|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
PARENT-CHILD INTERACTION THERAP Y FOR DISRUPTIVE BEHAVIOR IN CHILDREN WITH MENTAL RETARDATION: A RANDOMIZED CONTROLLED TRIAL By DANIEL MARC BAGNER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
Copyright 2007 Daniel M. Bagner
ACKNOWLEDGMENTS I would like to thank my mentor, Dr. Sheila Eyberg, for her guidance, support, and encouragement throughout graduate schoo l. My appreciation also goes to the undergraduate and graduate resear ch assistants in the child st udy lab for their hard work and devotion in helping to make a difference in the lives of children and their families. I would also like to thank my committee members, Drs. Stephen Boggs, Maureen Conroy, and Michael Robinson, for their general support as well as their spec ific suggestions and contributions. I would like to express my a ppreciation to the funding sources that made this study possible, including the National Institute of Mental Health (F31 MH068947); American Psychological Association, Divisi on 53; Childrens Miracle Network, Shands Hospital, University of Florid a; and Center for Pediatric Psychology and Family Studies, University of Florida. I also wish to ac knowledge the extraordinary amount of support, love, and encouragement I have received from my wife, Amy, and my family and friends. iii
TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF TABLES .............................................................................................................vi LIST OF FIGURES .........................................................................................................viii ABSTRACT .......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 2 METHOD.....................................................................................................................6 Participants ...................................................................................................................6 Measures .......................................................................................................................9 Screening Measures ...............................................................................................9 Measures of Child and Parent Functioning .........................................................11 Measure of Consumer Opinion ...........................................................................16 Study Design and Procedure.......................................................................................16 Treatment ....................................................................................................................19 Treatment Integrity .....................................................................................................22 3 RESULTS...................................................................................................................23 Observed Parent-Child Interaction .............................................................................25 Parent Report ..............................................................................................................26 Generalization of Treatment E ffects to the School Setting ........................................27 Consumer Satisfaction ................................................................................................28 Intent-To-Treat Analyses ............................................................................................28 Clinical Significance ...................................................................................................29 DISCUSSION ....................................................................................................................33 APPENDIX A ANALYSES US ING ANCOVA................................................................................38 B ANALYSES WITH TREA TMENT COMPLETERS................................................40 iv
C ANALYSES WITH MEDICALLY STABLE CHILDREN......................................42 LIST OF REFERENCES...................................................................................................44 BIOGRAPHICAL SKETCH.............................................................................................50 v
LIST OF TABLES Table page 2-1. Demographic Characteristics of Immediate Treatment (IT) and Wait-List (WL) Groups........................................................................................................................8 2-2. Inter-coder Reliability of the Dyadic Parent-Child Interaction Coding System-II (DPICS-II) for Mothers and Children with Comorbid Mental Retardation and Oppositional Defiant Disorder.................................................................................15 3-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments....................................................................................................25 3-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments.............................................................................................................27 3-3. Mean SESBI-R Scores for Teachers at Time 1 and 2 Assessments...........................28 3-4. Number of Families in the Clinically Significant Range at the Time 1 and 2 Assessments and Clinically Significant Change......................................................31 A-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments Using ANCOVA.......................................................................38 A-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments Using ANCOVA.................................................................................38 A-3. Mean SESBI-R Scores for Teachers at Time 1 and 2 Assessments Using ANCOVA.................................................................................................................39 B-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments With Treatment Completers......................................................40 B-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments With Treatment Completers................................................................40 B-3. Mean SESBI-R Scores for Teachers at Time 1 and 2 Assessments With Treatment Completers..............................................................................................41 C-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments With Medically Stable Children...............................................42 vi
