to review any aspects of the program they did not understand. Parents in the control group were assigned to one of the two treatment conditions after post assessments were completed.

Measures: Assessments were conducted at baseline, one month after treatment, and one year after treatment. Assessments consisted of parent reports of children's behavior problems (CBCL, ECBI Intensity); parent personal adjustment (using the PSI to measure parent depression, competence, isolation, spousal support, health and restricted role); parent bi-weekly reports of children's prosocial and negative behaviors and spanking (according to the PDR); blinded observations of parent-child interactions in the home (assessing parent praise, critical statements, no-opportunity commands, and affect, as well as children's total deviance); and teachers' reports of child maladjustment at school (using the PBQ). Baseline analyses revealed no significant differences between groups according to demographic or family background characteristics.

Outcomes
Posttest: At the one-month posttest, mothers in the IVM treatment group reported significantly fewer child behavior problems according to the ECBI intensity score, reduced stress levels, and less use of spanking compared to those in the control group, while IVMC mothers reported improvement in stress levels and spanking. Neither group indicated changes in children's behavior according to the CBCL, in mother observations of children's positive or negative behaviors according to the PDR, or in teacher reports of children's behavior at school. Home observations indicated some improvement during mother-child interactions. Compared to the control group, IVM mothers exhibited more praise and positive affect, IVMC mothers displayed more positive affect (with no significant changes in commands or criticisms), and IVMC children were observed to be less deviant. There were relatively few differences on the outcome measures between the two treatment conditions, but IVMC children were observed to be significantly less deviant than the children in the IVM group, suggesting that combined treatment was superior.

A follow-up assessment was conducted on 100 parents from all three conditions (including the waiting-list control group after they received treatment and the parents from the IVM group in Study 3) to determine maintenance of effects and characteristics of families who benefited from this approach to training. At the one-year follow-up, 39 percent of mothers, 43 percent of fathers and 41 percent of teachers continued to report children's behavior problems in the abnormal range. Single-parent status, depression, high negative life stress, and low SES status were significantly correlated with mother and teacher reports of child conduct problems as well as observations of child deviant behavior. These findings indicate that the self-administered program should not be the sole treatment used with highly stressed families.

Study 5

Evaluation Methodology

Design: Seven Head Start centers were included in the study and were randomly assigned to two groups:

• An experimental group in which parents, teachers, and family service workers participated in the intervention (BASIC and some teacher training), in addition to receiving the regular center-based Head Start program (n=296).
• A control group in which parents, teachers, and family service workers participated in the regular center-based Head Start program (n=130).

In the first year, 3 Head Start centers were randomly assigned to the experimental group, while two centers were assigned to the control condition. In the second year, the original experimental group received an additional year of the intervention; the original two control group centers were assigned to the intervention group; and the remaining two centers were assigned to the control condition (and would receive the intervention in the third year).

Families in the experimental condition were invited by Head Start staff (family service workers) to participate in the intervention, which was similar to the BASIC program described in previous evaluations (although slightly abbreviated). Parents in the intervention group received nine weekly, two-hour group sessions led by trained facilitators who were family service workers employed by Head Start. In the second year of the study, four parents who had emerged as group facilitators in the first year were trained as co-facilitators. All Head Start family service workers (teachers and aids) assigned to the intervention group participated in a two-day workshop to familiarize them with the BASIC program so that their classroom management style would be consistent with the strategies parents were learning. The teacher workshop also included instruction on increasing parents' involvement in the Head Start program and strengthening communication and interactions among teachers, family service workers, and parents. Those assigned to the control condition received the regular Head Start services, which included topics such as stress management, nutrition, and self-care.

Sample Characteristics: Families generally faced multiple risk factors, including low income (85% receiving welfare), single-parent status (55%), mother depression (42%), and substance abuse problems (28%). More than half of the children (53%) in the sample were boys. Approximately 37 percent represented minority groups, including Asian, Hispanic and African American families.

Measures: Assessments were conducted at baseline, immediate post-treatment, and one year after the intervention and included parent reports of children's behavior problems and social competence (CBCL, ECBI) and of their discipline approaches, including harsh discipline, inconsistency, limit-setting, and positive reinforcement (according to the Daily Discipline Interview, DDI). Blinded home observations of parent-child interactions were also conducted, assessing positive parenting techniques (including praise and positive affect), critical statements, commands, and non-verbal affect (valence), as well as children's deviance/noncompliance, negative affect/valence, misbehavior, poor conduct, positive affect, and positive behaviors. Teachers reported on children's social competence and externalizing behavior (using the Teacher Report Form, TRF); however, at the one-year follow-up, only a random sample was used (including teachers of 90 intervention and 46 control children). Parents and teachers both reported the frequency of parent's involvement in children's activities at school and their communication with teachers.

Analysis: MANOVAs were conducted for each set of dependent variables within each domain (parenting competencies, parent school involvement, and child social competencies and conduct problems). When the group by time interaction was significant, individual dependent variables were examined using univariate 2 (group) x 2 (time) ANOVAs. When the group x time interaction was significant, paired t-tests examined changes from pre- to post-assessment for each group separately.

Outcomes
Baseline Equivalence and Attrition: 426 families (296 experimental and 130 control) completed baseline assessments, 394 families (264 intervention and 130 control) completed immediate post assessments, and 296 families (69% from baseline) completed assessments at the one-year follow-up (189 intervention and 107 control). At baseline, the two groups in the original sample (n=542), as well as the sample that completed baseline assessment (n=426) were not shown to differ significantly on any risk factors, except that the control group included more minority children (47 percent versus 32 percent). There continued to be no significant differences between attriters and completers in risk factors other than ethnicity at the one-year follow-up in both samples. There were differences between the experimental and control groups on 4 of 29 baseline variables (mother report of limit-setting, teacher report of parent contract with school, and observations of child deviance and negative affect).

Posttest: At post assessment, observer reports indicated that intervention mothers improved on all four parent behaviors during interactions with their children, as mothers made significantly fewer critical remarks and commands, used less harsh discipline, and were more nurturing, reinforcing, and competent compared to mothers in the control group (who remained stable). Intervention mothers reported that their discipline strategies also improved, as they were more consistent, used fewer physical and verbally negative discipline techniques, and were more appropriate in their limit-setting techniques. In turn, the children of mothers in the intervention group exhibited significantly less misbehavior, noncompliance, deviance, and negative affect, and more positive affect, while the control children's behavior remained unchanged. Similarly, teachers reported that the intervention children showed increased social competence, while the control children remained stable. Although mothers did not report increased contact or satisfaction with their children's school, intervention teachers reported significant increases in parents' involvement and contact with school, whereas control teacher's reports remained stable.

Long-Term: One year later, when children were in kindergarten, improvements in the intervention mothers' parenting skills and in their children's affect and behavior were similar to those found at post assessment. Compared to their baseline scores, mothers' reports at one-year showed significant improvements in their discipline techniques, including decreases in harsh discipline style and inconsistency and increases in appropriate limit setting, versus no changes for the control group. Observers also noted significant differences between groups for all four mother behavior variables, with BASIC mothers demonstrating increases in positive affect, praise and physical positive behaviors; and decreases in negative affect and harsh or critical discipline styles. Control mothers showed no changes in these behaviors. Regarding child behavior, mother reports showed no change in ECBI or CBCL measures of child adjustment for either group at the one-year assessment, which was also true at post intervention, but observer reports indicated that BASIC children improved on four of six behaviors, including increases in positive affect and positive behaviors, and decreases in negative affect and misbehavior, compared to no changes for control children. There were no differences between groups according to a small sub-sample of kindergarten teacher reports of children's behaviors.

At post assessment, a test of clinical significance indicated that 69 percent of the high-risk mothers in the BASIC intervention showed a 30 percent reduction in critical statements, compared to 52 percent of the control mothers. Also, 73 percent of children showed at least a 30 percent reduction in negative and noncompliant behaviors at home, compared to 55 percent of the control group. However, neither of these differences remained significant at the one-year follow-up.

Study 6

Evaluation Methodology

Design: The sample was comprised of 78 clinic-referred families (half of whom were self-referred and half professionally-referred) and their 3 to 8 year-old children (58 boys and 20 girls). The majority of parents were Caucasian and low- to middle-class. Families were included in the study if children met the DSM-IV criteria for Oppositional Defiant Disorder and/or Conduct Disorder. They were then randomly assigned to one of two conditions:

• BASIC training (n=39 mothers and 30 fathers)
• BASIC + ADVANCE training (n=37 mothers and 27 fathers)

All families received the BASIC training, which was the standard 12-week, two-hour program described previously. After BASIC ended, 38 families were randomly assigned to the ADVANCE training condition, in which parents received 14 weeks of two-hour group training sessions focused on improving communication skills (replacing destructive styles of communication with effective ones such as active listening and expressive speaking), personal self-control (i.e., coping with anger, stress and depression and substituting positive self-talk for depressive, blaming self-talk), and problem-solving (with spouses, family members or employers); ways of giving and receiving support; and fostering children's problem-solving skills.

Measures: Families were assessed at baseline, immediately following BASIC treatment, and immediately following the ADVANCE intervention. Assessments included parent reports of child adjustment (CBCL and ECBI), parent competency (using the PSI to measure parent depression, competence, isolation, spousal support, health and restricted role), anger/aggression and marital adjustment. Observers blind to the study rated parent-child interactions at home, assessing parents praise, critical statements, and non-intrusive statements, and children's deviance (including whining, crying, physical negative behavior, yelling, and noncompliance). In addition, couples were videotaped in a laboratory discussing problems, and their interactions were coded by blinded observers completing the PS-I CARE assessment, rating problem-solving techniques, communication skills, and marital collaboration or engagement. Children were also assessed via a social problem-solving test.

Outcomes
Baseline Equivalence: Baseline analyses revealed no significant differences between the two groups in demographic characteristics or child behavior problems.

Posttest: When analyses were conducted after the ADVANCE program, child adjustment (according to the ECBI), social competence (using the CBCL), and deviance (according to observations) significantly improved for both treatment groups, compared to their baseline scores. Parent-child interactions also significantly improved for both treatment groups, with parents demonstrating fewer critical statements and more praise and reflective statements. In addition, both groups of parents reported decreases in distress and depression. There were some differences between groups at post assessment. Compared to the BASIC-only group, ADVANCE children showed significant increases in the total number of solutions generated during problem solving. Blinded observations of parents' marital interactions indicated significant improvements in ADVANCE parents' communication and problem solving skills, and collaboration, compared to parents who received only the BASIC program. However, there were no differences between ADVANCE and BASIC-only groups in parent reports of marital satisfaction, anger or stress levels; parent reports of children's behaviors (according to the CBCL and ECBI), or observations of child deviance.

In terms of clinical improvement, all children were in the abnormal ECBI range at baseline, but 41 (53.2 percent) moved into the normal range and 36 (46.8 percent) remained abnormal after the ADVANCE training. Similar results were found for the CBCL. Whereas 49 mothers (64 percent) reported abnormal CBCL behavior problems at baseline, 26 (53.1 percent) showed a change into the normal range at follow-up.

Results also indicated that fathers' improvement in marital communication skills significantly reduced their number of criticisms in parent-child interactions, and fathers' improved marital satisfaction was related to children's improvements in social skills. These results indicate that improving families' personal distress and interpersonal issues can greatly affect both parents' and children's skills. In addition, the fact that only one family dropped out of the ADVANCE program attests to its perceived usefulness by families. Moreover, all the families attended more than two-thirds of the ADVANCE sessions, with the majority attending more than 90 percent of the program.

Study 7

Evaluation Methodology

Design: Using a randomized control design, 14 Head Start centers (36 classrooms) were randomly assigned to two groups:

• An experimental condition in which parents, teachers, and family service workers participated in the prevention programs (BASIC + ADVANCE + SCHOOL + TEACHER) (n=191, 23 classrooms from 9 centers)
• A CONTROL condition in which parents, teachers, and family service workers participated in the regular center-based Head Start program (n=81, 13 classrooms from 5 centers).

Centers were chosen based on their willingness to participate in the study and to be randomly assigned to either intervention or control. Parents were recruited into the experimental condition by Head Start family service workers and teachers during summer and fall orientation. Originally, 328 families indicated interest in participating in the study (225 experimental and 103 control). Thirty-four families (15%) from the experimental centers and 22 families (21%) from the control centers did not complete spring posts assessments. This left 272 families (191 experimental and 81 control).

The experimental condition included four components: the standard BASIC parent training program, an abbreviated version of the ADVANCE and academic skills training (SCHOOL) programs, and a teacher training program (TEACHER). The first three components were held for parents of children attending Head Start and consisted of 16 weeks of two-hour group training sessions led by Head Start family service workers. The twelve-week standard BASIC program was offered to parents in the Head Start year, and a four-week abbreviated version of ADVANCE and SCHOOL was offered in the kindergarten year. Whereas the ADVANCE component targeted parent problem-solving skills and strengthening their relationships with partners, the SCHOOL program promoted parents' increased communication with teachers; techniques for encouraging children's reading, academic and problem-solving skills; arranging child "play dates;" and coaching positive peer play skills. The final component, the TEACHER training program, focused on improving teachers' classroom management and discipline skills, building relationships with students and parents, and helping instructors promote students' social and emotional competence. All Head Start teachers and aides in the intervention group received six monthly workshops sequenced over the first year of the intervention. Those in the control centers received the usual Head Start services, which included parent education in stress management, nutrition, self-care, and dental care.

Sample Characteristics: This study included a sample of 272 Head Start mothers, 272 four-year old children and 61 teachers. As in Study 5, many of the families in this study faced multiple risk factors for child conduct problems, including low education, low income, parent substance abuse, depression, and spouse or child abuse. The sample of children was 54% male with an average age of 55 months. 63% of the families represented ethnic minority groups, including predominantly African American, Hispanic, and Asian American ethnicities, and 52% of the families were single-parent households.

Measures: Assessments were conducted at baseline, post intervention, one year, and two years post intervention (through grade one). They included parent reports of children's behavior at home (using the CBCL and ECBI); teacher reports of children's conduct problems at school (using the Social Competence and Behavior Evaluation, SCBE, to measure social competence, emotional expression and adjustment difficulties; and the teacher ADHD checklist to assess the presence of attention deficit hyperactivity disorder); and blinded observations at home and at school (assessing children's noncompliance and deviance at home, and conduct problems and the amount of time engaged in activities at school). Construct scores combining observational and report data were calculated for negative (harsh discipline, critical parenting) and positive (monitoring, praise, consistent discipline, positive affect) parenting styles; parent-teacher bonding (primarily assessing increased communication between parents and teachers and parent involvement in school activities); child conduct problems at home (noncompliant, aggressive, and disruptive behaviors) and at school (including children's aggression, noncompliance, ADHD symptoms, engagement and social competence); and observer reports of teacher classroom management style (including teachers' positive techniques and harsh discipline, and classroom atmosphere).

Analysis: Analyses of covariance on the post-treatment construct scores using the corresponding pretest construct score as a covariate, was conducted. When these ANCOVAs showed significant group effect for the construct, then adjusted posttreatment means were examined to determine if the experimental group improved significantly more than the control group.

OUTCOMES:

Fidelity Monitoring: Monthly supervision was provided for parent group leaders coupled with close monitoring, direct observations and detailed training sessions and training manuals. 100% of the group leaders reported discussing all videotaped vignettes assigned and assigning all homework sessions described. For teachers, workshop checklists, standardized handouts, and videotaping of all teacher training was used to assure the integrity of the intervention.

Baseline Equivalence and Attrition: Analysis compared baseline equivalence of the experimental and control groups of the sample who completed both pre and post assessments (n=272). Analyses revealed several differences between groups at baseline, with the experimental group reporting significantly more risk factors than control group members. Specifically, the intervention group contained fewer boys, more minority members, higher rates of mother depression and stressful life events, and lower annual incomes. There were also significant differences in two of the six construct domains. Intervention mothers had higher negative parenting scores and lower positive parenting and bonding scores. Because of the differences on the two construct scores, analyses of covariance on the post-treatment construct scores using the corresponding pretest construct score as a covariate, was used.

Attrition analyses done at the one-year follow-up indicated no significant differences in dropouts between groups in terms of demographic and baseline risk factors. Differences between drops and nondrops on risk, demographic, and construct scores were examined separately for intervention and control groups. There were no differences on any variable in the intervention group, and only one difference between drops and nondrops in the control group. More families from the control group who reported children with behavior problems dropped out between pre- and post-assessments.

Posttest: At post assessment, intervention mothers had significantly lower scores on the negative parenting construct and higher scores on the positive parenting construct compared to control mothers. Parent-teacher bonding was marginally significant in the full sample and significantly higher in a subsample of intervention mothers who attended six or more intervention sessions than for control mothers. In terms of children's behaviors, intervention children showed significant improvements on the conduct problems school construct (including lower reports of hyperactivity and antisocial behaviors and more social competence), marginally significant results on the home construct, and children of mothers who attended six or more sessions showed significant improvement on the conduct problems at home construct, compared with control children. In addition, children who were in the "highest risk" category (based on observations of high rates of non-compliant and aggressive behavior) at baseline showed greater clinically significant reductions in aggressive and noncompliant behaviors than high-risk children in the control group. Teachers' behavior also improved, with instructors in the intervention group having significantly higher scores on the classroom management construct than control teachers.

Long-term: When baseline scores were compared to the one-year follow-up measures, some intervention effects remained. Only four of the original six construct scores were examined (negative and positive parenting, parent-teacher bonding, and child behavior problems at home). Observer reports of teachers' classroom management and children's behavior at school were not available. There were marginally significant effects in the predicted direction for the Positive Parenting and the Child Conduct Problems at Home constructs. Parents who attended more than nine sessions continued to have lower scores on the negative parenting construct and higher scores on the positive parenting construct. Clinically significant findings on children's conduct problems found at post assessment continued to be present one year later. This difference was significant with the whole sample, as well as with the children of parents classified as "attenders". Parent-teacher bonding was significant at the one-year follow-up, but in the reverse direction, with intervention mothers showing less bonding with teachers than control mothers. Analyses of the two-year follow-up results are currently being conducted.

Reid, M. J., & Webster-Stratton, C. (2001). Parent training with low income, minority parents: A comparison of treatment response in African American, Asian American, Caucasian and Hispanic mothers. Seattle, WA: University of Washington.

Effectiveness of the BASIC Program for Ethnic Minorities. Data from Studies Five and Seven (those using Head Start samples) were combined in order to compare the effectiveness of intervention according to ethnic group, with analyses differentiating between Caucasian, African American, Asian American and Hispanic participants. Results indicated that significant changes occurred regardless of the ethnicity of the family and all ethnic groups had high consumer satisfaction for the program. However, some differences emerged when scores on the consumer satisfaction survey were rank-ordered, with Caucasian mothers consistently rating the program somewhat more critically than the other three groups. In terms of attendance, Asian mothers had the highest rates and Caucasian parents attended significantly fewer sessions than Hispanic and African American parents. Further details about this study can be found under Study 31.

Study 8

Evaluation Methodology

Design: The Dina Dinosaur curriculum was evaluated in a randomized trial with 97 clinic-referred children (72 boys and 25 girls) ages four to seven. Children and their parents (95 mothers and 71 fathers) were randomly assigned to one of four groups:

• Child training only (Dina Dinosaur Curriculum) (n=27)
• Parent training only (BASIC + ADVANCE) (n=26)
• Combined parent and child training intervention (n=22)
• Waiting-list control group (n=22)

The Dina Dinosaur curriculum consisted of 18 weeks of two-hour sessions focusing on empathy training, problem-solving training, anger control, friendship skills, communication skills and overcoming difficulties at school. Children met in small groups (six children) which allowed participants to collaborate, share ideas, and develop bonds to each other. As in the parenting programs, videotapes were used to foster discussion, teach problem-solving skills, and model prosocial behaviors. Because young children are vulnerable to distraction, the intervention incorporated group practice, role plays, stories, puppet plays, home assignments, and clubs and incentives to strengthen motivation, hold children's attention, and reinforce key concepts. Facilitators used life-size puppets to model appropriate behavior and thinking processes. Parents and teachers were involved by helping with homework assignments and receiving regular letters outlining key concepts being taught and suggesting home or classroom reinforcement of particular behaviors.

Those assigned to the parent training condition received the BASIC and ADVANCE programs, including 22 weekly sessions that reviewed the concepts described in the earlier evaluations. Parents in the combined intervention received the BASIC and ADVANCE programs while their children participated in the Dina Dinosaur curriculum. Families in the waiting list control condition were randomly assigned to one of the three intervention conditions after eight or nine months.

Sample Characteristics: Families were primarily Caucasian (85%), 68% were two-parent, and the majority were low- to middle-income. Children were selected for inclusion based on meeting the DSM-IV criteria for ODD and/or CD.

Measures: Families were assessed at baseline, two months after intervention and one year after treatment. Assessments included parent reports of children's negative and positive behaviors at home (CBCL, ECBI, PDR), teacher reports of child behavior at school (PBQ), and child problem-solving testing (PPS-I CARE). Blinded observations were also conducted, including observations of parent-child interactions at home (assessing parent commands and criticisms, praise, positive affect, and negative valence, as well as children's deviance, positive affect/warmth), laboratory observations of children playing with a friend (using the PPS-I CARE to note problem-solving and conflict resolution skills), and laboratory observations of parents discussing their child's behavior (assessing parent's problem-solving and collaboration skills on the PS-I CARE). There were no significant differences between groups in demographic or family background characteristics at baseline, nor in parents or teacher reports of child misbehavior.

Analysis: A four-group analysis of covariance (ANCOVA) was used to evaluate treatment effects for each dependent variable, using pretreatment performance as the covariate. Significant effects were followed by preplanned contrasts (t statistics), comparing each treatment group.

Outcomes

Baseline Equivalence: There were no significant differences among the four groups on demographic or family background variables, nor were there any significant differences for parent reports of child behavior (CBCL, ECBI, PSI), teacher reports of child behavior, or observations of child behavior in the home or laboratory.

Posttest: At post treatment, all treatment groups improved on many of the parent and child behavioral variables, relative to the control group. More specifically, the treatment groups were superior to the control group according to parent reports of child adjustment (CBCL, ECBI), mother observations of children's positive and negative behaviors (according to the PDR), observations of children's conflict management skills during peer interactions, and parent stress. In contrast, there were no significant differences between treatment and control groups according to teacher reports of problem behavior or in observer reports of child deviance or positive affect displayed during parent-child interactions.

When the individual groups were compared to the control group and to each other, the child training program was found to have some unique beneficial effects on children's behavior. Children receiving the Dina Dinosaur curriculum only demonstrated significant improvements in observed conflict management skills when interacting with peers compared to those receiving parent training only, and those receiving child training (either alone or with parent training) demonstrated better problem-solving skills than the parent training only group. Those receiving parent training (either alone or with child training) demonstrated significantly more positive parenting behaviors and parent collaboration compared to control families and those receiving child training only. For example, those receiving parent training only improved on all four observed mother behaviors (including commands/criticisms, positive affect, praise, and negative valence), while those in the combined condition improved on three of four behaviors and those in the child training only condition improved on only one variable. Similarly, fathers in the parent training only condition improved on 3 of 4 variables compared to the control group, while those in the combined and child training only conditions had no improvements. Those receiving the combined parent and child training interventions demonstrated more mother praise compared to those in the control group and the child training only intervention.

All three conditions, relative to the control group, demonstrated improvements in child behavior problems according to both mother and father reports. When the three treatment groups were compared with each other, those receiving parent training only had better mother reports of children's problem behavior (according to the CBCL) and reduced stress compared to those receiving only child training. Mothers in all three conditions observed significantly fewer targeted negative behaviors and more prosocial behaviors at home than control mothers, but there were no significant differences among the three treatment conditions.

One Year (maintenance effects, no control group): Results obtained after one year demonstrate continued improvements in parent and child behaviors since post assessment, as well as the emergence of several additional significant findings. Comparing one-year measures to baseline scores, all three treatment groups had significantly fewer child behavior problems (CBCL, ECBI and PDR), better child problem-solving skills, less spanking, improved parent behavior during parent-child interactions (including fewer criticisms and commands, less negative valence, and more positive affect), and lower parenting stress levels. In addition, observers rated intervention children as demonstrating less deviance and more positive affect and physical warmth at home, compared to their baseline scores, a finding not found at post treatment. While teacher reports at the one-year follow-up show intervention children having increased behavior problems since the post assessment, those in the abnormal range on the PBQ at baseline demonstrated significantly improved behavior at the one-year follow-up, for all three treatment groups.

Overall, the combined parent and child training group appeared to have the most positive effects in the broadest array of behaviors. It was superior to the child training only intervention in improving parent behaviors in their interactions with children (particularly in mother praise and parent collaboration) and in reducing children's problem behaviors. In addition, it produced better results than the parent training only program in children's social problem-solving skills.

Moreover, analyses of clinically significant improvements (measured by a reduction in total child deviant behaviors at home) revealed that the combined parent and child intervention showed the most sustained effects in child behavior at the one-year follow-up. Children in the combined intervention group showed a 95 percent decrease in deviant behaviors since baseline, compared to reduction of 74 percent for those in the child only condition and 60 percent for those in the parent only condition.

Study 9

Evaluation Methodology

Design: Participants were recruited from families requesting treatment for their child's conduct problems at the University of Washington Parenting Clinic. One-third of the families were self-referred, the others were referred by professionals (teachers, physicians) in the community. Families entered the study in three 50-55 family cohorts in the fall of 1995, 1996 and 1997. Families of 159 4- to 8-year-old children with oppositional defiant disorder were randomly assigned to one of six conditions:

• parent training (PT; n=31),
• child training (CT; n=30),
• parent training plus teacher training (PT+TT; n=24),
• child training plus teacher training (CT+TT; n=23),
• parent/child/teacher training (PT+CT+TT; n=25),
• waitlist control group (n=26).

The waitlist control group received no treatment and had no contact with the research team or the team's therapists during the 8-9 month waitlist period. After the post-intervention assessment, control group families were offered the parent training program.

Of the entire sample that completed baseline assessments, four families dropped out before beginning treatment and refused to participate in post-assessments. The rest of the sample completed post-assessments regardless of how many sessions attended. There was no significant difference in drop-out rate by treatment condition. No teachers dropped out of the TT condition.

Sample Characteristics: The children in the study were 90% boys, 79% white, with an average age of 5 years and 11 months. The majority (74%) of the parents were married.

Measures: Baseline assessments were conducted in early Fall, post-intervention assessments were conducted approximately six months later in the Spring, for all six conditions in each of the three cohorts. Follow-up assessments were conducted one year later in the Spring for five of the six conditions, because the control group was treated after post-intervention assessments. Composite measures were calculated by adding multiple instruments in five areas:

Positive and Negative Parenting - Two parent-report measures (Parenting Practices Interview and Dyadic Parent-Child Interactive Coding System-Revised) and two observational measures (Coder Impressions Inventory and Parent Daily Discipline Inventory) were used to assess parenting styles and skills. The Parenting Practices Interview (PPI) has internal consistency alpha coefficients of .71 for harsh discipline and .66 for supportive parenting. The Dyadic Parent-Child Interactive Coding System-Revised (DPICS-R), which looks at positive parenting and critical statements, has a Cronbach's alpha of .78. The Coder Impressions Inventory-Parent (CII-P) is completed following a half-hour parent-child observation where parents are measured on 12 harsh-critical items (Cronbach's alpha = .89; interclass consistency coefficient = .54), 13 nurturing-supportive items (Cronbach's alpha = .88; interclass consistency coefficient = .67) and a rating on a 5-point scale of the observers perception of the degree to which a family needs help (interclass consistency coefficient = .64).

Child Conduct Problems at Home - Composite scores were calculated using one parent-report variable (the Eyberg Child Behavior Inventory - ECBI; Cronbach's alpha = .92) and four in-home observations including the DPICS-R coding system and the Coder Impressions Inventory for Children (CII-C).

Child Conduct Problems at School - The composite scores included two teacher report variables (scales from Teacher Assessment of School Behavior - TASB - Cronbach's alpha ranges between .62 to .91 ; and pertinent scales from the Teacher Rating of Perceived Competence Scale for Young Children - PCSC - reliability ranges between .70 to .90 for the subscales) and two summary scores from independent observations of classrooms.

Child Social Competence with Peers - The composite score includes two teacher-report variables (relevant scales from TASB and PCSC), one classroom observation (30-minute observations calculating conduct problems; Cronbach's alpha = .71) and one laboratory observation of the child with a peer (observation focused on Inappropriate Play scale from DPIS; internal consistency = .88).

Negative Classroom Management - A composite score of negative classroom management and atmosphere was computed via direct observation for each teacher using five variables: total teacher criticism; observation of classroom atmosphere (CAM, Cronbach's alpha ranging from .94 to .95) measured on a 10-item questionnaire; and three items from the Coder Impression Inventory - Teacher, which rates harsh techniques, nurturing techniques and percentage of time teacher is inappropriate.

Analysis: The procedures were consistent with an intent-to-treat analysis. The four families who dropped out before treatment began were followed up, but they refused to complete postassessments. Treatment effects for each measure were analyzed using six-group analysis of covariance with pretest scores as covariates for corresponding posttest scores. Planned comparisons contrasting each treatment condition with the control condition were conducted and then the addition of TT to CT and to PT was tested against CT alone and PT alone. Lastly, PT+CT+TT condition was tested against two factor conditions (CT+TT and PT+TT). Missing data were handled at two levels: an individual summary score was only computed if at least 60% of the items that made up the scale were present. Composite scores were also only computed if at least 60% of the summary scores in the composite were present. Cases were excluded from analysis on that composite if the composite score was missing at one of the time points. This resulted in sample sizes that fluctuated differently for different composite scores and treatment groups. The CT only condition had the most fluctuation at follow up for the father's positive parenting (n dropped from 30 to 22 at follow up).

Outcomes

Implementation Fidelity: Therapists conducting the parent or child group co-led their first group with a supervisor, followed a treatment manual for each session and documented fidelity to the program through a weekly protocol checklist of standards to be covered in each session. Group sessions were videotaped for feedback and analysis at weekly supervision meetings. Therapists received ongoing supervision, feedback and training throughout the study. Supervisors also randomly selected videotapes for integrity checks and the study reports that the analysis of checklists indicates high treatment integrity. Teacher training sessions were manualized, videotaped and the tapes were reviewed to ensure that training procedures did not vary across cohorts of teachers.

Attendance at the children and parent groups ranged between 90% and 100% for at least 15 sessions among the CT, CT+TT, PT, PT+TT and CT+PT+TT groups. All teachers attended the four days of training and all attended at least two meetings to work on individualized behavior plans for the child.

Baseline Equivalence and Differential Attrition: The study used ANOVA and chi-square analyses to compare baseline equivalence across all six conditions. No significant differences were found at baseline among all groups on demographic or family background variables. Further, there were no significant differences between conditions at baseline on any composite scores. From the entire sample that completed baseline assessments, only four families dropped out prior to the beginning of the study and refused to participate in postassessments. No information was provided about how these four families differed at baseline from those that continued in the study. At the two-year follow up, 9% of the families in the treatment groups (n=12 families) dropped out. Significantly more families in the CT condition dropped out. The children in these families who dropped out at two-year follow-up had fewer behavior problems at postassessment.

Post-test (Webster-Stratton et al., 2004): At immediate post-test, all treatment conditions resulted in significantly fewer conduct problems for children compared to controls. However, when compared with each other, there was little significant difference between treatment groups.

The following results were found in the five areas measured:

Negative and Positive Parenting: The six-group ANCOVA revealed significant effects for all four composite scores: mother negative parenting (F=7.26, p<.001); father negative parenting (F=6.65, p<.001); mother positive parenting (F=3.29, p<.01); and, father positive parenting (F=2.37, p<.05). Based on pre-planned comparisons across groups, when parent training was involved, most groups had significant differences between treatment group and control groups, but for CT or CT+TT, only a measure of mother's negative parenting was significantly different between treatment group and control group.

Child conduct problems at home: The six-group ANCOVA revealed significant differences among the groups. However, there were no significant differences between the different treatments for child behavior with mother or father.

Child conduct problems at school: The six-group ANCOVA was non-significant for the child negative problems at school composite score, however, preplanned comparisons indicated that all five of the treatment conditions showed significant treatment effects when compared with the control, though there were no differential effects of treatment conditions.

Child social competence with peers: The six-group ANCOVA was non-significant for the child social competence composite score, however, preplanned comparisons indicated that three of the treatment conditions that included CT showed significant treatment effects when compared with the control, though there were no differential effects of treatment conditions.

Teacher Classroom Management: The ANCOVA revealed significant group effects for the teacher classroom management composite score (F=5.39, p<.001). The three conditions that received TT, as well as the CT-only condition, showed significant treatment effects when compared to the control. There were no significant effects for other between-treatment comparisons.

One-year post intervention (maintenance effects, no control group): Because the control group was treated after postassessments, these follow up results were only focused on whether children and parents in the treatment conditions improved or maintained effects in the following eight areas: mother negative parenting; father negative parenting; mother positive parenting; father positive parenting; child conduct at home per father; child conduct at home per mother; child conduct at school; and, child social competence. Mixed design (Time x Condition) ANOVAs were computed for each composite score from postassessment to follow up. Seven of the eight areas were not significant at one-year follow up. Only child conduct at school was significant at one-year follow up (F=3.45; p<.01). However, seven of the eight measures maintained to the 1-year follow-up. School behavior of children in the PT+CT+TT deteriorated from postassessment to follow-up.

Two-years post intervention (Reid et al., 2003): At two-year follow-up, parents were asked if their children received additional services or treatment after completing the program, and what changes in medications their children received. Results indicated no significant differences for any of these variables among treatment conditions. On ratings of conduct problems at home, comparison among treatment groups showed that children in the PT+TT group had significantly better outcomes that in the PT alone condition (chi-sq=5.27; p<.02). On ratings of conduct at school, the number of children showing significant improvement according to teacher reports maintained or improved for all five treatment groups. However, there were no significant differences between treatment groups.

Study 10

Evaluation Methodology

Design: Participants were 105 self-referred families of clinic-referred children diagnosed with early onset conduct problems (age three to eight years) who had contacted the Lakehead Regional Family Mental Health Center in Thunderbay, Ontario. These participants who agreed to participate in the study were randomly assigned to three conditions:

• BASIC program (n = 46 families)
• Eclectic treatment normally offered (n = 46 families)
• Waiting list control group (n = 18 families)

All families were considered "clinical" cases with poor scores on behavior problem scale (ECBI). However, their mean score was lower than that reported by Webster-Stratton in previous studies, suggesting that these children's behavioral problems were less severe.

Families assigned to BASIC received 11 to 14 weeks of treatment in small groups. Treatment was similar to that described earlier, with the addition of occasional home or school visits if necessary to support the parenting program. Facilitators were trained by Dr. Webster-Stratton. Families in the eclectic treatment met with facilitators individually, at the facilitators' invitation. Meetings usually took place at the Center, but sometimes were held at home or in the community. Parents and children could also receive other services offered at the clinic. The therapeutic approaches utilized included ecological, solution-focused, cognitive-behavioral, family systems, and popular press parenting approaches.

Sample Characteristics: Boys were the identified child in 82% of the families. Study parents included 69 married or common-law couples, 38 single mothers, and one single father. Although there were differences between married and single parents, there were no significant differences between the two treatment groups and the control group at pretest on family income, percentage of single-parent families, measures of behavior problems in children, or mothers' self-report of depression.

Measures: Assessment consisted of mother reports of behavior problems on the Eyberg Child Behavior Inventory (ECBI), Child Behavior Checklist (CBCL), and Parent Daily Reports (PDR), as well as teacher reports on the Achenbach Teacher Report (TRF) and the Matson Evaluation of Social Skills with Youngsters (MESSY).

Analysis: Analysis of covariance (ANCOVA), using pretest scores as a covariate, was used to test the effects.

Outcomes

Posttest: At the four-month posttest, BASIC mothers reported significantly fewer child problems on the ECBI intensity and total problem scores, compared to the control group, but no significant differences for the Parent Daily Report or teacher reports. When the eclectic treatment was compared to the control group, only the ECBI problem score was significantly different, favoring the treatment group. Comparisons of the two treatment groups revealed that mothers in the BASIC group reported significantly fewer problems on the ECBI problem score, but no other significant differences were found (although participants rated BASIC treatment more favorably on a consumer satisfaction survey). In terms of clinical significance, 41 percent of the BASIC mothers and 74 percent of the eclectic mothers continued to report problems on the ECBI in the clinical range.

Study 11

Evaluation Methodology

Design: Participants were 152 parents who were recruited into the study by responding to advertising for a free parenting class focused on handling children's tantrums, noncompliance and hyperactivity. Of these, 81 actually volunteered to participate in the study and completed the pretest. After pretest, 25 subjects attrited and 3 were omitted from the analyses, leaving a sample size of 53. Average age of the target child was 6.1 years. Of these 53 families, 37 completed post assessments, and 29 remained at follow-up (four to six months after the program ended). Parents were randomly assigned to three conditions:

• Abbreviated BASIC parent program plus problem-solving group (n = 21)
• Abbreviated BASIC parent program plus extra facilitator discussion (n = 16)
• Waiting-list control group (n = 16)

Both intervention groups received an abbreviated version of BASIC which included viewing four tapes and meeting for ten hours in groups to cover the main topic areas. The problem solving group received six hours of training in problem solving skills, so that in each session one hour was devoted to BASIC and one hour to problem-solving. Those assigned to the facilitator discussion group met with the facilitators for six hours to discuss ways to apply the skills they had learned in BASIC to their children.

Measures: Assessments included parent reports of problems (ECBI and Parent Identified Problems Scale (PIP)) and their own knowledge and use of parenting skills.

Analysis: MANCOVAs and ANCOVAs were used to test for results, using pretest scores, parent education, and family size as covariates.

Outcomes

Differential Attrition: Attrition analyses revealed that those who dropped out of the study were more likely to be less educated, score worse on the Parenting Situation Test, and report more child problem behaviors. Moreover, those who dropped out between posttest and follow-up were more likely to be members of the problem-solving group (seven parents versus the one discussion group member). Otherwise, there were no significant differences between the two treatment groups at the 4-6 month follow-up.

Posttest: Posttest results indicated that both intervention groups improved on the parent situation test and parent punitiveness compared to controls, with no significant differences between intervention groups. With regard to child behaviors, intervention parents reported reductions in the frequency and disruptiveness of the PIP behaviors compared to controls, with no difference in treatment group scores. Parents in the problem-solving groups achieved greater reductions than the control group for the ECBI intensity scores, but both intervention groups showed reductions that placed children's scores in the normal range. Only the problem-solving group showed improvements in their views of their children and of the parental role, compared to the control group.

Long-term: At the four to six month follow-up, comparisons between intervention groups and controls were not possible, as the control group received delayed intervention. Analyses comparing the two intervention groups showed no differences between groups for any of the three outcome domains. When improvements over time for each group were analyzed, treatment effects were found for the parenting situation test and for all three measures of child behavior.

Study 12

Evaluation Methodology

Design: Twenty-four Chicago-area pediatric practices were stratified according to demographic characteristics and randomly assigned to one of three intervention groups. The active intervention was Incredible Years Parent training delivered either by nurses in one condition or clinical psychologists in another. The comparison condition was bibliotherapy. Parents were given a book on Incredible Years and received the same care that would typically be provided in the pediatric clinic, but did not attend treatment sessions.

Study participants were children ages 3.0 - 6.11 years and their parents. Eligible children had met DSM-IV criteria for ODD based on clinical consensus diagnoses, exhibited receptive language at or above the 24 month level, and did not have a DSM-IV diagnosis that superseded the ODD diagnosis (e.g., autism). There were several exclusionary criteria including: externalizing scores below the 90th percentile on the CBCL, being in treatment already, unable to be scheduled for a pretreatment interview, could not be reached after screening, or refused further participation. Families that screened high on externalizing participated in a second stage evaluation in which the pretreatment assessment battery was used to establish the ODD diagnosis. After some exclusions for not meeting diagnostic criteria for ODD, failing the language screen, refusing treatment, or being unavailable for treatment during office hours, 117 children with ODD started the study intervention (49 in the nurse-delivered intervention, 37 in the psychologist treatment group, and 31 in the comparison group).

Of those children entering treatment, 70.1% completed all three assessments (nurse-led 61.2%; psychologist-led 75.7%; and comparison 77.4%).

Sample Characteristics: The sample was 53% male, largely middle-class (60%) with both lower (10%) and upper (30%) classes represented. The average age was 4.6 years. Within the treated sample, 56% had no comorbidities, while 44.4% had at least one, the most common being ADHD (27.4% of the sample).

Measures: The parent-reported Eyberg Child Behavior Inventory (ECBI) and Child Behavior Checklist (CBCL) were used. The Peabody Picture Vocabulary Test was administered as a measure of single-word receptive language skills. The Rochester Adaptive Behavior Inventory was a semi-structured parent-completed interview and includes items on anxiety, mood, and disruptive disorders. Parental responses were reviewed by clinicians. There were also 15 minute interval parent-child interaction observations rating compliance to two types of commands.