C-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments With Medically Stable Children.........................................................42 C-3. Mean SESBI-R Scores for Teachers at Time 1 and 2 Assessments With Medically Stable Children........................................................................................43 vii
LIST OF FIGURES Figure page 2-1. ECBI Intensity and Problem T-Scores During Treatment..........................................20 3-1. Participant Flow Diagram...........................................................................................24 viii
Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PARENT-CHILD INTERACTION THERAP Y FOR DISRUPTIVE BEHAVIOR IN CHILDREN WITH MENTAL RETARDATION: A RANDOMIZED CONTROLLED TRIAL By Daniel M. Bagner August 2007 Chair: Sheila M. Eyberg Major: Psychology This study presents results of a randomized controlled trial examining the efficacy of parent-child interaction therapy (PCIT) with families of preschool-age children with co-morbid oppositional defiant disorder and me ntal retardation. Fo llowing an initial assessment, 30 clinic-referred families were randomly assigned to an immediate treatment (IT) or waitlist (WL) control group. Results indicated that mothers in the IT condition interacted more positively with thei r child and were more successful in gaining their childs compliance than mothers in the WL group. In addition, mothers who received treatment reported statistically and clinically significant improvements in their childs behavior and were highly satisfied w ith both the content and process of PCIT. These results provide prelimin ary evidence of the efficacy of PCIT with an underserved population in need of behavioral intervention for disruptive behavior. ix
CHAPTER 1 INTRODUCTION Disruptive behavior among preschool-age children is a growing concern because of its high prevalence and poor prognosis (Loeber, Burke, Lahey, Winters, & Zera, 2000). In the general population, the preval ence of disruptive behavior disorders, including oppositional defiant di sorder (ODD) and conduct diso rder, is estimated to be between 2% and 16% (American Psychiat ric Association [APA], 2000). Disruptive behavior is also the most common reason for referral of young children to mental health services (Frick, 1998; Reid, 1993). Evidence-bas ed treatments have been found effective for these disorders (Eyberg, Nelson, & Boggs in press), but children with mental retardation (MR) are typically excluded from these outcomes studies and represent a high priority for study (Kazdin, 2002). Although few studies have evaluated prev alence rates of co -morbid disruptive behavior disorders in children with MR even higher prevalence rates have been estimated among these children than in th e general population (Benson & Aman, 1999). Jacobsons (1982) large-scale New York study indicated that approximately 40% of children with mild MR and 47% of children w ith moderate MR had si gnificant disruptive behavior. In addition, in one community mental health c linic, Benson (1985) found that almost half of the children and adolescen ts with MR had been referred for conduct problems. Part of the reason that so few studi es have examined prevalence rates of disruptive behavior disorders in children with MR may be related to the more general 1
2 controversy surrounding diagnosis of psyc hiatric disorders in this population. Professionals have disagreed as to whether significant disrupti ve behavior is indicative of a distinct disorder or a feature of the presentation of MR (Kobe & Mulick, 1995). According to the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition, Text Revision (DSM-IV-TR; APA, 2000), there is no evidence that the nature of comorbid mental disorders is any different in individuals with or without MR. Further, because behavior problems are not qualitative ly different in children with and without MR, authors have suggested that treatmen ts developed for use with non-handicapped children may be suitable for use with child ren with MR as well (Kobe & Mulick, 1995). For many children with MR, treatment for their disruptive beha vior is critical. Disruptive behavior often prevents these ch ildren from participating in important educational and community activ ities (Durand, 2001) as well as rehabilita tion activities for associated disorders such as speech, occupational, or physical therapy. Further, Benson and Aman (1999) have suggested that behavior problems contribute to physical safety concerns for children with MR and those around them, higher need for supervision, reduced oppor tunities for independent functioning, and disrupted interpersonal relationships. As a result, ca regivers of children with MR have been increasingly turning to mental health prof essionals for treatment of their childrens disruptive behavior (Kobe & Mulick, 1995). Individual behavioral proce dures have historically play ed a central role in the management of children with MR (Durand, 2001). The most common behavioral techniques used to manage behavior problems have been positive reinforcement and time-out procedures (Benson & Aman, 1999). Differential reinforcement has been a