Analysis: Analysis was intent to treat as missing data was compensated with values from previous assessments: At post-treatment, 21.4%, 32.7%, and 16.2% of values were moved forward for nurse-led, psychologist-led, and comparison conditions respectively. At follow-up, 15.4%, 10.8%, and 16.1% of values were moved forward, for the same groups respectively. Linear mixed modeling procedures were used to assess treatment-related changes over time, with treatment group, trials, and the group by trials interactions treated as a fixed effect and pediatric practice as a random effect. The combined therapist-led groups were compared with the comparison group, and the two therapist-led groups were compared with one another.

Outcomes

Baseline Equivalence: There were no differences by group for child's age, gender, race, parent's marital status, social class, maternal or paternal education, child's receptive vocabulary, or the number of children exhibiting comorbidity.

Posttest: All three groups improved over time on parent-reported measures of symptoms, however, there were no significant treatment differences among the three groups. There were also no differences on command compliance, child-rearing knowledge, or use of other mental health services.

There was a dose effect, with greater improvement in the two treatment conditions relative to the comparison on both the Eyberg and CBCL externalizing parent-reported scales. On the Eyberg, attending seven or more sessions was better than no treatment. On the CBCL, attending nine or more sessions was better than no treatment.

Study 13

Evaluation Methodology

Design: Study participants were 124 clinic-referred, low-income parents of children diagnosed with ODD/CD who were referred for treatment to mental health clinics in the greater London area. In total, 340 referrals to the study had been received, but only 148 were assessed as eligible. Of these, 24 dropped out prior to assignment, leaving the sample of 124. Dropouts were compared with starters, but did not differ significantly by age, gender, or severity of initial conduct symptoms. Clinics chosen were well run with good reputations, typical of the best practice in NHS. Families were randomly assigned to three conditions:

• BASIC parent program (n = 62)
• Standard treatment (n = 30)
• Waiting list control (n = 32)

There were no significant differences between groups at the start of the trial. The BASIC program was the 12-week package described above and was modified by being dubbed into an English accent. Practitioners in mental health clinics were trained to deliver the treatment. The standard treatment was the service typical of good practice in the NHS. This treatment consisted of the parent and child being seen together by one or two workers for five or six sessions.

Measures: Assessments were conducted at baseline, post treatment (at six months), and at one-year follow-up. Those on the waiting list were reassessed after five or six months, and offered standard treatment. Assessments included parent reports of depression and child behavior problems (Child Behavior Checklist-CBCL, Parent Daily Report-PDR, Strengths and Difficulties Questionnaire-SDQ, Parent Account of Child Symptoms-PACS). Overall rate of reassessment at posttest was 114 out of 124 (92 percent). The dropout rate was 17 percent for the BASIC program and 34 percent for standard treatment (a significant difference).

Outcomes

Posttest: Results indicated that BASIC was more effective than standard treatment for conduct symptoms on all reports of antisocial behavior (PACS Conduct Scale, CBCL Total Score and Externalizing Score, SDQ Total Deviance and Conduct Symptoms, PDR). The improvements made by the BASIC group at posttest were maintained at the one-year follow-up. Compared with waiting list control, the effect was large, and both clinically and statistically significant. Symptoms of hyperactivity improved significantly according to the interview measure but not according to the SDQ. Peer relationship problems did not show significant improvement with either treatment. Mother depression improved significantly with the BASIC program but not with the standard treatment. Parent satisfaction was very high for the BASIC program, and dropout rates were lower than for the standard treatment condition. Cost effectiveness analyses indicated that both treatments cost about the same; however, BASIC had twice the amount of contact hours (15 vs. 7.5), and the effectiveness was more than double the standard treatment using the main measure (PACS). The authors suggested that standard treatment cost 587 pounds per half standard deviation reduction of aggression, whereas BASIC cost 267 pounds.

Study 14

Dr. Laurie Miller and colleagues at New York University have conducted a pilot study with 30 families of children (age two to four years) who have a sibling with a documented history of antisocial behavior (i.e., court-documented delinquent or criminal behavior, or clinical diagnosis of conduct disorder). Preschoolers were randomly assigned to intervention or control groups. The intervention consisted of 50 parent training sessions using the BASIC and ADVANCE programs (1 1/2 hours, 2 times/week) and 50 social skills groups using the Dinosaur School themes, biweekly 1/2 hour parent-child play practice activities and 24 home visits (1 1/2 hours biweekly). Assessments included blind ratings of videotape parent-child interactions, and blind ratings of peer interactions in preschool and parent reports. Results indicated that families demonstrated high rates of attendance and reported high levels of satisfaction with the program. Intervention mothers showed improvements in parenting practices and child outcomes. An NIMH funded randomized controlled intervention trial with 100 preschool-aged siblings of adjudicated delinquents is currently underway (1997-2002).

Study 15

In a second study by Dr. Miller and her colleague Dr. Rojas-Flores, 45 Hispanic families enrolled in a home-based Head Start program were randomly assigned to an intervention or control condition. The intervention consisted of the Spanish-language version of the BASIC program held in ten sessions over three months. Results indicated high levels of attendance and high levels of satisfaction with the program. Preliminary analyses indicated significant improvements in parenting practices and child outcomes.

Study 16

Evaluation Methodology

Design: This study combined the Incredible Years Parenting Program, which deals only with parents, with added components of home visits, child groups, and parent-child interaction groups. All group sessions were audiotaped or observed and led by trained professionals. Ninety minute parent and child groups were conducted separately, followed by 30 minute parent-child interaction groups. In child groups, children engaged in peer playgroups while leaders sought to enhance social competence through the teaching of social skills such as cooperation and sharing. The parenting group curriculum followed the Incredible Years Training Series which, through the use of problem solving discussions and role-playing, focuses on reinforcing positive behaviors and actions of children and using consistent, nonphysical disciplinary measures. Parents were also allowed to observe certain child groups to witness group leaders effectively reinforcing their children's behaviors. Parent and child interaction activities consisted mainly of free-play, arts and crafts, or reading, and allowed parents to practice their learned skills while leaders observed and provided feedback.

Home visits occurred twice monthly for 90 minutes and were designed to help parents effectively transfer and use skills learned on-site into their own homes.

After completing a baseline assessment, families were randomly assigned to the prevention program or a no-intervention, control condition lasting one year. Intervention treatment consisted of 50 parent-attended group sessions held twice weekly for 9 months and 10 in-home visits. Those in the control condition were not contacted until post-intervention assessments were administered to all participating families. Follow-up assessments were conducted six months later.

Sample Characteristics: Participants were recruited through family court, child psychiatric outpatient clinics, and court-mandated after-care outpatient services for adult first-time criminal offenders. Records were screened for presence of a preschool aged child or sibling and potential participants were given study details and contact information through the mail or by their caseworker or clinician.

While 62 families were deemed potentially eligible, only 50 (80%) were deemed eligible through the initial telephone screening interview, 40 of those (80%) consented to the study and 75% (n=30) completed the initial assessments and were randomly assigned to the intervention (n=16) and control groups (n=14). The sample of 19 boys and 11 girls with a mean age of 44 months was 67% African American and 33% Hispanic American. Preschoolers meeting DSM-III criteria for CD and pervasive developmental disorders (as assessed by a clinical interview with child psychologist or psychiatrist) and severe to profound mental retardation were excluded. Parents with current substance use disorders, psychotic disorders, or medical conditions that would not allow for study participation were also excluded.

Measures:

Baseline Characteristics: The Family Status Interview was administered to assess family demographics. The SCID (Structured Clinical Interview for DSM-III-R) was administered to parents to assess anxiety, mood, substance, eating, and somatization disorders and screen for psychotic, anti-social personality, and conduct disorders. Parental stress was measured with the Parental Stress Index and parental IQ was estimated using the Wechsler Adult Intelligence Scale - Revised. Disorders affecting index relatives, specifically CD, ODD, ADHD and substance abuse, were assessed using the Parent as Respondent Informant Schedule.

Parent and child behavior: Videotaped parent and child interactions consisted of free play (7 minutes), puzzle completion (5 minutes), and cleanup (3 minutes). The Global Impressions of Parent Child Interactions (GIPCI) was used to code parental responsiveness and affection. Interclass correlation coefficients were calculated to estimate inter-rater agreement and resulted in .70 for the GIPCI coded data. The Dyadic Parent-Child Interaction Coding System (DPICS) was used to code parental behaviors which were then combined to form summary scores. Positive Parenting scores consisted of behaviors coded as physical positive, unlabeled praise and labeled praise, and Negative Parenting scores consisted of behaviors coded as critical statements and physical negative. The interclass correlation coefficient for DPICS data was .80, but it should be noted that several categories ("acknowledgement" and "descriptive and reflective questions and statements") were eliminated due to unreliability in the measurement by those coding the tapes.

Child behavior problems were assessed through use of the Child Behavior Checklist (CBCL) that requires parents to rate their children on various problem behaviors. Raw scores were converted into Externalizing and Internalizing T scores.

Attendance and satisfaction: The number of intervention groups and home visits were recorded for attendance. Parents completed a consumer satisfaction questionnaire using Likert scales to rate the overall program and the difficulty or usefulness of program methods, learning techniques, and abilities of group leaders.

Outcomes:

Baseline Equivalence: No significant differences between control and intervention group participants were found on the basis of demographic and clinical characteristics of the child, relatives and parents. The characteristics found suggest that participant recruitment and selection successfully identified and involved high-risk children. The sample overall contained high rates of family risk factors, such as parental antisocial behavior, parental history of substance abuse, low SES, and single-parent status, that are predictive of child delinquency.

Posttest:

Attendance and Satisfaction: Intervention group attendance rates averaged at 55% (M=28/50). Completers were defined as those intervention-participating families who attended 5 or more group sessions (n=13). These completers received all 10 home visits. Average satisfaction ratings are as follows: 6 to 7 for usefulness of material covered, 6 to 7 for ease of implementation, 6 to 7 for satisfaction with child's progress, and 6 to 7 for recommendation to others. None rated the usefulness of any program aspect below a 4 (neutral).

Retention Rates: All intervention families (n=16) and 86% of the control families (12 of 14) were reassessed at post-intervention. 88% of the intervention families (14 of 16) and only 43% of the control families participated in the follow-up assessment. Examining the baseline characteristics of those completing follow-up assessment revealed significant group differences for caregiver's education and age (p=.10). Follow-up dropouts were more likely to have higher education (11 years vs. 10 years for those retained) and to be older (39 years vs. 33 years).

Intervention Effects: Results of repeated measures analysis of variance revealed several significant findings. There was a strong, significant Group X Time interaction effect for parental responsiveness and affection. Control parents showed decreases in these behaviors over time, suggesting a positive intervention effect. While there were no significant interaction effects for negative parenting behaviors, the Group X Time interaction for positive parenting behaviors showed somewhat strong, but insignificant effects (F=1.88, eta²=.075). Results also indicate strong, significant interaction effects for Externalizing scores on child behavior problems. Those in the control group showed increases over time while those in the intervention group showed decreases. These analyses were repeated using random regression models to include children with missing data. Results were consistent with the initial analyses.

Follow-Up: Because follow-up results were based on a small number of children in the control condition (n=6), analyses should be considered with caution. There was a marginally significant Group X Time interaction for CBCL Externalizing scores (F=2.46, p=.098).

Study 17

Evaluation Methodology

Design: Day care centers were eligible for selection if they met the following criteria: (a) 90% of enrolled families met income-eligibility requirements for subsidized child care, (b) the centers were licensed by the Department of Child and Family Services, (c) the centers served families with two- and three-year olds, (d) the centers were located in Chicago, and (e) the administrative staff consented to participation.

In order to obtain equivalent experimental conditions, centers were assigned to groups of centers of like size, ethnic composition, percentage of single-parent families, median income, and day care center quality. Once grouped, the centers were then randomly assigned to one of three experimental conditions: parent training delivered to parents and teachers in separate groups (PT + TT; n = 4), parent training delivered to teachers (T; n = 4), or Control (C; n = 3) [Gross et al., 2003, p. 262-263]. The authors state, however, that program effectiveness was evaluated across four conditions, those previously listed and also a parent-only group receiving the parent training (PT).

Parents were eligible for participation in the study if they met the following criteria: (a) the parent was the legal guardian of a two- or three-year-old child enrolled in a participating day care center and (b) all baseline assessments were completed. If a parent had more than one eligible child, the younger child was selected for inclusion.

Seventy-eight parents (30%) were in the combined PT + TT condition, 75 (28%) were in the PT condition, 52 (20%) were in the TT condition, and 59 (22%) were in the C condition. Over the course of the study, 21.2% (n = 56) of the sample dropped out, leaving a final sample of 208 parent-child pairs. Attrition was higher in the two PT conditions (n=47) than in the C and TT conditions (n=9).

Teachers were eligible for participation if they (a) worked in a classroom with two- and three-year-old children and (b) completed baseline assessments prior to initiating Teacher Training in their center. At baseline, 112 teachers participated in the study with 60% (n = 67) head teachers and 40% (n = 45) teachers assistants. Thirty-one percent (n = 35) of the teachers left the study prematurely, leaving a final sample of 77 teachers.

Sample Characteristics: The analysis sample of 208 was primarily African American (57.2%), followed by Latino (29.3%), White (3.4%), and Multi-ethnic and Other (10.1%). Parents were primarily employed full-time (56.7%) or part-time (13.0%), in school (10.6%), working and going to school (8.2%), unemployed (4.3%), and other (7.2%). Parents with high school education or greater constituted 42.8% of the sample, and 57.1% of the sample had some college or vocational training.

Measures: Parenting self-efficacy was measured using the 38-item Toddler Care Questionnaire (TCQ). This questionnaire assesses parents' efficacy in managing a range of tasks and situations relevant to raising small children.

Parent discipline strategies were measured using the Parenting Scale, which includes 30 discipline situations in which parents are asked to rate the degree to which their discipline strategies are more typical of one anchored strategy than another.

Parent behavior was assessed from a 15-minute videotaped parent-child free-play session using the Dyadic Parent-Child Interactive Coding System-Revised (DPICS-R). Parental behavior was coded for key parent and child behaviors targeted in the parent program, including the frequency of commands, critical statements, and physically coercive behaviors. Two normally distributed composite variables were then created: total commands and positive parent behaviors.

Parent stress was measured in three ways: depression, everyday stress, and neighborhood problems. Depression was assessed using the 20-item Center for Epidemiological Studies Depression Scale (CES-D), which is designed to measure depressive symptoms in the general population. Eighty-one of the parents fell into the depressed range on the CES-D. Everyday stress was measured using the 20-item Everyday Stressor Index (ESI), which asks parents to report the extent to which they are bothered by such stressors as financial concerns, employment problems, parenting concerns, family responsibilities, and interpersonal conflicts. Neighborhood stress was measured by the 22-item Neighborhood Problem Scale (NPS), which asks parents to report the degree to which each of 22 environmental stressors is viewed as a problem in their neighborhood. Environmental stressors presented in the NPS include drugs, gangs, guns and violence, unemployment, etc.

Children's behavior problems were assessed using three sources of information that were analyzed separately: parent and teacher reports of problem behaviors at home and in the classroom and an observational rating of problem child behaviors using a 15-minute videotape of parent-child play session. Parent-reported child behavior problems were assessed using the Eyberg Child Behavior Inventory (ECBI), a 36-item inventory for children containing two scales: the Intensity Scale and the Problem Scale. Fifty-one of the toddlers fell into the clinical range on behavior problems (using a criterion score of 15 on the Problem Scale). Teacher-reported child behavior problems were measured using Kohn's Problem Checklist (KPC), a 49-item rating scale designed for use by teachers in day care centers. Child behavior is analyzed along two dimensions: apathy-withdrawal and anger-defiance. Thirty-nine of the children received high classroom behavior scores. Observer-rated child behavior problems were assessed using a ratio of aversive child behaviors to positive child behaviors created from eight DPICS-R items. The ratio of the total number of aversive to positive child behaviors was calculated and then transformed using a logit transformation to create a variable called observer-rated negative child behavior.

Finally, consumer satisfaction was measured at post-intervention using an end-of-program questionnaire. The same questionnaire as used by Webster-Stratton was administered in order to assess parents' and teachers' perceptions of (a) improvements in children's behavior, (b) program difficulty, and (c) program utility.

Parents and children were assessed by family interviewers four times over a 15-18 month period (baseline, immediate post-intervention, 6-months post-intervention, and 1-year post-intervention). Parent assessments took place either at the center or parent home, based on participant preference. All questionnaires were read aloud to participants in either English or Spanish, unless the parent preferred to complete them independently.

Analysis: The data were analyzed using growth curve modeling in order to measure change over time. Time periods included baseline, postintervention, six-month follow-up, and one-year follow-up. The three polynomial terms included in the initial regression model were (a) a linear effect, (b) a quadratic effect, and (c) a cubic effect. In addition, a one-way ANOVA (one tailed) with planned contrasts was run in order to test for significant changes in scores from baseline to one-year follow-up and from post-intervention to one-year follow-up.

Outcomes
Baseline Equivalence and Attrition: At baseline, all dependent measures were compared by experimental condition and training group nested within day care center. The four experimental conditions were equivalent on all measures except the teacher rating of child behavior problems using Kohn's Problem Checklist (KPC). When baseline scores were examined for equivalence by parent group within day care center, teacher ratings of child behavior problems in the classroom was the only significant variable.

Over the course of the study, 31.2% (n=35) of teachers dropped out of the study. There were no significant differences between teachers who dropped and those who remained. Among parents, 21.2% (n=56) dropped from the study. Parents who dropped had significantly lower overreactive discipline scores than parents who remained, indicating that dropouts were less likely to use harsh and coercive discipline strategies with their children than those retained in the study. Parents who remained were also more likely to be Latino. Attrition was unrelated to other parent-child outcomes or demographic variables to parent stress.

Posttest:
Parent Outcomes: After controlling for the effects of ethnicity, depression, and everyday stress, parents who received parent training reported a 2.1 point greater increase in self-efficacy scores at each time point up to one-year post-intervention compared with parents who did not receive training. Planned contrasts of mean scores revealed a significant improvement in parent self-efficacy among parents who received parent training from baseline to the one-year follow-up.

After controlling for the effects of ethnicity, depression, and everyday stress, parents who received PT reported a 1.0-point linear drop in coercive discipline from baseline to post-intervention compared with parents in the C and TT groups, suggesting an immediate linear parent training effect on use of coercive discipline not seen in the C and TT conditions. The subsequent combination of linear and curvilinear effects, however, indicates that the intervention was effective in reducing coercive discipline at post-intervention, but the effect disappears by the six-month follow-up. Further analysis indicated a significant improvement in coercive discipline strategies among PT and PT + TT parents from baseline to post-intervention, followed by a significant backslide of effects from post-intervention to one-year follow-up.

For the variable positive parent behavior, significant effects were noted for ethnicity and for linear and quadratic parent training effects. The significant linear effect indicates an immediate post-intervention effect of PT leading to more positive parenting behaviors. The significant negative quadratic term indicates a slight attenuation of effects at the follow-up assessments. Analysis of changes in mean scores with planned contrasts revealed a significant improvement in observed positive parent behavior among PT and PT + TT parents from baseline to post-intervention without statistically significant backsliding from post-intervention to one-year follow-up.

After controlling for ethnicity, parents in all three experimental conditions had significant reductions in their use of commands during free-play as compared to parents in the control condition. Further analysis revealed a significant drop in parents' observed use of commands among PT, TT, and PT + TT parents from baseline to post-intervention, as well as significant long-term effects from baseline to one-year post-intervention.

Summary of Parent Outcomes: After controlling for parent stress, PT and PT + TT parents reported higher self-efficacy and less coercive discipline and were observed to have more positive behaviors than C and TT parents from baseline to post-intervention, with effects retained one year postintervention for parent self-efficacy and positive parent behaviors.

Child Outcomes: There were no intervention effects on parent-rated ECBI Problem Scores or for observer-rated negative child behavior. From baseline to post-intervention, results from three sets of chi-square analyses indicated that both PT and TT led to significant improvements in high-risk classroom behavior problems relative to controls. The combined PT + TT condition was not significantly better than either PT or TT alone; however, a larger percentage of the high-risk children in the PT and TT condition moved to the low-risk group than did their control group counterparts.

Although most of the sample's classroom behavior improved from post-intervention to the one-year follow-up, significant effects were found for PT and TT. Surprisingly, children in the combined PT + TT group fared significantly worse than did children in the single PT or TT groups.

It can be concluded that adding teacher training to parent training provides no extra benefit, and the only benefit for teacher training alone was demonstrated in improvements in classroom behavior problems among high-risk children.

Consumer Satisfaction: Although the overall parent-reported level of satisfaction with the program was very high, and parents perceived their children's behavior as having improved after completing the program, many parents reported it was difficult to attend the meetings or complete the homework assignments.

Although the overall teacher reported level of satisfaction with the program was somewhat less enthusiastic in terms of improving children's classroom behavior, the program was rated "very helpful" by 98% of teachers, and 81% reported they would "strongly recommend" the program to others.

Study 18

Evaluation Methodology:
Design: This evaluation is based on student data for 22 closed K-3 Program cases out of a total 482 served. These students entered the K-3 Program between April 1996 and March 1998. Cases had to have been open at least four months to be included in this analysis.

Measures: Change in child behavior was measured using the Eyberg Child Behavior Inventory and the Sutter-Eyberg Child Behavior Inventory. These instruments were completed by parents and teachers when students entered, completed, and/or left the K-3 Program. Teachers completed the child behavior change forms at the end of the school year.

The program also measured parents' ability to establish and maintain positive behavior guidelines for their children. Tests were taken pre- and post-program to evaluate the "parenting practices" scores. A voluntary component of this initiative included video and training material for parents.

The program also attempted to measure academic progress outcomes; however, because Delaware has no statewide test to assess student academic progress from kindergarten through second grade, teacher ratings of student progress in reading and mathematics were used in the evaluation. The rating systems used "Satisfactory" as approximately A's, B's, and C's, while "Academically at Risk" students indicated grade levels of D's, and "Unsatisfactory" represented students producing work equivalent to F's.

Analysis: The analysis of outcomes is based primarily on information obtained from pre- and post-program teacher surveys. Also included in the analysis was a comparison of pre- and post-program parent school involvement surveys to provide a measure of how often parents contact classroom teachers and are involved in a child's activities at school.

Outcomes:
Students whose families moved, parents who were non-compliant, and families which became active with the Division of Family Services (DFS) or the Division of Child Mental Health Services (DCMHS) accounted for 54% of case closure reasons.

The average teacher child behavior intensity scores, which indicate the frequency of inappropriate or disruptive behaviors, decreased for all students from pre- to post-program measurement. For those students for whom post-program parent scores were available, the average child behavior intensity scores also decreased from the pre- to the post-program.

Lower scores represent improvements in child behaviors. The average teacher behavior scores decreased from 163.2 pre-program to 147.6 post-program, while the average parent behavior scores also decreased from 124.4 pre-program to 104.4 post-program. Both of these decreases are statistically significant. A child behavior intensity score of 127 is considered the threshold for higher frequencies of inappropriate or disruptive child behaviors.

Parents were sampled pre- and post-program, and had a statistically significant improvement in their rates of participation at school activities. Teachers were also tested regarding improvements that they witnessed in parental involvement and reported a statistically significant improvement.

With regard to parenting outcomes, statistically significant improvements in "parenting practices" were found for parents involved in the program, regardless of whether they attended video parenting classes.

On academic outcomes, at the end of the year, for students above the 127 threshold score separating higher and lower frequencies of inappropriate or disruptive behavior: 1) 56%, 22%, and 22% scored satisfactory, at-risk, and unsatisfactory, respectively, for English. 2) 64%, 22%, and 14% scored satisfactory, at-risk and unsatisfactory, respectively, for Math. For students below that threshold their corresponding rates on English were 57%, 23%, and 20%. While on math these students were reported as having rates of 61%, 22%, and 17%.

Assessments were also made of the comparative absence rates between the K-3 Program and statewide absence rates. It was found that the average absence rate for Kindergarten to Third Grade was 8.5. To analyze if the K-3 Program experienced significantly different results, the program evaluators measured how many students involved in the K-3 Program missed 11 days or more (1 standard deviation away from the mean statewide). It was found that 54% reached this benchmark, constituting serious attendance problems.

Additionally, the K-3 Program evaluators sought to determine the program's effects on disciplinary incidents. Overall, 22% of the total number of students in the K-3 Program schools who received out-of-school suspensions were participants in the K-3 Program.

Twenty focus groups were conducted with K-3 teachers, nurses, principals, guidance counselors, and parents. The results were very positive and indicated strong support of the K-3 Early Intervention Program.

Study 19

Evaluation Methodology:

Design: Participants were systematically recruited from monthly census reports at one child welfare agency in New York City. Parent pairs were selected depending on the eligibility of the foster child. Authors selected 152 potentially eligible children and subsequently conducted interviews with caseworkers and biological and foster parents to assess study criteria. On the basis of these screening criteria, authors excluded 48% (72/152) of the children who did not meet the study criteria and 10% (15/152) who refused. Authors excluded children with documented developmental disabilities or an official report of sexual abuse because they require specialized interventions. Authors also excluded biological or foster parents with a known mental handicap and those who did not speak English or Spanish. From those eligible, authors enrolled 81% (64/79) parent pairs, who were randomly assigned to intervention (n= 40 parent pairs or 80 individuals) or to a usual care comparison condition (n= 24 parent pairs or 48 individuals). Enrollment did not differ by parent (biological vs. foster) or condition (intervention vs. usual care.)

Biological and foster parents received three assessments: one at baseline (before intervention), one at the end of the intervention (3 months after baseline) and one at follow up (3 months after the end of intervention). The two-component intervention consisted of a 2 hour Incredible Years parenting course and a 1 hour co-parenting program offered to biological and foster parents on 2 separate week days for 12 consecutive weeks. Biological and foster parents were eligible to participate if the foster child met the following criteria: substantiated history of child maltreatment, residence in a non-kinship foster home, and a goal of reunification.

The two-component (parenting and co-parenting) 12-week intervention was offered at the agency by a trained bilingual (English and Spanish) team from the agency mental health unit. The team (parent leaders) delivered the group intervention in pairs. The parenting component was offered to groups of 4 to 7 parent pairs for 2-hour sessions by using a manualized version of the Incredible Years "Parents and Children Basic Series Program." The program is comprised of four units: play, praise and rewards, effective limit setting and handling misbehavior. Strategies included videotaped vignettes, role plays and homework.

The co-parenting component was offered to individual families (biological and foster parent pair and target child) in a separate session by using a newly devised curriculum. During this session, parent pairs had the opportunity to expand their knowledge of each other and their child, practice open communication, and negotiate inter-parental conflict regarding topics such as family visitation, dressing and grooming, family routines, and discipline.

Sample Characteristics: The final sample consisted of 128 parents (64 biological and foster pairs) of maltreated children placed in short-term foster care; parents were mostly mothers, except for 7 biological fathers and 1 foster father. Sample was primarily Latino (57%) and African American (33%); approximately 50% were foreign born, had less than a high school diploma, and were never married. Approximately one third worked outside the home. Children were between the ages of 3 and 10 years and were placed in regular foster homes for an average of 8.4 months at baseline. Prior to placement, most children resided in apartment buildings in inner city neighborhoods. There were fewer neglected (71%) and more abused (29%) children in the intervention than in the usual care condition (100% and 0%, respectively).

Measures: A variety of measures were utilized; measures were administered to both biological and foster parents. The "Practicing Parents Interview" was used to assess discipline attitudes, beliefs and practices. Co-parenting relationships were assessed by using items from the "Family Functioning Style Scale" and from the "Family Adaptability and Cohesion Scale." Externalizing and conduct problems were assessed by using items from the "Child Behavior Checklist" and the "Eyberg Child Behavior Inventory." The key outcome measures included self-reports of parenting practices, co-parenting and child externalizing problems.

Analysis: Preliminary analysis was conducted to assess baseline differences between biological and foster parent psychosocial characteristics (age, ethnicity, education, employment status, etc.) and to assess baseline differences across study condition (intervention vs. usual care). Primary analysis included all randomized participants whose outcome scores (discipline scores, co-parenting, and child-externalizing problems) were subjected to general linear model analysis of covariance (ANCOVA) at two endpoints: at the end of the intervention and at follow-up. Baseline scores for each dependent measure were used for the covariate in the analysis. Intervention main effects (parent groups combined) and interaction effects (parent X study condition) were analyzed in the primary analysis. Additionally, a secondary analysis examined mediators and moderators of change by using separate ANCOVA's for each dependent variable.

Outcomes:

Intervention comparisons with parent groups combined: There was a significant difference in the intervention group compared to usual care conditions on positive discipline at the end of the intervention, and again at follow-up. There was a significant difference in the intervention group compared to the usual care condition on clear expectations at follow-up. There was a significant difference in the intervention group compared to the usual care condition on co-parenting flexibility, co-parenting problem solving, and co-parenting total at the end of intervention. Although not statistically significant, intervention families reported children as having lower Child Behavior Checklist externalizing T-scores and Eyberg Child Behavior Checklist total T-scores when compared to the usual care group at follow-up.

Intervention comparisons by parent: Condition X Parent interactions were also examined to test for differential intervention effects by parent. Only one significant condition by parent interaction was found at follow up: Biological parents retained intervention gains for positive discipline more than did foster parents.

Completers vs. Non-completers: ANCOVA analysis showed that Incredible Years completers reported higher positive discipline than did non-completers at follow-up. There was a Condition X Ethnic Status interaction for harsh discipline at the end of the intervention: African American parents reported more improvement in harsh discipline than did Latino parents. Initial level of child conduct problems (high- vs. low-risk) did not moderate intervention effects.

Study 20

Thirty-five parents monitored in a child protection service for child neglecting behaviors participated in either the intervention group (n=26) or were on the waiting list (n=9). Parents who were monitored at the Montreal Youth Center for their child neglecting behaviors were identified by their caseworker to participate in Incredible Years. The first group of parents who registered were assigned to the intervention group (received PTP plus regular services). Nine parents who registered later were put on the waiting list for the next year. They made up the control group (received regular services during the study period). 

The program lasted 16 weeks, was in a group format and aimed to develop a harmonious parent-child relationship, to support parents in learning and consistently applying effective practices, and to improve problem solving and communication skills within families and with teachers. A repeated measures design was used to test the program's effect on parenting practices, parents' feelings of self-efficacy, parents' perception of their child's behavior, and parents' satisfaction. Parents were tested twice, during a 19-week interval, before and after the parent training program. Analyses of variance comparing intervention and control groups with repeated measures revealed that the program had a positive impact on parenting practices (harsh discipline, physical punishment, praise/incentive, appropriate discipline and positive verbal discipline) and parents' perception of their child's behavior (frequency of behavioral problems and number of problematic behaviors). No change on clear expectations from parents, or on parents' self-efficacy was observed.

Study 21

Evaluation Methodology

Design: The study took place in 11 Sure Start areas in north and mid-Wales, which are high-risk, disadvantaged areas in which the government has invested money for preventive parenting support for families of preschool children. The 12-week parenting intervention was provided. The Eyberg Child Behavior Inventory was administered to socially disadvantaged families, with a child aged 3 or 4 years. Eligibility for the program was determined if the child lived with the primary caregiver, scored above the clinical cut off on either the Eyberg problem or intensity scale, and the primary caregiver was able to attend group times. A total of 153 families were eligible and consented to take part. Families were randomized on a 2:1 basis, with 104 to intervention and 49 to the wait list control. Twenty were lost to the follow-up six months later, 18 of whom were in the intervention group. These 20 were included in the intent-to-treat analysis, with the assumption of no change in their baseline scores.

Measures: The Eyberg Child Behavior Inventory was the primary measure used to assess number and intensity of conduct problems. The Strengths and Difficulties Questionnaire assessed conduct and hyperactivity problems, as well as the Conners Abbreviated Parent/Teacher rating scale and the Kendall Self Control rating scale. There were home observations using the Dyadic Parent-Child Interaction coding system which assessed positive and critical parenting and deviant child behaviors. The Parenting Stress Index assessed stress levels of parents, the Parenting Scale measured parenting competencies, and the Beck Depression Inventory measured parental depression.

Analysis: ANCOVAs taking account of area, treatment, and baseline response values were used. As previously mentioned, the analysis was intent-to-treat.

Outcomes
Baseline Equivalence and Attrition: The intervention group had a smaller percent of boys and was somewhat more likely to be headed by a single parent than the control group. The 18 lost from the intervention group at follow-up did not differ significantly from those intervention families who remained in the study. This was not tested in the control families because of the small number lost (n=2).

Posttest: Children in the intervention group had significantly reduced antisocial and hyperactive behavior and increased self-control compared with the control group children. Intervention group parents also perceived intensity of problems in siblings as less severe at follow-up. Although intervention families had been above the level of clinical concern at baseline, at follow-up, they were below the level of clinical concern. Observational results corroborated the questionnaire findings. There were also reductions in parental stress and depression levels, and improvements in parenting competencies in the intervention parents compared with the control parents.

Study 22

Evaluation Methodology

Design: Children were randomly assigned to treatment or a wait-list control condition. The entry criteria were: (1) child between the ages of 3-8, (2) no debilitating physical or intellectual impairment, or history of psychosis and not already receiving psychological treatment, (3) primary referral for conduct problems that continued for more than six months, (4) parent report symptoms were clinically significant on the Eyberg Child Behavior Inventory, and (5) met diagnostic criteria for ODD and/or CD. Participants were recruited from families requesting treatment at the University of Washington Parenting Clinic.

Sample Characteristics: The final sample included 142 boys and 39 girls between the ages of 3.0 and 8.5 years at intake. The sample was 79% European American, 9% Latino/Hispanic American, 6% Asian American, 4% African American, 1% Native American, and 1% Unspecified. Mothers ranged in age from 24-65; fathers 26 to 77 years. Most (73%) were currently married or living with their partner. Most (96%) had completed high school. Forty percent of mothers and 52% of fathers had completed a 4-year college degree, and 61% of mothers and 93% of fathers were currently employed.

Measures: Only parent reports were utilized in this study; however, there are both mother and father ratings. The parent-reported Child Behavior Checklist Internalizing Subscale assessed depression (mother and father ratings used), the parent-reported Child Mood subscale of the Parenting Stress Index determined depressed mood of child, and the mother- and father-reported Eyberg Child Behavior Inventory determined total behavior problems. Also, the Parenting Stress Index assessed child behavior problems and parental adjustment to yield a total stress score for the parent. A Parenting Competence subscale and Attachment subscales were also used.

Analysis: Intent to treat ANOVAs (pre vs. post x intervention condition) were conducted, controlling for baseline externalizing problems in all outcome analyses. All analyses were repeated using only children who were depressed at baseline. Most analyses were done separately by parent informant, due to the large amount of missing data from fathers. Mother-father ratings were combined for tests of clinical significance and in mediation tests using Structural Equation Modeling.

Outcomes
Baseline Equivalence and Differential Attrition: The two groups looked fairly similar at baseline, although there were more partnered families in the control group. Authors state none of the demographic or key study variables at baseline were significantly different between the two groups. There was complete data for mother ratings at baseline and post-treatment, so no attrition analysis was needed. There were no known differences between fathers who completed ratings and fathers who did not on mother-rated CBCL Internalizing and PSI Child Mood. They were also similar on demographic variables.

Posttest: Mother reports of CBCL Internalizing and depressed mood were significantly lower in the intervention group, compared to the control group. This was also true in the sample of children with elevated levels of depression at baseline. Father reports were not significant for the CBCL in the full sample, but were significant for the depressed subsample of children. Father reports on depressed mood were marginally significant. There was also some evidence of clinical significance favoring the treatment group, especially with children in the severe-clinical range moving to the normal range. Using both mother and father reports, 31% of children in the treatment group who were rated in the severe-clinical range moved to normal, compared to 0% in the control group (marginally significant).

There were also significant treatment effects for the proposed mediators of parent-reported parent competence, attachment, and role restriction.

Mediation: The mediation model tested the hypothesis that treatment condition would be significantly related to perceived parenting effectiveness (mother and father ratings on the Parenting Competence and Role Restriction subscales of the PSI) and internalizing and externalizing symptoms at post-treatment, controlling for baseline ratings of these constructs. There was a direct treatment effect on post-treatment internalizing symptoms, a direct effect on self-perceived parenting effectiveness at post-treatment which, in turn, had a direct effect on post-treatment internalizing symptoms. The total indirect effect of treatment condition on internalizing symptoms was significant. This provides evidence that parenting effectiveness partially mediated the effect of treatment on internalizing symptom improvement.

Study 23

Evaluation Methodology

Design: The evaluation occurred in five urban sites in Great Britain that specialized in offering the Incredible Years. Seventy-six families were randomized to the Incredible Years parenting intervention or to a wait-list control condition. Control group families were offered the intervention six months later. The 14-week parenting intervention was delivered weekly to groups of 10-12 parents in two-hour sessions. Children were ages 2-9, referred for help with conduct problems with a score above clinical cutoff on the Eyberg Problem Scale, with parents who were able to attend group and communicate in English. Exclusion criteria included: child severely disabled, in temporary care, parent drug addicted, or previous attendance at one of their programs.

Assessments took place in the home at Time 1 (pre-intervention and pre-randomization), Time 2 (post-intervention 6 months later), and Time 3 (follow-up 18 months post-recruitment). It should be noted that the 18 month follow-up did not include the control group, as these families had already received the intervention.

Sample Characteristics: A high proportion of the sample was boys living in poverty, with a single parent showing signs of depression. Conduct problem scores were high.

Measures: Parent report and observations were used to measure program effects. Parent-child interaction was recorded using a video camera in six structured settings in the home designed to assess child and parent behaviors. The parent-reported Eyberg Child Behavior Inventory Problem and Intensity subscales were used assessed child conduct problems. There were also measures of parenting skill, confidence, and mood. The Beck Depression Inventory assessed parental depression.

Analysis: ANCOVAs were used, controlling for baseline scores, in an intent-to-treat analysis.

Outcomes
Baseline Equivalence and Differential Attrition: There were no significant differences between groups on demographic factors, parent-reported parenting skills, depression or child behavior problems. There was some variation between groups on observation measures. There was 93% retention at six months, and 90% at 18 months. Families lost to follow-up did not differ significantly from those retained.

Posttest: At posttest, there were intervention effects for the parent-reported Eyberg problem and intensity scales (child behavior), and on observed child negative behavior. There were also intervention effects on the sibling Eyberg Problem scores. Intervention effects were also found for parenting skills and confidence, specifically less use of negative strategies and increases in positive parenting, and sense of competence and parenting skills. There was no effect for parent depression.

Long-term: At the 18-month follow-up, the changes were maintained, as evidenced by no significant change from post-intervention to follow-up on within-group paired-sample t- test. There was no control group at this point, as they had received the intervention.

Mediation: Using hierarchical multiple regression, with change in child negative behavior as the dependent variable, the effect of treatment was entered as Step 1 (significant), and Step 2 shows that the effect is attenuated when positive parenting is entered as a mediating variable. The significance of the change was tested using the PRODCLIN program, and showed a significant partial mediation effect. Change in child negative behavior was not correlated with change in parent sense of competence or depression.

Study 24

Evaluation Methodology

Design: Children ages 4-8 referred to two child psychiatric outpatient clinics for treatment of oppositional or conduct problems were studied. Children were screened using the ECBI 90th percentile as a cut-off score. Exclusionary criteria included gross physical impairment, sensory deprivation, intellectual deficit, or autism, and children receiving other psycho-therapeutic interventions, as well as children on ADHD medication more than six months prior to study entry.

In all, 136 children were randomized to the BASIC parent program (PT; n = 51), PT combined with child therapy (PT + CT; n = 55), and a waiting list control group (n = 30) who were offered the intervention after six months. Child therapy was based upon the Incredible Years Dinosaur School Program, which sought to provide child social skills, conflict resolutions skills, playing and cooperation with peers.

Sample Characteristics: All children had a diagnosis of ODD (n=111) or threshold ODD (n=16), and 18.9% received a subthreshold (n=14) or definite CD diagnosis (n=10). Also, 35.4% (n=45) met diagnostic criteria for ADHD.

Measures: Drugli and Larsson (2006) examine a diverse set of outcomes reported by parents, children, and teachers.

Parents were the sole informants for the other studies, although there are both mother and father reports. The parent-reported Eyberg Child Behavior Inventory (total intensity and problem subscales), was the prime outcome measure. The parent-reported CBCL (internalizing, aggression and attention subscales) was also used. The Parent Practices Interview, modified from the Oregon Social Learning Center's discipline questionnaire to apply to young children was used to measure inconsistent discipline, harsh discipline, and positive parenting. The Parent Stress Index consisted of 101 items to assess parents' perceived stress related to both child behaviors and parenting.