3 useful technique to reduce behavior problems in children with MR as well (Handen, 1998). The behavior problems targeted by these techniques for children with MR, however, have typically been limited to aggr essive behaviors rather than disruptive behaviors more generally (Benson & Aman, 1999). Parent-training interventions have also been implemented for children with MR. Some studies have demonstrated that parent training is superior to individual treatment with the child (Handen, 1998). This may be the case, in part, because parents are able to learn behavioral techniques and apply them to their child s behavior on a daily basis (Harris, Alessandri, & Gill, 1991). Parent-training interventions have focused not only on parent skill acquisition, but also on education about lifelong support issues (Walters & Blane, 2000). There have also been some studies examining parent-t raining interventions targeting disruptive behavior in children with developmenta l disabilities, but most of these studies used single subject or multiple baseline designs (Roberts, Mazzucchelli, Taylor, & Reid, 2003). Recently, a randomized controlled trial of the Positive Parenting Program (Triple P; Sanders, 1999), an evidence-based parent -training intervention, was adapted for use with developmentally delayed children and show n to improve disruptive behavior in this population (Roberts, Mazzucchelli, Studman, & Sanders, 2006). However, no other evidence-based parent-training interventions for disruptive behavior have been conducted with children with MR. The need for furthe r randomized controll ed trials of early interventions that focus specifically on decrea sing disruptive behavi or is critical for several reasons. First, replicati on is important to ensure that other evidence-based parenttraining interventions lead to similar effects. For example, Roberts and colleagues did not
4 find significant changes in noncompliance or parenting stress, two outcomes that are common following other evidence-based trea tments for disruptive behavior with nondelayed populations. Second, the adapted Triple P program targeted the specific needs of delayed children rather than disruptive behavior more ge nerally. Implementing evidencebased treatments for disruptive behavior wit hout modifications may lead to different effects and warrants re-examination with delayed children. Last generalization of treatment effects to the school setting for child ren with delays has not been investigated. When children with MR present with significant disruptive behavior it is difficult for teachers to teach them other life skills (e .g., academic skills, social skills). Until their disruptive behavior is manageable, many child ren will be unable to benefit from other efforts in the school setting. Parent-Child Interact ion Therapy (PCIT) is an empirically supported treatment for disruptive behavior in preschool-a ge children (Eyberg et al., in press) that is designed to change parent-child interaction patterns and thereby change childrens disruptive behavior. In PCIT, parents are taught specific skills to increase the childs pro-social behavior and decrease the ne gative or undesirable behavior Clinically, PCIT has been used successfully to treat the disruptive beha viors of children with mild and moderate MR, but the effectiveness of this treatment for children with MR has not been examined in a controlled trial. For children with MR, ge neralization of the cha nges in their cooperative behaviors with their parents to new situati ons, such as school, is critically important. Disruptive behavior can prevent children w ith MR from participating in additional treatment programs (Durand, 2001). Improvements in childrens compliance and social
5 skills during PCIT have been shown to generalize to their behavior at school as measured by teacher rating scales and observational m easures of classroom behavior (Funderburk et al., 1998; McNeil, Eyberg, Eisens tadt, Newcomb, & Funderburk, 1991). The purpose of this study was to evaluate the efficacy and gene ralization of PCIT with children with co-morbid disruptive behavior and MR. We hypothesized that in comparison to families in a waitlist (WL) control group, families in the PCIT group would show after treatment: (a) more positive and effective parenting behaviors during parent-child interactions; (b) greater child compliance during observed parent-child interactions; (c) fewer child disruptive behaviors as rated by parents (d) improved scores on a parent self-report measure of parenting stress; and (e) fewer child disruptive behaviors as rated by teachers.
CHAPTER 2 METHOD Participants Participants were 30 female primary careg ivers and their 3to 6-year old child. Power calculations, based on an earlier study of PCIT (Eisenstadt et al., 1993) and using power = .80 and a 2-sided comparison with al pha = .05, determined that an effective sample size of 8 families per group would be sufficient to detect expected differences. All adult caregivers living in the home were encouraged to participate in treatment due to recent findings that father involvement leads to better maintenance of treatment gains (Bagner & Eyberg, 2003). However, data for this study were collected only from the primary caregiver, who in all cases was the mother. Most of the children (80%) were referred by pediatric health care professionals (e.g., developmental pedi atricians, clinical child psychologists, speech pathologists, a nd occupational therapis ts), although 10% were referred by the childs teacher and 10% were self-referred after s eeing flyers announcing the study. Inclusion criteria were, first, a di agnosis of ODD accord ing to criteria recommended by Jensen et al. (1996) for optimal caseness. The child was required to meet criteria for ODD on the Diagnostic Inte rview Schedule for Children-Fourth EditionParent Version (DISC IV-P; Shaffer, Fisher Lucas, Dulcan, & Schwab-Stone, 2000) and to score above a T score of 64 on the Aggressive Behavi or Scale of the Child Behavior Checklist for 1 to 5 year olds (CBCL 15; Achenbach & Rescorla, 2000). Children were also required to meet di agnostic criteria for mild or moderate MR based on the 6