Analysis: Associations between categorical variables were analyzed with chi-square tests. Differences in group means between treatment conditions posttreatment and at 1-year follow-up (no control group, maintenance effects only) were analyzed with ANCOVAs using pretreatment scores as covariates. Imputation was used for subjects with one missing value on any of the three assessment points, baseline, posttest, and 1-year follow-up. Findings are reported for completers only, but authors state that intent-to-treat analysis found identical results for the primary ECBI measure.

Outcomes

Baseline Equivalence and Differential Attrition: There were no significant differences among the groups on any of the demographic variables, diagnostic status, the primary outcome measure (ECBI), or use of medication because of ADHD. At posttest, the sample included 127 (of the original 136) children (2 dropped and 4 failed to complete PT; 3 failed to complete the combined treatment, and 2 controls did not complete a posttest assessment). At the one-year follow-up, the wait list control group was no longer available as they had received treatment; however, the PT condition included 40 and the combined condition 48 that completed the assessment. Because intent-to-treat analysis showed identical results for the primary measure, the ECBI, outcomes were reported only for completers.

Posttest:

Drugli and Larsson (2006)

At posttest, children in the combined parent and child training had significantly lower aggression scores, significantly lower clinical levels of aggression, and significantly higher problem-solving skills. At follow-up, most gains were not maintained. These outcomes did not improve significantly from parent training alone.

No significant benefits of either intervention emerged for several other measures: attention problems, internalizing problems, social competence, peer interactions, student-teacher relationships, and parent-teacher involvement.

Larsson et al. (2009)

ECBI Scales: Mothers and fathers in PT reported lower intensity scores for the children, and fathers reported significantly fewer problems than wait-list controls. There were no differences between the combined condition and wait-list controls. There were also no significant differences between the two interventions.

CBCL Scales: Mothers in the PT condition reported less aggression, better attention, and less internalizing problems than controls. There were no significant differences reported by fathers in the PT condition vs control. In the combined condition vs. controls, only mother-reported aggression was significant, with the intervention lessening aggression.

Parenting Practices and Stress: Parental use of positive strategies increased after treatment in both conditions (reported by both mothers and fathers). The use of harsh and inconsistent discipline decreased in both conditions as reported by mothers, and fathers reported less inconsistent discipline in the combined condition only. Mothers' experience of stress was less in both conditions, although fathers reported less stress in the PT condition only.

One-Year Followup (maintenance effects only, no control group): No differences between PT and the combined condition were found on any parenting measures, other than fathers in the combined condition reported a significant reduction in harsh disciplining of the child as opposed to those in PT.

About 80% of the children no longer received an ODD diagnosis at the one-year follow-up. From posttreatment to one-year follow-up, mothers reported an improvement in clinical significance, with children in PT improving from 57.1% to 63.2%, and in the combined condition 54.9% to 65.6%.

Overall, it is questionable that the child therapy combined with parent training added much.

Long-term Follow-up

Drugli et al. (2010)

A telephone assessment was completed with 54 of the 99 treated families (54.5%) 5-6 years after treatment ended, to determine diagnostic status of the children. The mean age of the children was 12.1 years. There were no significant differences between the original sample and the long-term follow-up sample on demographic variables, however, there were significantly more participants from the combined PT+CT condition (63% of sample) than the PT only group (37% of sample). Analyses were conducted on combined treatment conditions and only data returned from mother reports were used for all time points due to low father response rate.

Independent t-tests indicate that 33% of the children (n=18) had ODD and/or CD diagnoses at the 5-6-year follow-up. The majority of these children also had a diagnosis of ADHD or anxiety/depression. Of those who did not have an ODD/CD diagnosis (n=36), more than half (n=20) had diagnoses of ADHD, anxiety/depression, or both. About a third (35.2%) of the sample did not have any psychiatric diagnosis at the long-term follow-up. Five (9.3%) of the children received an ODD/CD diagnosis at both long-term follow-up time points, while half of the children (51.9%) changed diagnostic status between long-term follow-up periods (24.1% deteriorating and 27.8% improving), and 38.9% of children maintained a non-diagnostic status at both time periods.

Variables that significantly predicted (at pre-treatment) a diagnostic status of ODD/CD at 5-6-year follow-up included female child, single parent mother, a diagnosis of CD, and higher levels of child internalizing problems. Variables at post-treatment that significantly predicted ODD/CD diagnostic status at the 5-6-year follow-up were high levels of child internalizing and externalizing problems, high levels of mother's depressive symptoms and maternal stress in parenting.

Study 25

Evaluation Methodology

Design: Parents of 16,002 4-year-old children were sent a Child Behavior Checklist (CBCL) in the mail; more than half of the parents returned the questionnaire and 503 children who scored above the 80th percentile on the Aggressive Behavior scale of the CBCL were invited to participate in the study. Using a quasi-experimental design, families were matched on the child's gender, level of aggression, IQ, parent's education level, stress level and address density of the place of residence of the family, and then placed based on geographic reason in either the intervention group or the control group. 72 families (26% of the selected families) enrolled in the intervention group and 110 families (47% of the selected families) enrolled in the control group, however, after the matching procedure, only 72 families remained in the control group. The aggressive behavior score of children whose parents agreed or refused to participate in the study was not significantly different, neither in the intervention group nor in the control group. The control group parents were blind to their condition and were allowed to access regular services for their child's behavior.

The BASIC and ADVANCE curriculum were delivered in 18 2-hour sessions to 6-11 parents in each group. After the termination of the IY program, two booster sessions were offered at 3 months and 6 months post intervention.

None of the families dropped out during the intervention, however, five families (3 from control, 2 from intervention) were lost from posttest to follow-up at year 2.

Sample Characteristics: The children in the sample were primarily Caucasian, 71% male, with an average age of 4 years and 2 months. The average age of the primary caregiver was 35 years.

Measures: The 19-item aggressive behavior scale on the Child Behavior Checklist (CBCL) was used during recruitment. A parent-rated measure of child behavior - the Eyberg Child Behavior Inventory (ECBI) - was used to assess occurrence of conduct behavior problems at baseline, posttest and one- and two-year follow up. The Dyadic Parent-child Interaction Coding System - Revised (DPICS-R) was used by observers to rate 20-minutes of parent-child interactions. Parent and child behavior was first video-taped and then coded by coders blind to the condition for Critical Statements, Labeled Praise, Conduct Problems and Compliance. A parent-rated questionnaire - Parent Practices Interview (PPI) - comprising 6 scales was used to measure parenting skills and discipline styles of the parents.

Analysis: Intervention effects were evaluated using a repeated measures ANOVA using Helmert contrasts of the time x group interaction. Comparison between pretest versus all subsequent assessments, posttest versus all subsequent assessments and follow up at year one versus follow up at year 2 was conducted. Due to low level of attrition, no data was imputed. If a scale score was missing for a family, that scale score of the matched family was removed from the analysis as well. Scale scores were excluded from the analysis if more than 25% of the items of a scale of measures were missing. The study used an intent-to-treat analysis.

Outcomes

Fidelity Monitoring: Staff providing the Incredible Years intervention received 3-days of training, practice with a pilot group and supervision from IY trainers to deliver the curriculum. Intervention sessions were videotaped and reviewed during weekly meetings to ensure that the program was delivered with fidelity. A quarter of these videotapes were also peer-reviewed. The IY manual was used throughout and parental evaluations and checklists for group leaders were completed after each session.

For parents, the attendance rate was reported at 78% and an average of 14 of 18 sessions was attended by at least one parent.

Baseline Equivalence and Attrition: The intervention and control groups did not differ on any demographic characteristics except age of the child, where the children in the control group were an average of 2 months older than the intervention group. There were also no significant differences between intervention and control groups on the parent-rated measures; however, on observed behavior of parents and children, the following differences were significant: parents in the intervention group had a higher level of Critical Statements used and children in the intervention group showed more conduct problems than children in the control group.

Families lost to follow-up did not differ in their initial level of aggression from those retained. No further information regarding differential attrition was provided.

Outcomes
Post-test and Long-term Follow-up : Sustained effects were found on 6 of the 12 outcome measures of child and parent behavior. However, the program more clearly influenced parent behavior than child behavior.

Pretest versus all later assessments: On parent-ratings of their parenting skills using the PPI, repeated measures ANOVAs revealed several significant differences indicative of sustained effects in parenting skills between the intervention and control groups for the following scales: appropriate discipline; harsh and inconsistent discipline; and, praise and incentives.

Observation of parenting using DPICS found a significant difference between intervention and control group for the change in scores on critical statements and labeled praise between pretest and all subsequent tests.

For parent-rated child behavior on the ECBI, no significant differences between the intervention group and control group were found. For the observer-rated child behavior on the DPICS, one of two outcomes differed significantly across groups.

Posttest versus subsequent assessments: Only one of the 12 measures (parenting praise) was significant when comparing post-test result to all later assessments.

Follow-up Year 1 versus Follow-up Year 2: When comparing results between follow up at first year versus follow up at second year, the only significant difference between intervention and control groups was on child compliance; however, this was in the wrong direction, with control group children showing a larger increase in compliance scores than intervention children.

Effect size : No effect sizes were reported.

Mediating Effects : Improvement of parenting skills was examined as a mediator of improvement in child conduct. Results indicated that a decrease in critical statements from parents due to the IY parent program during the intervention led to a decrease in conduct problems in children two years after termination of the intervention.

Study 26

Evaluation Methodology

Design: Data was collected from participants of the Pathfinder Early Intervention Project (PEIP), which was established in 2006 and funded through the Department for Children, Schools and Families (DCSF) in the UK. Incredible Years was one of three evidence-based programs delivered to parents of children ages 8 to 13 years in six authorities of the PEIP. Those authorities were used for this study. While 375 children (ages 1 to 16) and their parents were actually involved in the PEIP, the data collected for this evaluation included only the primary parents of the 280 children (75%) who fell within the targeted age range of the study. All children and their parents evaluated received the IY program, and there was no control group. Participants in the study were recruited through referrals from Child and Adolescent Mental Health Services (CAMHS) and other agencies, or through self-referrals from advertisements. Parents received the intervention primarily through the schools. The participating children were identified as being at high risk of developing conduct disorder because of evidence of significant behavioral difficulties as well as other at-risk family characteristics (socio-economic disadvantage, young and/or single parents, unemployment, and adult mental health problems).

Modifications to the curriculum were developed by the program designer to adjust the material to the older targeted age group. This included using the 12-week BASIC program and adding content from the ADVANCE component, to deliver a total of 16-17 sessions. Most of the program facilitators had not previously delivered the IY program and backgrounds included nurses, psychologists, psychiatrists and social workers. Mentor or trainer support provided 3-day basic leader training, additional training specific for the target age group, and supervision. The six authorities conducted a total of 54 groups that included parents of children within the 8-13 age range (mean of 9 groups per authority). Both parents of the targeted child were invited to participate, but most children (94%) were accompanied by one parent. There were generally between 8 and 10 parents per group (mean = 9.5).

Sample Characteristics: The primary caregiver for all the children was a female, and 95% of children lived with a biological parent. The majority (63%) of parents left school before the age of 16 and the mean age at birth of the first child was 21.9. The majority (61%) of families received state benefits and 62% of children received free school lunch. More than half (55%) received a weekly income below the poverty line. Family history of crime was at 35% and depression after the child's first year of life was 61% (up from 38% during the first year). Problems with drugs or alcohol were reported in 28% of families. The majority of child participants were male (67%), and 53% of children were receiving additional help at school. The mean age of the children was 10.3 years.

Measures: Measures were collected at pre- and post-implementation and were administered by the group leaders. Four outcome measure instruments were used. The Eyberg Child Behavior Inventory (ECBI) collected data on child problem behaviors (numbers and frequency). Both the Intensity and Problem scales were used. The Strengths and Difficulties Questionnaire (SDQ) was used to assess the occurrence of behaviors associated with conduct problems. The Beck Depression Inventory (BDI) was used to measure parental depression (severity and symptoms). Finally, the Arnold-O'Leary Parenting Scale was used to measure parenting competencies (laxness, overreactivity, and verbosity).

Analysis: Data were analyzed using paired t-tests, and used an intent-to-treat design. Mediator and moderator analyses were used to interpret results.

Outcomes

Baseline Equivalence and Differential Attrition: Paired pre- and post-test data were available for between 45 and 51% of children depending on the measure. Both these (matched sample data) and the Intent-To-Treat (ITT) data were reported. Differential attrition analyses indicated no significant differences between groups who completed post-intervention assessments compared to those who did not on either demographic or outcome measures. At baseline, mean scores on both the Intensity and Problem scale of the ECBI exceeded the clinical cut-off, with 70% and 82% of children within the clinical range, respectively. On the Total Problem score of the SDQ, 74% of children fell within the 'abnormal' range, and 71% of children exceeded the clinical cut-off on the Impact Supplement score. 78% of parents scored at or above the cut-off for mild depression on the BDI.

Posttest: Child behavior outcomes on the matched sample data indicate a significant reduction from pre-test in both the post-intervention Intensity and Problem scores on the ECBI. There was a 29% reduction in the number of parents reporting their child's behavior in the clinical range (from 71% to 42%). ITT analysis also showed significant reduction in scores. Effect sizes (from ITT) were moderate (.3 on Intensity and .5 on Problem). Total scores on the Strengths and Difficulties Questionnaire indicate a significant post-intervention reduction on the paired t-tests, as well as significant reductions on the conduct problems and hyperactivity subscales. There was also a significant increase in positive social behaviors from pre- to post-test. Parents reported a 23% reduction in children's behavior in the clinical range (from 74% to 51%). Impact scores also decreased significantly. ITT results were also significant, and small effect sizes were reported (.2 on SDQ Total and .1 on SDQ Impact).

Parental depression was significantly improved, as paired t-test results indicate mean scores moving from the 'moderate to severe' range at baseline to the 'mild to moderate' range at post-test. Caregivers' scores exceeding the clinical cut off for depression at baseline dropped from 78% to 40% (a 38% reduction). ITT results were also significant, with medium effect sizes (.4). There was also a significant outcome on the paired t-tests on the parenting skills measure, with significant improvements on both the total score and all subscale scores. ITT results were also significant, and effect sizes were large (.6).

Mediator Analyses: A mediation analysis was conducted to determine the mediating effects of parenting skills on child behavior. Results indicate significant mediation.

Study 27

Evaluation Methodology

Design: Participants were 149 families who were recruited into the study using existing service systems beginning in early 2008, including public health service waiting lists, local schools, community-based agencies, and self-referral. Eligibility criteria included parents rating their child above the clinical cutoff on either the Intensity or Problem subscale of the Eyberg Child Behavior Inventory. Children were between the ages of 32-88 months. From the 149 families recruited, 137 parents participated in the study. Implementation occurred through community-based organizations or Family Resource Centers that provide statutory-funded, individual or group services and support for vulnerable families experiencing a variety of difficulties, including socioeconomic disadvantage, social isolation, mental health issues, substance misuse, community conflict, and domestic violence.

Participants were blindly and randomly allocated on a 2:1 basis to either the parent training intervention (n=103) or a waiting-list control (n=46) group. The assessment was conducted in two waves six months apart.

Sample Characteristics: Approximately one third of parents were single parents, and about two-thirds were classified as at risk of poverty. The mean age of mothers at first birth was 25, and about one-third of parents did not graduate from school. The majority of parent participants had a male child. The majority of families (between 65-68%) rated 2 or higher (out of 6) on socioeconomic disadvantage (variables included employment status, parental status, size of family, parental education, quality of housing, and levels of criminality in the participants' area of residence). Also, the majority of children (51% of controls and 65% of intervention) were at risk for conduct disorder, meaning they had a risk factor score of 2 or more out of 5 (risk factors included single parent, teenage parent, parental depression, family poverty, and parental history of drug use or criminality).

Measures: Measures were collected at baseline and again at six months (three months post-intervention). Four scales were used to measure child behavior: The Eyberg Child Behavior Inventory (ECBI) collected data on child problem behaviors (numbers and frequency). Both the Intensity and Problem scales were used. The Strengths and Difficulties Questionnaire (SDQ) was used to assess child problem behavior and socioemotional well-being (5 subscales relating to emotional symptoms, conduct problems, hyperactivity, peer problems, and prosocial behavior). The Conners Abbreviated Parent Rating Scale was used to measure hyperactive-inattentive behaviors, including restlessness, overactivity, emotional reactivity, and inattention. The Social Competence Scale was used to measure child social functioning, including emotional self-regulation and prosocial behaviors. All child measure instruments have high internal consistency alpha scores.

Two measures were used to assess parental well-being: The Beck Depression Inventory (BDI) was used to measure parental depression (severity and symptoms). The Parenting Stress Index-Short Form was used to obtain a measure of parent stress and functioning.

Observational measures were also collected using the Dyadic Parent-Child Interactive Coding System-Revised. Observers continuously coded behavior based on the frequency of a given behavior during parent-child interactions. Observations were conducted primarily in the participants' homes. Four summary variables were created for analysis, including child problem behavior, child positive behavior, postive parenting, and critical parenting. Observational data was only collected on wave 2 participants (n=80) due to observer training occurring during the first wave of implementation.

For cost analysis data at the 12-month follow-up, parents were asked to completed a Service Utilisation Questionnaire (SUQ) to provide information on their child's use of health, social, and special educational services (services related to childhood conduct problems) during the previous 6-month time period.

Analysis: An intent-to-treat design was used and ANCOVA was used to determine program effectiveness. Parents were nested within parent training groups. Observational data were analyzed using the last observation forward for those who did not complete observations at follow-up. A secondary analysis of observational data was also carried out, excluding only those who were lost to follow-up. High interrater reliability occurred for all variables (ranging from .88 to .97). Effect sizes for the ANCOVA were also calculated. Missing values at follow-up were replaced using multiple imputation.

Outcomes

Implementation Fidelity: Group facilitators received 3 days of training on the intervention and all had previous experience in delivering IYBP in community-based settings. During course delivery, all group facilitators received weekly supervision and support from a certified independent IY trainer and attended weekly meetings to assess progress and address issues. The group sessions were videotaped and randomly reviewed by a certified trainer to evaluate treatment delivery. Implementation fidelity was monitored through facilitator-completed self-evaluation checklists. The checklists indicated that 90% of all the material was covered across the 14 sessions. 76% of the first cohort of participants (n=53) attended at least seven sessions (mean = 10.8), compared to about half (52%) of the second cohort (mean = 6.6).

Baseline Equivalence and Differential Attrition: There was no statistically significant difference between groups at baseline. Analyses conducted on participants who were lost to follow-up (n=12) also showed no statistically significant differences from those who remained in the study, with the exception that children of parents lost to follow-up had statistically significantly higher Social Competency Scale scores. By the 12-month follow-up, 84% of the initial sample remained (n=87) and there were no significant differences between those lost to follow-up and those who remained on socioeconomic or demographic variables, nor were there differences on measures of parenting behavior or well-being. Children of parents lost to follow-up had significantly higher levels of social competence.

Posttest: There were significant program effects on all parent-reported child behavior outcomes at post-test. Additionally, effect sizes on these measures were near medium (SDQ Total = .48) to large (ECBI Intensity = .70; ECBI Problem = .75; Conners Hyperactivity = .92; Social competence = -.83). Problem behavior scores in the intervention group dropped from clinical levels at baseline to within the normal range at follow-up. One of two observed measures (child problem behavior) was found to be significant, favoring the implementation group, and had an effect size of 1.07. Three of the four parental measures were found to be significant at post-test. Results show that there was a significant decrease in the frequency of critical parenting in the intervention group, compared to control group parents, and significant improvements in parental stress levels and parental depression, compared to control group parents. The effect sizes on these measures were small (depression = .39) to medium (stress = .69; critical parenting = .63).

Follow-up: At 12-month follow-up, program effects on child behavior and adjustment seen at 6-months were maintained. Observational data showed a return to baseline levels on child behavior. At 12 months, a significant reduction in problem sibling behavior emerged. Parental well-being and psychosocial functioning improvements seen at 6 months were maintained at the one-year follow-up. Observational data on positive parenting practices, significant at 6-months, also showed maintenance of program effects. There was also a significant positive program effect on the measure of marital conflict, seen both a 6- and 12-months.

Costs of service use: There was a substantial reduction in the use of child services from baseline to 6 months, with maintenance of reductions at 12-month follow-up. Cost of service use amongst the intervention group at 12 months was 60% of that reported at baseline.

Study 28

Evaluation Methodology

Design: Participants for the study were recruited through teachers and counselors at local schools, doctor's offices, mental health professionals and community parent publications. A total of 204 parents called to inquire about the study because their child either had symptoms of, or was diagnosed with, ADHD. After an initial phone conversation with a researcher, 156 parents completed a structured interview with a clinician to assess for ADHD and ODD (oppositional defiant disorder). Of the 103 children who met the inclusion criteria, 99 (96%) attended an intake and were randomly assigned to a treatment condition (n=49) or a waitlist control condition (n=50). A total of five families (3 from control and 2 from treatment) dropped out during the study (5% attrition rate). Data were gathered at baseline and immediately follow the intervention. Post-test data were gathered on 96 of the 99 families.

Intervention: The IY program has been described previously in detail and comprises 20 weekly, 2-hour parent training sessions and separate child training sessions. During this study, the newest version of the curriculum (revised in 2008) was used. New material focusing on coaching, predictable schedules, emotional regulation strategies and problem-solving strategies were included. New vignettes demonstrating effective parental responses to child behavior were used with parents.

Sample Characteristics: A majority of the children were male (75%) with an average age of 64 months. 73% were White; mothers in the sample had an average of 15.6 years of education; fathers had an average of 15 years of education.

Measures: Assessments consisted of previously used measures including 27 scales completed by parents and teachers and 13 scales completed by blinded observers:

Parent reports of child behavior problems were collected using the CBCL, ECBI Intensity and Total Problems scales; Conners' Parent Rating Scale - Revised (a 57-item instrument that assesses ADHD; alpha=.91-.93); and, Social Competence Scale. Parent reports of parenting behavior were collected using the Parenting Practices Inventory which measures appropriate discipline, praise and incentives, monitoring, harsh and inconsistent discipline, and physical punishment. Teacher reports of child behavior were collected using the Teacher Report Form and Conners' Teacher Rating Scale - Revised. Independent observations of parent-child behavior were conducted in a lab and measured using the Dyadic Parent-Child Interactive Coding System - Revised. Independent observations of child behavior in the classroom were conducted by blinded observers using the Coder Observation of Child Adaptation - Revised. Child problem-solving and feelings assessment was conducted using Wally Problem Solving Test and Wally Feelings Test, though it was unclear who conducted these assessments with the children.

Analysis: Data were analyzed using repeated measures analyses of variance. Results for condition-by-time interactions were presented, as were differences in posttest means.

Intention-to-treat: The study used data on all families who provided valid assessments, regardless of dose of intervention received.

Outcomes

Implementation Fidelity: Parent satisfaction questionnaires were completed by parents at the end of the program. Both mother (92%) and father (85%) attendance at sessions was high. Fidelity was also monitored by videotaping groups and reviewing them during weekly supervision. Protocol checklists were completed after each session by group leaders.

Baseline Equivalence: No statistically significant differences between conditions were observed for any demographic variables, however no test for significance was conducted for between-group differences on outcome variables at baseline.

Differential Attrition: Attrition was minimal in the study, though no information about differential attrition was provided.

Posttest: Tests of significance are presented in two forms. First, the authors test for mean differences across groups at posttest (with no baseline controls). Second, using repeated measures ANOVA, baseline controls, and condition-by-time interactions, they test group differences in the change in outcomes from pretest to posttest. Since the two tests often give different results, both are summarized.

Mean differences across posttest (between group differences):

Of the 27 measures assessing child behavior from the perspectives of mother, father and teacher, 6 mother-report measures and 2 father report measures revealed significant posttreatment between group differences. The six mother-report measures were CBCL aggression, CPRS hyperactive, ECBI intensity, ECBI problems, emotional regulation and social competence. The two father-report measures were emotional regulation and social competence.

Of the 13 measures conducted by observers in lab and school settings, 4 measures revealed significant posttreatment between group differences. The four measures were praise and coaching during free play and praise and coaching during task-driven observations. None of the school peer observations or observations of child behaviors was significant for between group differences at posttreatment.

Of the 10 measures completed by parents rating their own behavior, none of the measures was significant for between group differences at posttreatment.

Of the two measures that assessed child problem solving and feeling language, one (problem solving) was significant for between group differences at posttreatment.

Condition x Time differences

Of the 27 measures assessing child behavior from the perspectives of mother, father and teacher, 19 measures revealed significant condition x time effects.

Of the 13 measures conducted by observers in lab and school settings, 5 measures revealed significant condition x time effects.

Of the 10 measures completed by parents rating their own behavior, 4 measures revealed significant condition x time effects. All four were mother-report measures. None of the father-report measures were significant.

Of the two measures that assessed child problem solving and feeling language, both were significant for condition x time effects.

Study 29

Evaluation Methodology

Design: Using a randomized control design, 14 elementary schools in the Seattle area were first matched on variables such as size, geographic location and demographics of the children and then matched pairs were randomly assigned to intervention and control conditions. Parents of all children in kindergarten were invited to participate in the study and 77% (n=1,152) of families completed consent forms. A subsample of moderate- to high-risk indicated students (n=433) were selected from each classroom based on parent or teacher reports of behavior problems and baseline (n=340, 78%), mid-program at year one (n=293, 67%) and posttest at the completion of the two-year intervention (n=252) data were gathered from this subsample. Within intervention schools, half of the indicated sample from each classroom was randomly assigned to receive the 2-year classroom intervention only (CR), the other half also received the parent intervention (PT+CR). The control schools received neither classroom nor parent intervention.

Sample Characteristics: Children were on average 67 months old and 59% were male, 38% were White, 14% were African American and 20% were Hispanic. The sample was a disadvantaged one with a high proportion living in poverty and 51% receiving financial assistance. 26% of the mothers had not completed high school and 26% reported symptoms of depression.

Measures: Assessments consisted of previously used measures and were gathered through home observations (using three scales from DPICS-R and four scales from CII), parent report (6 measures) and teacher report (3 measures).

Analysis: A mixed-design analysis of covariance with three conditions (control, CR, PT+CR), using baseline scores as covariates and the posttest and follow-up scores as repeated measures was used. Since schools were randomized to the intervention and control conditions, the analysis should have adjusted for clustering within schools and measured the intervention at the school level. However, the analysis did not make these adjustments.

Intention-to-treat: Data were gathered on all families regardless of dose of intervention received.

Outcomes

Implementation Fidelity: Group leaders completed protocol checklists to record their fidelity to the program. Attendance rates were also measured for the PT+CR group, and 28% of the parents did not come to any of the sessions.

Baseline Equivalence: No significant differences at baseline were found on any parent or child outcome measures across the three conditions. Of the 35 demographic variables measured for children and parents, only three were significantly different at baseline between conditions. There were significantly more Caucasians and fewer Asian families in the control condition compared to the two intervention conditions. The mothers in the control condition reported significantly more depressive symptoms that the mothers in the PT+CR condition.

Differential Attrition: No significant differences among conditions were found for percentage of families who completed the study. Further, families who dropped out of the study were not significantly different from those who did not drop on any demographic or outcome variables.

Mid-program assessment and posttest: Researchers presented results that looked at the changes from baseline to posttest. Main effect of condition was found on 10 of the 16 measures. Planned contrasts revealed only one significant difference between control and CR (teacher report of externalizing); however when comparing control to PT+CR, 11 of the 16 measures were significantly different at the p<.05 level. Four measures were significantly different when comparing CR to PT+CR. Few of the tests showed significant time-by-condition interactions, which appears to indicate that posttest differences were similar to mid-program differences.

Study 30

Evaluation Methodology

Design: This study combined assessments from three cohorts of Head Start families who had participated in previous prevention studies (see Study 5). The three cohorts were enrolled in 1993, 1994 and 1997. The combined sample was studied using a quasi-experimental design where Head Start centers were first matched on several variables and then randomly assigned to a treatment or control condition.

Previous studies have used this same sample and therefore no information regarding sampling procedures or attrition was provided in this study. There were a total of 882 participants in the sample, of which 607 were in the intervention group.

Within the total sample, 58% of mothers were identified as "indicated mothers" based on independent observations of them making 10 or more critical statements to the child during the home observation. 28% of the sample of children were "indicated children" identified by teachers using two different measures, and in the total sample, 34% of families had neither an indicated mother or child, 42% had an indicated mother only, 8% had an indicated child only, and 16% had both an indicated child and mother.

Intervention: The IY program has been described previously in detail, however in this study, the first two cohorts' teachers were provided with two days of training and parents were provided with weekly 2.5 hour sessions for 8-9 weeks. In the third cohort, teachers received 6 days of training and parents received 12 weekly 2-hour sessions.

Sample Characteristics: A majority of the children were under the age of 5 (86%) and 53% were male, 51% were White, 19% were African American and 10% were Hispanic. 84% of families reported a gross income of $20,000 or less.

Measures: Assessments consisted of previously used measures including the ECBI, DPICS-R and CII all of which are described in detail in previous write-ups.

The purpose of this study was to examine parent and child moderators of outcomes, program engagement effects, and predictors of engagement and therefore program engagement, program benefit and program attenders were also measured. Program engagement was assessed by combining measures of a) the number of parent training sessions attended by the mother, b) the percentage of homework completed, and c) the group leader's rating of the mother's engagement in group discussion. Program benefit was identified as a 30% reduction in the number of critical statements made by the mother between pre and post intervention.

Analysis: Program effectiveness was modeled using structural equation modeling. The models focus on the combined influence of indication, engagement, and intervention rather than the intervention alone.

Intention-to-treat: No information provided.

Outcomes

Implementation Fidelity: No information provided in this study.

Baseline Equivalence and Diferential Attrition: No information provided in this study.

Results:

Mothers of indicated and non-indicated children had similar rates of attendance. Further, indicated and non-indicated mothers also had similar rates of attendance. Program engagement and attendance in turn influenced child outcomes; specifically, program engagement reduced conduct problems in children as detected by observers (not mothers) and increased prosocial behaviors.

Figures 3-6 supply for the most straightforward interpretation of the combined influences of condition, baseline indication, and engagement. Figures 3 and 4 focus largely on differences among indicated children:

• Conduct problems dropped among indicated children of intervention mothers who attended, but not among indicated children of intervention mothers who did not attend, indicated controls, or non-indicated controls.
• Prosocial behavior increased among indicated children, regardless of attendance of the intervention mother, but not among indicated or non-indicated controls.

Figures 4 and 5 focus largely on differences among indicated mothers:

• Conduct problems dropped among children of indicated intervention mothers who attended but not among indicated intervention mothers who did not attend, children of indicated control mothers, and children of non-indicated control mothers.
• Prosocial behavior increased among children of indicated intervention mothers who attended but not among indicated intervention mothers who did not attend, children of indicated control mothers, and children of non-indicated control mothers.

In summary, indicated children and children of indicated mothers appeared to benefit most from the intervention, particularly when the mothers attended intervention sessions.

Study 31

Evaluation Methodology

Design: This study combined assessments from three cohorts of Head Start families who had participated in previous prevention studies in the Puget Sound and Seattle area (see Studies 5 and 30). The three cohorts were enrolled in 1993, 1994 and 1997. The combined sample was studied using a quasi-experimental design where Head Start centers were first matched on several variables and randomly assigned to a treatment or control condition. Because parents within schools were not randomized, the intervention and control groups differed on several baseline characteristics, making the study a quasi-experimental design rather than a randomized controlled trial.

Previous studies have used this same sample and therefore no information regarding sampling procedures or attrition was provided in this study. There were a total of 14 Head Start centers in the treatment condition and 9 in the control condition. In this particular study, the sample size was 634 families who provided data at baseline and post-intervention, and who reported their ethnicity as African American, Asian, Caucasian or Hispanic. Assessments were conducted at baseline, post-intervention and one-year follow-up. Data collected included child and parent behavior as measured by parent interview and independent, blinded home observations of the parent-child interactions. The study reported no information on attrition from baseline to posttest but noted that 78.3% of the posttest control group completed the follow-up and 73.1% of the posttest intervention group completed the follow-up.

Sample Characteristics: The children in the study included 343 boys (54%) and 291 girls (46%) with a mean age of 56 months. The mothers had an average age of 30.5 years and 54% were single mothers. Forty-two percent of the children represented minority groups based on mother report, and 19% of the sample was African American, 11% Hispanic and 12% Asian.

Measures: Assessments consisted of previously used measures and were gathered through home observations (using three scales from DPICS-R and four scales from CII), parent report (5 measures using ECBI, CBCL, P-COMP, PPI and INVOLVE) and teacher report (1 measure regarding parent-teacher involvement).

Analysis: Specific to this study that focuses on the relationship between ethnicity and treatment response, hierarchical linear modeling with time nested within individuals was used to control for center effects when examining ethnicity effects.

Intention-to-treat: All participants were included in the sample regardless of dose of intervention received.

Outcomes

Implementation Fidelity: Program developer provided ongoing supervision in content and techniques to parenting coaches on a weekly basis and Head Start coleaders on a monthly basis. Weekly records were kept by group leaders and examined by program developer. Direct observation of groups was also conducted by the program developer. However, no figures on fidelity were reported.

Baseline Equivalence: Several significant differences were found between intervention and control groups on demographic and outcome variables at baseline. Mothers in the intervention group reported significantly lower incomes, greater histories of mental illness and higher scores for depression and anger than control mothers. Further, mothers in the intervention group reported using more harsh and inconsistent discipline than control mothers, and this difference was confirmed by independent observers. Finally, intervention children exhibited more negative behaviors and scored higher on the ECBI than control children. When groups were examined by ethnicity of the child, significant group differences were found for single parent status (Hispanic lowest at 34%, African American highest at 76%); education (59% Hispanic mothers versus 14% African American mothers under 12 years of education); criminal history and child protective services involvement.

Differential Attrition: While dropouts in the control group were more likely to exhibit higher levels of family risk factors and have children with more behavior problems than dropouts in the intervention group, the study reported no significant condition x attrition interactions on any of the dependent variables at post-treatment. Also, the authors state that there was no "differential dropout at postassessment or 1-year follow-up by minority status in either of the original studies." This study does point out that there was differential enrollment in the study with 28% of minority and 17% of Caucasian mothers choosing not to participate in baseline assessments.

Results: The focus of this study was to evaluate differences in responsiveness to treatment among ethnic groups. Based on the analysis conducted on the 51 ethnicity effects (comparisons of three minority groups to whites for 17 outcome variables), only three reached statistical significance. Since these significant results could easily occur by chance, they indicate that the ethnic groups responded similarly to treatment.

Estimates across all race and ethnic groups summarized the overall program benefits: The intervention group did significantly better at posttest than the control group on 12 of 17 outcomes, and did significantly better at follow-up on 6 of 17 outcomes. The findings also noted clinically significant improvement at posttest on three outcomes: parent critical behavior, child problem behavior, and child positive affect. However, only the improvement in child positive affect persisted through the follow-up.

Study 32

Evaluation Methodology

Design: The study combined data from six randomized clinical trials conducted with 21 separate cohorts of 3- to 8-year-old children over the previous 20 years. The resulting sample included 514 families who participated in IY treatment-outcome research on oppositional defiant disorder and conduct disorder. A total of 317 families received parent training only (PT), 60 received child training only (CT), 38 received PT+CT, 24 received PT+Teacher training (TT), 23 received CT+TT, and 25 received PT+TT+CT. Twenty seven families were assigned to a waitlist control condition and received PT after their post-assessment.

Sample Characteristics: Children were on average 5.4 years old and 78% were male, 88.5% were White, 4.8% were African American and 3.9% were Hispanic.

Measures: Assessments consisted of previously used measures and were gathered through home observations (DPICS-R) and mother- report (ECBI, CBCL and PSI).

The purpose of this study was to evaluate predictors/moderators of outcome and therefore the following additional measures were examined: Parenting Stress Index, Dyadic Adjustment Scale, Beck Depression Inventory, parental substance abuse, comorbid child psychopathology, additional familial predictors (maternal education level, maternal age, maternal relationship status, social class, family size) and additional child predictors (children's age, sex, conduct problems).

The secondary purpose of this study was to evaluate mediators of outcome and therefore the following additional measures were examined: parenting style (as measured through home observations using DPICS-R and self-reported Daily Discipline Inventory), and treatment dose.

Analysis: Data were analyzed using latent growth curve models.

Intention-to-treat: Data were gathered on all families regardless of dose of intervention received. Full information maximum likelihood estimation was used when data were missing. However, analyses were also conducted in which families with missing data were dropped and the pattern of results from the latent growth curve models remained unchanged.

Outcomes

Implementation Fidelity: No information provided in this study; however details can be found in other studies from which this study uses data.

Baseline Equivalence and Differential Attrition: No information provided in this study.

Results:

The following significant predictive relationships were found:

• According to mother-report models, better child outcomes were observed when more treatment components (i.e., PT, CT, TT) were delivered to families. However, this relationship was not found in behavior observation models.

The following significant moderational relationships were found:

• According to mother-report models, when marital satisfaction is low, children of mothers who received PT improved more at one-year follow-up than children who received an intervention without PT. According to mother-report of child behavior, more positive treatment responses were observed in children with comorbid symptoms of anxiety/depression.
• According to home observation models, maternal depression, social class, paternal substance abuse, marital status and comorbid attention problems moderated treatment response. In most cases, including PT and/or CT was more effective, except in the case of comorbid attention problems, where including TT was more effective than not including TT.

The following significant mediators of outcome were found:

• According to mother-report models, both self-reported verbal criticism and harsh parenting predicted and mediated outcomes. This means that children of mothers who scored low on harsh parenting and verbal criticism at baseline and still improved during treatment had best treatment responses. Similar findings were revealed for verbal criticism and ineffective parenting as measured by home observations.

Study 33

This study was presented in an article that described the implementation and evaluation of three programs delivered in Birmingham, UK. A parallel randomized, controlled trial of Incredible Years BASIC was conducted using a pre-post test design with the parents of 161 children between the ages of 3 and 4 years who were showing symptoms of a conduct disorder as evidenced by reaching the "high-need" threshold on the parent-completed Strengths and Difficulties Questionnaire. Participants were identified through referrals from other agencies, child centers or self-referrals. Control participants received services as usual.

The sample comprised 101 males and 60 females with a mean age of 44 months at baseline; 50% of the families relied on benefits as their main source of income. The sample was randomly assigned to the intervention (n=110) and control (n=51). Baseline data were gathered prior to randomization on all 161 children and at immediate posttest on 147 children (attrition rate of 8.7%). Values were imputed for missing data.

Child outcomes were measured using two parent-completed assessments, the Strengths and Difficulties Questionnaire (7 measures) and the Eyberg Child Behavior Inventory (2 measures). Parent outcomes were measured using the Arnold and O'Leary Parenting Scale (4 measures).

Because several differences in outcome variables were found at baseline, analysis was conducted using analysis of covariance controlling for children's baseline scores, age and sex of the child, and center from which the family was recruited. The study adhered to the intent to treat principle and included results from participants regardless of dose received or dropout status.

Of the 13 outcome measures analyzed, 8 were significantly different between the intervention and control groups at posttest. Specifically:

• Four of the seven measures of the Strengths and Difficulties Questionnaire (conduct problems, peer problems, total difficulties, impact on others) were significantly improved in the intervention group compared to the control.
• One of the two measures of the Eyberg Child Behavior Inventory (intensity scale) was significantly improved in the intervention group compared to the control.
• Three of the four parenting measures assessed using the Arnold and O'Leary Parenting Scale (total, verbosity, and overreact) were significantly improved in the intervention group compared to the control.

Effect sizes reported as Cohen's d were small to medium and ranged between 0.31 and 0.50.

No follow-up data were gathered and the intervention was provided to the control group upon completion of the posttest.

Study 34

Evaluation Methodology

Design: Participants for this study were from the 1993 and 1994 cohorts of seven Head Start centers in Seattle described in Study 5. There were 481 participating families (335 in the intervention condition and 146 in the control condition) who had answered a key question about their prior history of report for child maltreatment. All children were attending Head Start, and the parents of the children in the intervention condition attended an 8-9 week version of the Incredible Years parent-training program. Intervention participants attended an average of 5.61 2-hour sessions. Nearly 79% completed at least half of the sessions. The IY program was delivered primarily by trained Family Support Workers in the Head Start programs, as well as a few parents who emerged as strong group leaders. Program facilitators completed 4 days of training workshops and also received ongoing weekly supervision. Facilitators co-lead their first parenting group session with a staff member from the University of Washington Parenting Clinic.

Sample Characteristics: The majority of families (56.3%) were single parent (mother) and nearly all children (95%) lived with their biological mother (the remaining 5% lived with other relatives). There were slightly more boys than girls in the sample (248 vs. 233, respectively) and the mean age of the children was 4.7 years. About 40% of the sample represented ethnic minorities, including African-American, Hispanic/Latino, Asian American and Native American. The majority of families (86.2%) were receiving financial aid, and the median family income was approximately $10,000 per year. Among the families that completed the question "have you ever been reported for child abuse (or have you ever abused your child)?", 20.8% answered affirmatively.

Measures: Measures were collected at baseline (fall), with post assessments collected in the spring. Additional follow-up was conducted 12-18 months after the initial baseline assessment when the children were in kindergarten. Assessments included parent interviews, questionnaires, and in-home observations conducted by raters blinded to condition. Parent and child behaviors were measured using the Dyadic Parent-Child Interaction Coding System-Revised (in-home observations), a Coder Impression Inventory (completed after home observations), and the Eyberg Child Behavior Inventory (completed by parents). Reliability checks on the observer measures were completed weekly and alphas ranged from .67 - .88.

Analysis: Longitudinal analyses only included participants with data at post or follow-up. In the intervention condition, this still included individuals who did and did not participate in the IY intervention, consistent with an intent-to-treat analytic approach. Analysis was completed using multi-level, random-effects modeling. Analyses were stratified by reported history of maltreatment. Interaction terms were also used to assess reported history of maltreatment as a potential moderator of intervention effects. Post and follow-up time points were nested within individuals, and individuals were nested within Head Start sites. Dose response relationships were also examined using exploratory analyses.

Outcomes

Program Fidelity: Program delivery was closely monitored to ensure appropriate delivery of the intervention, including program content, skill-building methods, and use of a collaborative approach with participating parents.

Baseline Equivalence and Differential Attrition: There was a loss of 21% (103 families) from baseline to post-intervention follow-up. There were no significant differences between conditions on rate of attrition (23% of intervention and 17% of control families), nor were there significant differences between conditions with respect to primary outcome measure, child welfare contact. There were, however, significant baseline differences between groups on some of the observation measures, where mothers who reported a history of child welfare contact issued more critical statements, were less nurturing/supportive to children, and had less competent discipline efforts than mothers who did not report child welfare contact. These differences were to be expected.

Posttest: At both post-intervention assessments, parents in the intervention condition with no history of child welfare contact saw significant improvements in several positive parenting characteristics, compared to control condition parents, including: praise/positive affect/physical positives, nurturing/supportive parenting, and discipline competence. There were also significant reductions among this group in negative parenting practices, compared to their counterparts in the control condition, including total critical statements and harsh/critical parenting. Reported effect sizes were small to moderate. Among parents who reported a history of child welfare contact, the pattern of effect size estimates was consistent with the intervention parents with no history of child maltreatment. For some variables, the estimated effect was larger for intervention families with a history of child maltreatment, but not significant. This was potentially affected by differences in the sample sizes for the two groups. On measures of child behavior, the children of intervention parents who did not report a history of child welfare contact showed reductions in negative nonverbal affect and improvements in positive affect. However, there were no significant child behavior outcomes among the children of intervention parents who did have a history of child maltreatment.

Dosage outcomes indicated significant improvements across all parenting indicators among intervention families with higher attendance. Number of sessions attended was not associated with improvements in child indicators.

Study 35

Evaluation Methodology

Design:

Recruitment /Sample size: Forty-three private-practice pediatric groups and federally qualified health centers in eastern Massachusetts were invited to participate in the study, and 12 agreed to participate (though the study reports that 11 pediatric practices-seven private practices and four federally qualified health centers-participated in the trial). Participating practices conducted a behavioral screening for all children between 22 and 42 months of age using the Infant-Toddler Social-Emotional Assessment Scale. Parents of children scoring at or above the 80th percentile of the screener were eligible. Parents were excluded if they could not speak English or Spanish well enough to participate, or if they reported that their child had a diagnosis of pervasive developmental disorder or global developmental delay. Eligible parents were asked for permission to be contacted by study staff. Of 830 parents who consented to contact, 345 (42%) enrolled in the study. Of those enrolled, 290 (84%) completed measures at pretest and were assigned to conditions. However, 17 subjects dropped out after assignment and were not included in any of the assessments or the analysis.

Study type/Randomization/Intervention: With the aim of assigning 6 to 12 parents each to a parent-training group and a waitlist control group, participants were randomly assigned to condition once enough participants were enrolled at a practice. If fewer than six participants could attend on the selected evening, control group participants were randomly reassigned to that intervention group. In six practices (three private practice and three health centers) too few participants were identified within three months to constitute both an intervention and control group, thus all participants at these practices were assigned to the intervention group creating a third condition: a non-random intervention group. With the loss of 17 subjects immediately after assignment, 150 participants were randomly assigned to intervention (n=89) and control (n=61) groups, and 123 were assigned to the non-random intervention group. Parents in intervention groups completed an abbreviated, 10-week version of the Incredible Years program for parents, meeting for two hours per week for 10 weeks, most often within the pediatric office.

Assessment/Attrition: Parent-report questionnaires were completed at baseline, posttest (immediately upon program completion), and six-month and one-year follow-ups. Parent-child interactions were videotaped (to be coded later) at baseline, posttest, and one-year follow-up. Attrition rates are not provided for each assessment, but rather for the number of participants who were followed up. Among the non-random intervention group, 41% did not provide follow-up data. Attrition rates for the randomized groups were 19% for the intervention, and 18% for the control groups. There is no information on the initial loss of 17 subjects.

Sample characteristics:

Children included in the study were primarily male (62%). Seventy-four percent were White, 12% African American, 1% Asian, and 12% were reported as having a race of Other. With regard to ethnicity, 18% were Hispanic. Participating parents were primarily female (96%), and 70% were married. Thirty-four percent of parents reported an education level of high school or less, 19% some college, 27% a college degree, and 19% a graduate degree. One-third of participants (33%) reported a family income greater than or equal to $100,000, 24% reported $50,000 to $99,999, 17% reported $20,000 to $49,999, and 26% reported less than $20,000.

Measures:

Validity of measurements: All measures and scales have been used in prior published work and reliability and validity has been established.

Primary outcomes: The 30-item Parenting Scale was self reported by parents and used to assess negative parental discipline styles.

The 36-item Early Childhood Behavior Inventory (ECBI) was reported by parents and used to assess the presence and intensity of child disruptive behaviors, and includes both Problem and Intensity sub-scales.

Twenty-minute observations of parent-child interactions during standardized tasks (including free play, problem solving, and behavioral inhibition) were conducted at each pediatric practice and video recorded. Videotaped sessions were scored by independent, blinded coders using a 75-item Coder Impression Inventory (CII). Coded interactions provided measures of negative parenting, child disruptive behavior, and an overall parent-child interaction scale (corrected). There were some problems with the measures, however. Poor videotape quality prevented use of more detailed codes and measures. Also, because scores approached the lowest values at posttest and follow-up, the CII measures were skewed. The corrected parent-child interaction scale consists of the subset of items with the least skew.

Analysis:

Statistical methods/baseline control/correct unit of analysis: The analysis used a mixed-effects regression model, including demographic variables that differed among conditions at pretest as covariates and accounting for clustering of time points within participants and for participants within pediatric practices. Analyses of main effects included a categorical study condition variable and an ordinal point variable for time to test whether standardized mean differences in outcomes at follow-up significantly differed between conditions.

The study reports results from an intent-to-treat analysis based on a method of multiple imputation-to include running primary analyses across 10 imputed data sets, adjusting for additional variance across imputations.

Intention-to-treat: The study appears to have attempted to assess all subjects and uses an intent-to-treat analysis relying on a process of multiple imputation.

Outcomes:

Implementation fidelity: All sessions were videotaped, and three from each 10-week series were randomly chosen and coded for content, delivery, and degree of parent participation. Session checklists and video reviews indicate that more than 90% of content and delivery elements of the program protocol were followed throughout all groups. Additionally, parents completed 81% of assigned home activities. However, among participants randomly assigned to the intervention group, 20% attended less than three group sessions, and 41% of participants in the non-random intervention group attended less than three sessions.

Baseline Equivalence: There were no differences among conditions on outcome variables at baseline. For participants in randomly assigned groups, there were no differences in demographic variables at baseline; however, participants in the non-random intervention group were more likely to report minority race/ethnicity, lower levels of education, lower family income, and being a single parent.

Differential attrition: A stepwise logistic regression found that some demographic (but not outcome) variables predicted incomplete data: marital status, parent age, child age, and non-White or Hispanic race/ethnicity. The lack of information on the initial loss of 17 subjects was not examined for differential attrition.

Post-test and long-term effects:

Parent-report questionnaires: At posttest, six-month follow-up, and one-year follow-up, participants in both the randomly assigned and non-randomly assigned intervention groups had significantly greater reductions than participants in the waitlist control group in all parent-reports: the Parenting Scale, the ECBI Problem Scale, and the ECBI Intensity Scale.

Videotaped parent-child interactions: At posttest, participants in the randomly assigned intervention group had significantly greater decreases in coder-rated negative parenting, child disruptive behavior, and negative parent-child interaction (corrected) than participants in the waitlist control group. At one-year follow-up (interactions were not videotaped at six-month follow-up), only the negative parent-child interaction rating (corrected) was significantly different between the two groups.

At posttest and one-year follow-up (interactions were not videotaped at six-month follow-up), participants in the non-randomly assigned intervention group had significantly greater decreases in coder-rated negative parent-child interaction only.

Study 36

Evaluation Methodology

Design: This meta-analysis of the Incredible Years Parent Training coded effect sizes and measured design characteristics for 50 studies. The inclusion criteria included: (a) effects of the program were examined immediately after intervention; (b) the effectiveness was examined by comparing an intervention group to a comparison group; (c) the study reported at least one quantitative measure of disruptive or prosocial child behavior, which was measured equally among participants; and (d) sufficient empirical data was reported to enable the calculation of standardized mean difference effect sizes or standardized mean difference effect sizes were reported in text.

The 50 studies came from a search that initially identified 231 citations for review but eliminated all but 39 for not meeting eligibility requirements. Given that some reported on multiple studies and that some overlapped, the 39 citations produced 50 studies for analysis. These 50 studies included 4745 participants.

Sample: Twenty-two studies (44%) were identified as treatment studies, 12 studies (24%) as selective prevention, 11 studies (22%) as indicated prevention, and 5 studies (10%) could not be classified. Thirty-seven studies (74%) were considered as having used the standard program rather than a variant. In 17 studies (34%) the program's developer, Dr. Webster-Stratton, was an author or co-author.

Assignment was random in 28 studies (56%), random after blocking or matching in 13 studies (26%), and non-random in eight studies (16%). The intervention condition was compared to a waiting list condition in 24 studies (48%). The comparison group received nothing in 17 studies (34%) and received an alternative treatment in seven studies (14%).

Measures: Study characteristics were coded by six coders using a detailed coding scheme. A subsample of studies coded independently by two coders showed high intercoder reliability (average correlation = .96 for continuous variables and average kappa = .92 for categorical variables).

The key measure of effect size, Cohen's d, was calculated for five outcomes: 1) disruptive child behavior, 2) prosocial child behavior, 3) parent reports of child behavior, 4) teacher reports of child behavior, and 5) observer reports of child behavior. Multiple effects sizes for the same outcome in a single study were averaged. The first two outcomes appear to be measured across diverse informants, while the last three outcomes appear to be measured across multiple outcomes.

Measures of study characteristics served as predictors. These included measures of study context, intervention characteristics, child characteristics, family characteristics, and methodological features.

Analysis: The analysis examined effects sizes for each of the five outcomes separately. The authors state, "To avoid double/triple counting of participants and samples contributing too much to the effect size mean, the number of participants in each sample was divided by the number of occasions that this sample was included in the meta-analysis for standard errors and inverse variance weights."

Outcomes

The overall weighted effect sizes for the five outcomes are as follows:

• disruptive behavior: .27 (p < .001) across the set of 50 studies,
• prosocial behavior: .23 (p < .001) across the set of 26 studies,
• parent reports of child behavior: .30 (p < .001) across the set of 49 studies,
• teacher reports of child behavior: .13 (p = .001) across the set of 25 studies,
• observer reports of child behavior: .37 (p < .001) across the set of 23 studies.

When the calculations excluded studies using alternative treatments to more precisely compare the program to no treatment, effect sizes changed little. A measure of study rigor had no relationship with effect size for any of the outcomes. Involvement of the developer in the study also did not influence effect size.

Only for parent reports did the effect sizes differ significantly across studies (i.e., reflect more than randomness). Initial severity of child problem behavior proved strongest in increasing the effect size, and reflects the greater success in treatment than prevention studies. The variation in effect sizes for this outcome had no relationship to use of the standard program or a variant. Although not a significant predictor of effect size, the number of sessions attended by parents had large positive effects.

Study 37

Evaluation Methodology

Design: Parents of children between the ages of 3 and 8 years referred for antisocial behavior to mental health services in four local clinics in South London were recruited for the study. Using a permuted block design, in each center during a three-month period, all eligible referrals were allocated to one condition (intervention or waitlist control). Of the 430 referrals, 32.8% (n=141) were enrolled in the study (intervention n=90; control n=51). Among those that did not participate in the study (n=289), 62 (21.5%) refused, 67 (23%) could not be contacted, 124 (43%) were ineligible and the rest stated they no longer needed services. Attrition was 22% at posttest and control participants were provided the intervention after posttest assessments were conducted.

Additional follow-up data were gathered, from intervention participants only at 1-year post-intervention . Of the 73 intervention families who remained at posttest, 59 (81%) were contacted at 1-year follow-up. At the 7-9 year follow-up, 94 families were allocated to parent training (73 initially allocated and 21 from the wait-list control group who received the intervention after six months) and 26 to usual management (10 originally allocated and 16 who were randomly allocated after six months from the wait-list control condition).

Sample Characteristics: The average age of the sample of children was 5.6 years and 71.5% of the sample was male. The study reported that most families were poor and disadvantaged and that the children had serious behavior problems, a mean antisocial behavioral severity above the 97th percentile.

Measures: Nine measures of child behavior and one measure of parent behavior were gathered. Parents and interviewers were blind to the condition when baseline measures were collected but not later. Assessments consisted of previously used measures and were gathered through parent-interview (parent account of child symptoms that measures conduct problems and hyperactivity); questionnaires (Strength and Difficulties questionnaire that measures conduct problems and total deviance; Child Behavior Checklist that measures externalizing problems and total problems; Parent Daily Report that measures total number of problems per day; and, Parent Defined Problems questionnaire that reports a mean score across three parent-identified problems); child-interview (diagnosis of oppositional defiant disorder); and, video-taped observation of parent-behavior for the use of inappropriate commands among a random sample of 20 video-taped observations.

Analysis: Analysis of variance over time was used with age and sex as covariates and controls for baseline outcomes. The study complied with the intent-to-treat principle by analyzing the full sample regardless of dose received. In Scott et al. (2001), one analysis examined subjects with complete data at both pretest and posttest, while another analysis (labeled intent to treat) examined all subjects with baseline values carried forward for those with missing posttest data.

Paired t-tests were used to examine changes in the intervention condition from baseline to follow-up and posttest to follow-up. However, no comparisons could be made with the control condition.

OUTCOMES

Implementation Fidelity: Mean attendance was 9.1 sessions and 18% of the families who completed the study attended four or fewer times. The study reported that intervention sessions were videotaped and therapists received weekly supervision to ensure adherence to the manual. No further information was provided.

Baseline Equivalence: The study reported that the conditions did not differ significantly on any variables at baseline, however it appears that the tests compared three groups - intervention with posttest data, control with posttest data, and subjects from either group lost to follow-up - rather than all randomized control and intervention subjects. Further, it is unclear if the study examined all outcome variables at baseline.

Differential Attrition: Although attrition rates differed by condition (intervention 19%; control 27.5%), the study did not provide specific information on differential attrition at posttest. The baseline comparisons, which separated subjects lost to follow-up from others, might provide some evidence, and the authors say there were no differences across groups. Again, however, not all outcome variables were examined.

At 1-year follow-up, compared to the intervention participants who could not be contacted, those who were contacted had significantly lower baseline levels of antisocial behavior and the parent was significantly less likely to be an ethnic minority.

At 7-9 year follow-up, the study reported that "dropout was not significantly associated with any pre-intervention variable."

Posttest: Using an intent-to-treat sample with estimated missing data (baseline outcome carried forward), all nine child-behavior measures were significantly different between the intervention and control at posttest in the expected direction. The parent-behavior measure coded using video-taped observation for the subsample of 20 parents was also significantly different between conditions with intervention parents demonstrating a higher ratio of praise to inappropriate commands.

Effect sizes were reported as Cohen's d and ranged between small-medium (0.31) and large (0.92).

Scott (2005) reported on long-term changes among intervention subjects only. Among the intervention group at 1-year follow-up, aside from hyperactivity and peer-problem measures, all other measures improved significantly from baseline. However, these changes were not significant when comparing posttest to 1-year follow-up measures, suggesting that most improvements occurring at post test were maintained. Only one measure, parent-defined problems, had significantly improved from posttest to 1-year follow-up.

At 7-9 year follow-up, intervention participants had significantly improved scores, compared to a usual management group, for oppositional symptoms, antisocial personality traits, and antisocial behavior. Effect sizes were medium to large: 0.91 for oppositional defiant symptoms, 0.70 for antisocial personality traits (Scott et al., 2014).

Study 38

Evaluation Methodology

Design: Recruitment. The study recruited parents and children aged 5-7 from an inner city London authority with an ethnically diverse and socioeconomically deprived population and from an authority in Southwest England with a largely white and middle-class population. Recruitment occurred in 11 primary schools in London and 56 primary schools in the South West. There were 3675 children screened for antisocial behavior by inviting parents and teachers to complete a short set of questions about child behavior. The screening used measures of conduct problems and oppositional defiant disorder, selecting a cutoff equal to a level of conduct problems reached by the highest 15% of a national population. In total, 2665 children had parent or parent and teacher screens, of which 1190 (45%) met the screen criteria.

Of the 1190 children meeting the screening criteria, 395 parents consented to a baseline assessment and met other eligibility requirements of English fluency, interest in participating in the program, no child developmental delay, and no safeguarding concerns. Then 325 (82%) completed the assessment, but only 210 (65%) were randomized. Those not randomized were listed in Figure 1 as dropping out of the study (59 were no longer interested in taking part), excluded (55 could not attend), or ineligible (1 had too low a score on a measure of antisocial behavior). The randomization appears to have occurred after the assessed subjects dropped out or were excluded.

Compared to those who did not consent, the 395 consenting parents had children with more severe problems, were referred to the study rather than recruited, and came from more disadvantaged families. However, the 210 randomized subjects differed little from the 395 consented subjects.

Assignment. The 210 subjects were randomized to four conditions: 1) the Incredible Years Parent Training to improve parenting, reduce child behavioral problems, and improve child and parent relationships (12 weeks); 2) the Supporting Parents with Kids' Education in Schools (SPOKES) program to improve parents' ability to support child reading development (10 weeks); 3) both Incredible Years and SPOKES programs in combination (22 weeks); and 4) a control group provided with a telephone information service that identifies information to parents about services that are appropriate for concerns about their child. The randomized sample sizes equaled 55 for the Incredible Years group and 54 for the control group.

Assessment/Attrition: The posttest occurred 9-11 months after the pretest, and a follow-up occurred 12 months after the posttest. The study collected full or some data on 80% of the families at post assessment and 70% at follow up.

Sample characteristics: The study sample of children with conduct problems came from families with greater social disadvantage than the local population. Of the sample children, 55% were boys, 24% had special education needs, 31% had a single parent, 37% had a parent who left school at age 16 or earlier, and 28% had a parent who never worked or was unemployed. On average, the children were 6 years old.

Measures: The study used five measures of parenting, two primary measures of child behavior (one for antisocial behavior and one for reading), and three secondary measures of child behavior (two for antisocial behavior and one for reading). Two measures, the Parental Account of Child Symptoms and the BAS Word Reading, were designated as primary outcomes. All researchers doing assessments were blind to condition, but all three measures of child behavior problems came directly or indirectly from parents who helped deliver the program.

The Alabama Parenting Questionnaire consists of two subscales reflecting positive and negative parenting behavior. The measures come from self-reports on parenting practices such as parental involvement, monitoring and supervision, consistency of discipline, corporal punishment, and other discipline practices.

The Interview of Parenting Practices measures 1) use of praise, and 2) use of aversive discipline. It is a semi-structured interview with parents about their own parenting. For each topic area, the parent gives detailed examples from the previous week and then the investigator rates the parent behavior. The study states that the instrument has satisfactory reliability and good validity when compared with directly observed parenting practices and other assessments of parenting. Its advantage over self-report questionnaires is that the interviewer makes the judgment of the parenting using objective criteria based on detailed descriptions.

The Reading Strategies Interview measures the time parents spend reading with their child and the strategies they use to create the right environment to help the child with any difficulties. The different strategies for enabling a positive atmosphere and appropriate support for reading were summed from five questions, each scored 0-2.

The Parental Account of Child Symptoms, a primary outcome, uses a semi-structured interview in which the interviewer assesses the severity and frequency of the child's disruptive behavior based on detailed parent accounts of several common situations. The measure is well validated and predicts poor child outcomes. The eight items are each rated for severity (0-3) and frequency (0-3) on a four-point scale. The mean score of all eight items is computed to yield the total Antisocial Behavior (AB) score.

The BAS Word Reading measure, a primary outcome, comes from the British Ability Scales. It is an individually administered standardized test of the child's ability to read single words.

The Visual Analogue Scale provides the opportunity for parents to report the nature and intensity of their child's difficulties on a 10 cm scale.

The Eyberg Child Behavior Inventory consists of 36 items designed to assess conduct problems, as reported by parents. It measures the frequency with which problems occur as well as the number of problems. This questionnaire has well-established validity.

The Reading Comprehension Test of the Wechsler Individual Achievement Test is an individually administered standardized test of the child's ability to understand a passage and answer questions based on it.

Analysis: The analysis used the General Linear Model with repeated measures and covariates. Tests compared each of the three intervention groups to the control group and compared pretest outcomes to posttest and follow-up outcomes. Other tests for moderation and mediation used standard methods.

The study states that it used an intent-to-treat strategy based on group assignment, irrespective of what type of treatment the subjects actually received. However, it gives little information on the source of attrition or efforts to follow subjects.

Outcomes

Implementation fidelity: About 72-76% of subjects in the three intervention conditions attended at least half the sessions.

Baseline equivalence: The conditions did not differ significantly on children's gender, age, or special educational needs (although p = .056 for age). The parents in the IY group were significantly less likely to leave school at age 16 or younger, but they did not differ significantly on being single parents or unemployed/never worked. The study did not report on equivalence of the outcome measures.

Differential Attrition: Attrition across the four conditions ranged from 13% -17% at posttest and 22%-30% at follow-up. Overall, the control group showed the lowest attrition, and those who were attending the fewest sessions were most likely to drop out. Otherwise, the study presented no tests of significance or additional analyses of differential attrition.

A memo to Blueprints from the lead author addressed differential attrition by comparing baseline measures of the Incredible Years and control groups for the analysis sample of cases with valid posttest and follow-up data. For the posttest sample (N = 36-44 for the control group and N = 42-45 for the Incredible Years group), 5 of 13 tests showed significant baseline differences. The five differences involved higher scores on child problem behavior measures and on poor parenting in the Incredible Years group. For the follow-up sample (N = 36-42 for the control group and N = 39-41 for the Incredible Years group), 4 of 13 tests showed significant baseline differences. The four differences again involved higher scores on child problem behavior measures and on poor parenting in the Incredible Years group.

Posttest: The results covered 10 outcome measures relating to parenting, antisocial child behavior, and child reading. For parenting, the Incredible Years group significantly improved 1 of 5 measures: negative parenting (d = .52). For antisocial behavior, the Incredible Years group significantly improved on 3 of 3 measures: scores from the semi-structured interview (d = .65), the Eyberg assessment (d = .60), and the Visual Analogue Scale (d = .65). For reading, the Incredible Years groups significantly improved 1 of 2 measures: British Ability Scales word reading (d = .25).

Long-term: The 1-year follow-up results differed little from the posttest results with one exception. The significantly greater improvement in the Visual Analogue Scale of child behavior at posttest for the Incredible Years condition dropped to insignificance at follow-up. The other outcomes differed significantly across conditions at posttest remained significant at follow-up.

Moderation: The study reported no significant moderation effects for child antisocial behavior, indicating similar benefits of the program for all subjects.

Mediation: To test whether changes in parent behavior mediated changes in child outcomes, the study examined two mediators: changes in parent use of aversive discipline and changes in parent reading strategies. Reduction in use of aversive discipline appears to mediate the influence of the reading condition on antisocial behavior (but not the influence of the Incredible Years condition).

Study 39

Evaluation Methodology

Design: Using a randomized controlled design, 936 5-6-year-old children from eight schools in London were screened for antisocial behavior using a parent and teacher questionnaire. High-risk cases (n=279; 40.8%) scoring above a pre-determined cut-off were invited to participate in the study. Of these, 128 families (46%) agreed to participate, however 16 were unable to attend and therefore a total of 112 families (40%) were then randomly assigned within each school to the intervention (n=61) or control (n=51) condition. Data were gathered at baseline, at 4-months post-intervention, and 4-7 years post-intervention. Attrition at 4-month follow-up was low at 2.6% with 3 participants dropping out of the intervention condition. At 4-7 year follow-up, attrition was approximately 20%.

Sample Characteristics: The average age of the sample of children was 5.2 years and 70% of the sample was male. The study reported that 34% of the children were ethnic minorities, 52% lived in single-parent households and most families were poor and disadvantaged.

Measures: Thirteen parent behavior measures and seven child behavior measures were gathered. Parents and assessors were blind to the condition when baseline measures were collected but not later. Interviewers and coders were blind to condition at follow-up as well.

Assessments consisted of previously used measures and were gathered for parent behavior through parent interviews (play, praise, rewards, consequences, time-outs, harsh discipline, warmth, criticism); questionnaires (appropriate and positive versus harsh and inconsistent); and, video-taped observation of parent-behavior (positive attention, seek cooperation, gives commands). For child behavior, measures were gathered through parent-interview (antisocial behavior, ADHD symptoms, emotional symptoms, oppositional defiant diagnosis); questionnaires (Eyberg Child Behavior Inventory and Teacher-rated oppositional symptoms); and a test for reading ability.

Analysis: Analysis of covariance/multiple regression was used, and the study accounted for the effect of school- and group-level clustering and controlled for baseline outcomes. Complete case analysis was conducted and then repeated using maximum likelihood methods. The study complied with the intent-to-treat principle by analyzing the full sample regardless of dose received.

OUTCOMES

Implementation Fidelity: Facilitators received extensive training, weekly supervision, completed treatment adherence schedules after each session, and received weekly written feedback from participants. The median attendance was 15 out of 28 sessions; 75% of parents attended 5 or more sessions and were labeled attenders.

Baseline Equivalence: The study reported that no significant differences between conditions were found for any demographic variables. It is unclear if the study examined equivalence for all outcome variables at baseline.

Differential Attrition: No significant differences were found for baseline conduct scores between families who participated in the study and those who did not. Attrition was low at 4-month follow-up (2.6%).

At 4-7 year follow-up, attrition was 20% and the study reported that "dropout was not significantly associated with any pre-intervention variable."

Posttest: Five of the seven child-behavior measures were significantly different between the intervention and control at 4-month follow-up in the expected direction. Specifically, intervention children improved on reading scores, parent-reported antisocial behavior and ADHD symptoms, and the questionnaire-based child behavior inventory. Further, rates of oppositional defiance among intervention children halved on reassessment and were unchanged in control children.

Program Effect on R&P Factors: Nine of the thirteen parent-behavior measures were significantly different between the intervention and control at 4-month follow-up in the expected direction. Specifically, intervention parents used more play, praise, rewards, and time-out, and less harsh discipline, showed increased warmth, attending to, praising and cooperation and decreased criticism towards their child.

Effect sizes were reported as Cohen's d and ranged between small-medium (0.31) and medium-large (0.63).

Long-term Effects: At 4-7 year follow-up, no intervention effects were found for parent or child outcome measures.

Study 40

Evaluation Methodology:

Design:

Recruitment: The study used two methods to recruit mothers with disruptive children. It included participants who were referred to the program from outpatient clinics for child and adolescent psychiatry for disruptive child behavior. A total of 53 families were referred and 45 participated in the program. In addition, in order to include more ethnic minority families, the study recruited families from schools. A total of 268 families were contacted in recruitment and 109 met criteria for participation.

Assignment: Families were randomly assigned to either the Incredible Years treatment (n=107) or a waitlisted control group (n=47). The ratio of random assignment was 2 to 1, in order to include more families in the treatment immediately. The waitlist design allowed the control group to enter the program after the posttest.

Attrition: Assessments occurred at pretest, posttest, and 3-month follow-up for the intervention group only. The study reported that 23% of the data was missing in the final analysis. Forty-two families had missing data at one of the time points and an additional 22 families dropped out of the study completely.

Sample: Children in the sample were 3 to 8 years of age (mean=5.59). The sample was comprised of 41% Moroccan immigrants and 19% Turkish immigrants, the remaining families were either Caucasian or an unspecified "other." Of the ethnic minority mothers, 78% were not born in the Netherlands and many reported problems speaking or understanding Dutch (approximately 66%). Variation in education level was large, but the mothers had on average 13 years of education (primary and secondary school).

Measures: The study included 15 outcome measures reported mostly by mothers, who helped deliver the program, but also by teachers. The measures had good reliability.

For the first three outcomes, the Eyberg Child Behavior Inventory (ECBI), a mother-reported measure of child disruptive behavior, separately reports incidence of problem behavior and severity of the behavior. Mothers also reported children's aggressive behavior toward other children using an adaptation of the Teacher Rating of Aggression.

For the next six outcomes, the Strengths and Difficulties Questionnaire (SDQ) is a mother- and teacher-report of children's problem behavior (total problems, conduct, and hyperactivity/inattention). It is not clear if teachers were blind to child condition.

The Parent Practices Interview measured five parenting practices of the mothers: appropriate discipline, harsh and inconsistent discipline, physical punishment, praise and incentives, and clear expectations. Last, the Parenting Stress Index measures maternal parenting stress (presumably mother-report).

Also, the study conducted diagnostic interviews of parents regarding child behavior, however the results of those do not appear to be reported.

Analysis: The main outcomes were analyzed using analyses of covariance to compare the outcome among the treatment group as compared to the control group. Additional analysis tested the moderator effects of socioeconomic status and the recruitment mechanism.

Intent-to-Treat: The study used multiple imputation to include all randomized cases. These results are more conservative when compared to the analysis of the sample with complete data and are reported in the tables.

Outcomes

Implementation Fidelity: The study discussed measures to assess fidelity, but noted only that attendance rates were 66%-72%.

Baseline Equivalence: The study reported no significant differences between the treatment and control group at baseline, but used an overall multivariate analysis of variance rather than tests for each of the baseline measures.

Differential Attrition: The study reported significant attrition, and noted that several baseline measures predicted missing data, but did not formally test for differential attrition.

Posttest: At posttest, the study reported significant treatment effects for 4 of 9 child outcomes.For mother-reported measures of problem behavior (ECBI), there was significant improvement in the treatment group as compared to the control group for the intensity (p<.05) and incidence of problem behavior (p<.001), but not for aggression. In addition, treatment mothers reported improvement in child conduct (SDQ, p<.05) and teachers reported improvement in hyperactivity or inattention (SDQ, p<.05).

Risk and Protective Factors: In addition, the study reported significant treatment effects for 2 of 6 risk and protective outcomes. Treatment mothers reported a decrease in harsh and inconsistent discipline (p<.05) and praise and incentives (p<.001).

Although the 3-month follow-up had no control group, the study reported that posttest effects were maintained (i.e., no significant changes occurred from posttest).

Effect sizes were reported as moderate, ranging from a Cohen's d of .38-.57 for child behaviors and from .59-.64 for parent behaviors.

Long-term Effects: No long-term follow-up was conducted.

Study 41

Evaluation Methodology:

Design:

Recruitment: This Dutch study used a nationwide screening within all penitentiary institutions or via organizations working with formerly incarcerated women. Mothers who had either recently been released or would be released within the following 3 months, were caregivers to children between the ages of 2 and 10, and saw their children at least 2 weekends per month were invited to participate. A total of 183 women were considered eligible and 113 consented to participate.

Assignment: Women were randomly assigned to either the treatment or control conditions in a ratio of 2 to 1. However, in 2 of 6 phases of recruitment, there were not sufficient numbers to assign in this ratio and randomization was suspended while all participants were assigned to the treatment group.

Attrition: Assessments occurred at four time points: pretest, 4 months and 5 months into the program, and posttest. Of the 113 women assigned to a condition, 86 were assigned to the treatment and 27 to the control. In the control condition, 23 women were included in the analysis (85%). In the treatment condition, a total of 68 (79%) were included in the analysis from both those who attended sessions and those who did not. Reasons for exclusion included having no data or not meeting inclusion criteria in retrospect.

Sample: The education level among mothers in the sample was low, with 73.6% having completed lower secondary education or less. A minority was native Dutch (23.6%) and the rest originated from other countries. The average sentence of participating mothers was 11.1 months and the majority was convicted of drug-related offenses (57.5%).

Measures: The study included 8 outcome measures, 7 reported by mothers, who helped deliver the program, and one by teachers. The measures had good reliability.

For the first 3 outcomes, the Eyberg Child Behavior Inventory (ECBI), a mother-reported measure of child disruptive behavior, separately reports incidence of problem behavior and severity of the behavior. Teachers reported children's aggressive behavior toward other children using an adaptation of the Teacher Rating of Aggression.

The next 5 outcomes focused on parenting and used the Alabama Parenting Questionnaire (APQ), a mother-report of the most important aspects of parenting behaviors related to disruptive behavior problems in children: involvement, positive parenting, poor monitoring/supervision, inconsistent discipline, and corporal punishment.

Analysis: The study used a multilevel analysis including three-level models, assessments (Level 1) were nested within children (Level 2) and children were nested within families (Level 3). For positive parenting and corporal punishment, two-level models were used because of non-significant variance at the child level. The models controlled for baseline outcomes with group-by-time interactions. However, one-tailed tests were used for significance of the interactions.

Intent-to-Treat: The study reported intent-to-treat and per protocol analyses separately, with the intent-to-treat analysis using all available data and including 19 mothers who attended no sessions.

Outcomes

Implementation Fidelity: Group leaders reported covering 98% of program activities.Average attendance over 12 group sessions was 7.7 and 3.2 of 4 home visits.

Baseline Equivalence: The study stated that there were no baseline differences between children in the intervention and control groups for parenting and child outcomes at baseline but did not give details or mention baseline measures for demographics.

Differential Attrition: The study does not discuss differential attrition. It did note that only one baseline difference (child age) was found between children with and without teacher data.

Posttest: None of the child outcomes was significant with two-tailed tests. The study reported significant (one-tailed) treatment effects for 1 of 3 child outcomes.For mother-reported measures of problem behavior (ECBI), there was significant improvement in the treatment group as compared to the control group for the intensity (p<.05, one tailed) of problem behavior. The study reported a trend approaching significance (p<.10, one tailed) for the number of problem behaviors and aggressive behavior.

Risk and Protective Factors: In addition, the study reported significant treatment effects for 1 of 5 self-reported parenting behaviors. There was significant improvement in the treatment group as compared to the control group for inconsistent discipline (p<.001).

Long-term Effects: No long-term follow-up was conducted.

Study 42

Evaluation Methodology:

Design:

Recruitment: Children were either 1) clinically referred by pediatricians, child psychiatrists or psychologists, 2) self-referred by parents, or 3) screened in preschool settings. All participating children were at-risk for disruptive behavior based on the assessment of their caregiver. In addition, children and families were formally evaluated using parent reports and a laboratory-based mother-child interaction observation. Exclusion criteria included formal diagnoses of neurological or developmental disorder, severe developmental delay, or if a child was undergoing any pharmacological or psychotherapeutic intervention.

Assignment: The study used a stratified random assignment of children and their families to either the treatment condition or waitlist control group, balancing for age and gender. The study began by using a 2:1 allocation to allow more families to receive the intervention, but switched to a 1:1 allocation to ensure large enough numbers in the control group. A total of 455 children were initially screened and 197 fulfilled the inclusion criteria. Of these families, 73 were excluded (50 did not complete the evaluation process, 8 met exclusion criteria, and 15 were unwilling to participate). In total, 124 children and their families were randomly assigned, 68 to the treatment condition and 56 to the control condition.

Attrition: Assessments by condition occurred at two time points: baseline and posttest (conducted 6 months after baseline). In addition, 12 and 18-month data were collected only on treatment families with booster sessions provided 9 and 15 months after baseline. Among treatment participants, 65 (96%) completed the posttest whereas 49 (87.5%) of control group participants completed the posttest. Overall, 114 participants (92%) were administered a posttest.

Sample: The average age of the children was 55.86 months (with ages ranging from 3 to 6 years of age) and a majority (73%) was male. The families were mostly of medium socioeconomic status, and the majority of caregivers were mothers (98%) who were married or living as married (80%) with a mean age of 35.35 (SD = 5.50). Almost half (48%) of the mothers had a university degree. The study does not provide information on race/ethnicity of the sample, but the research was conducted in Portugal.

Measures: The study included 4 measures of child behavior (rated by parents) and 2 additional observer-rated measures of child behavior. The observations were videotaped and coded by a trained rater who was blind to condition. The study also included 3 measures of self-reported parent behaviors and 2 observer-rated (with raters once again blind to condition) measures of parent behavior. In addition, the study included measures of parental stress and parental depression as baseline measures for the mediation analyses, but not outcomes.

The parent-rated outcome measures of child behavior included the Strengths and Difficulties Questionnaire Hyperactivity (sample α=.66) and Conduct (sample α=.46) scales. The study also used the Preschool and Kindergarten Behavior Social Skills (sample α=.88) and Externalizing (sample α=.90) scales.

The parent self-reported measure used in the mediation analysis included the Parenting Scale (sample α=.74), which indicates dysfunctional discipline practices. The mediation analysis also used the Parenting Sense of Competence Satisfaction (sample α=.76) and Efficacy (sample α =.79) subscales.

In 25-minute sessions, trained observers rated caregiver and child interactions in a lab setting, engaged in free play. The observers used the Dyadic Parent-Child Interaction Coding System to rate child pro-social behavior, child deviance and non-compliance, positive parenting, and critical parenting in 5-minute segments. The study reported 76% inter-rater agreement for 20% of double-recorded sessions. Intra-class correlations for each subscale were .53 for Child Pro-Social Behavior, .92 for Child Deviance and Non-Compliance, .97 for Positive Parenting, and .91 for Critical Parenting. The Child Pro-Social Behavior and Child Deviance/Non Compliance subscales were used as outcome measures, whereas the Positive Parenting and Critical Parenting subscales were included in the mediation analyses.

Analysis: The study used a General Linear Model for repeated measures analysis of variance to study the effects of the interaction between condition and time, followed by pairwise comparisons with Bonferroni adjustment. In addition, the Greenhouse-Geisser sphericity correction was performed and reported for multivariate analyses.

Intent-to-Treat: The study did not follow an intent-to-treat model.

Outcomes

Implementation Fidelity: The study reported that program deliverers were trained and followed program protocol, but did not provide results of data collected using these quantitative measures to assess level of fidelity. The study reported 78% attendance with 59 caregivers attending 9 or more sessions.

Baseline Equivalence: The study reports no significant differences between conditions on demographic or outcome measures at baseline.

Differential Attrition: Overall, 8% of the participants dropped out of the study from baseline to 6-month follow up. While more families in the control group were lost compared to the treatment group, this difference was not significant. There were, however, statistically significant differences between completers and attritors on some demographic variables. When compared to retained caregivers, those who were lost had fewer years of schooling and lower SES, were non-married in a higher proportion, and evaluated their children as having higher social skills. The study does not report treatment by outcome differential attrition, however the treatment and control groups remained equivalent after removing data for attritors.

Posttest: The study found significant time-by-condition effects for all 4 parent-reported measures of child behavior at posttest, which were not considered independent measures since parents both received the program and rated outcomes of the intervention. According to these assessments, however, compared to control children, children in the treatment group showed significantly greater reduction in hyperactivity, conduct problems, externalizing behaviors, and a greater increase in social skills from pretest to posttest. In addition, the condition-by-time interaction for the observer-rated, indepedent measure of pro-social behavior was significant in favor of the treatment group, but there were no significant differences in the measure of deviance. Moderator analyses showed the effect on social skills was found to be more significant for children aged 4 and 5 than children aged 3.

Risk and Protective Factors: The study reported a significantly larger decrease in self-reported negative parenting practices and the increase in mothers' perceptions of self-efficacy with parenting, between pretest and posttest among treatment caregivers as compared to control caregivers. However, the treatment by time interaction was not significant for feelings of satisfaction. In addition, the condition-by-time interaction for the observer-rated measure of positive parenting was significant in favor of the treatment group, but no differences were detected for critical parenting.

Long-term Effects: Not reported by condition.