7 criteria of the American A ssociation for Mental Retarda tion (Lambert, Nihira, & Leland, 1993). Specifically, the children were required to obtain a standard score between 40 and 75 on a measure of intellectual functioning and to show significan t deficits in at least two areas of adaptive behavior according to the primary caregivers report on the Adaptive Behavior Scale-School, 2 nd Ed. (ABS-S: 2; Lambert et al.). Inclusion criteria also required that the primary caregiver obtain a Standard Score of 75 or higher on a cognitive screening measure. Exclusion criteria for children included major sensory impairments (e.g. deaf, blind) and autism spectrum disorders. Familie s suspected of child abuse at the initial assessment or during the course of treatment were also excluded and were reported to the state as required by law. Children on psycho active medication to c ontrol their behavior were not excluded if they were stabilized on the medication and dosage for 1 month or longer at the time of the initial evaluation. A ny family meeting exclusionary criteria was referred to appropriate services outside the research project. Twenty-eight children were initially screened out of the study for the following reasons: 14 children scored higher than 75 on the measure of intellectual functioning, 8 children scored below a T score of 65 on the CBCL Aggressive Behavior subscale, 1 child did not show significant deficits in at least two areas of adaptive behavior 1 child met criteria for an autism spectrum disorder, 1 child was deaf, 1 primary caregiv er scored below a Standard Score of 75 on the cognitive screening measure, 1 family d eclined participation, and 1 family did not show up for their screening assessment. Childrens mean age at the time of the screening assessment was 54.13 months (SD = 10.15). Most children (60%) met diagnostic cr iteria for mild MR (IQ score between 55
8 and 75), and most (77%) were boys. The raci al/ethnic composition of the families was 67% Caucasian, 17% African American, 13% Biracial, and 3% Hispanic. The mean Hollingshead (1975) score was 41.30 ( SD = 14.14) indicating that, on average, families fell in the lower to middle income range of socioeconomic status. Families were randomly assigned to the immediate treatment (IT) group ( n = 15) or the WL group ( n = 15). The IT group contained 10 ( 66%) two-parent families, and the WL group contained 12 (80%). Of the 10 twoparent families in the IT group, 6 (60%) had fathers who attended at least one treatment session. As shown in Table 2-1, there were no significant demographic differences between groups. Table 2-1. Demographic Characteristics of Immediate Treatment (IT) and Wait-List (WL) Groups IT a WL b Characteristic M SD M SD t (28) 2 (1) p Child age (months) 52.40 8.81 55.87 11.38 0.93 -.359 Child sex (% female) 20.00 -26.67 --0.19 .666 Child ethnicity (% minority) 26.67 -40.00 --0.60 .439 Child FSIQ 57.53 11.01 60.80 11.25 0.80 -.428 Maternal age 35.20 8.57 37.67 7.22 0.85 -.401 Paternal age c 38.50 11.05 38.33 9.64 0.04 -.970 Maternal IQ 99.40 13.95 99.60 14.44 0.04 -.969 Hollingshead 37.40 13.24 45.20 14.35 1.55 -.133 Marital status (% two parent) 66.67 -80.00 --0.68 .409 Distance (miles) 32.64 22.98 36.01 21.72 0.41 -.683 a n = 15. b n = 15. c There were 22 fathers included in th e group comparisons resulting in a df of 20. The study completers included the 10 IT families and the 12 WL families that successfully completed the Time 2 assessment. Of the 10 IT families that completed the Time 2 assessment (which occurred 4 months after the Time 1 assessment), 8 families successfully terminated treatment and were considered treatment completers, and 2
9 families that completed the Time 2 assessm ent subsequently dropped out of treatment before meeting treatment termination crit eria. The study dropouts included the 5 (33%) IT families and the 3 (20%) WL families that did not complete the Time 2 evaluation. Overall, the study dropout rate was 27% a nd the treatment dropout rate was 47%. The treatment dropout rate is somewhat lower th an other child treatment outcome studies, which have reported that between 50 and 75% of children and families referred to parent behavior management programs discontinue treatment prematurely (Kazdin, 1996; Wierzbicki & Pekarik, 1993). Comparison of the study dropout ra tes between the two groups was not statistically significant, 2 (1, N = 30) = .68, p = .41. Measures Screening Measures Diagnostic Interview Schedule for Ch ildren-Fourth Edition-Parent Version (DISC-IV-P; Shaffer et al., 2000) The DISC-IV-P is a structured diagnostic interview for administration to parents. It includes all comm on mental disorders of children included in the DSM-IV that are not dependent on speci alized observations or test procedures. Individual modules of the in terview can be administered separately, and 1-week testretest reliability for administration to pare nts of 9 to 17 year old children has been reported at .54 for ODD (Shaffe r et al.). The DISC-IV-P was used, along with the CBCL Aggressive Behavior subscale, to screen children for inclus ion based on the presence of ODD. Wechsler Preschool and Primary Scale of Intelligence-Third Edition. (WPPSIIII; Wechsler, 2002). The WPPSI-III assesses cogn itive ability in children ages 3 to 7. The WPPSI-III Performance IQ, Verbal IQ, a nd Full Scale IQ (FSIQ) have test-retest reliability coefficients ranging from .84 to .93, and the intersubtest correlations are all
10 moderate to high (Wechsler). Children were administered the core subtests (i.e., 4 subtests for the 3-year-olds and 7 subtests fo r the 4to 6-year-olds) to derive a FSIQ score, which was used as one component of the diagnosis of MR; children were required to obtain a FSIQ score of 75 or below. Adaptive Behavior Scale-School: Second Edition (ABS-S: 2; Lambert et al., 1993). The ABS-S: 2, a revision of the orig inal Adaptive Behavior Scale (Lambert, Windmiller, Cole, & Figueroa, 1975), is a questionnaire measuring adaptive behavior in individuals between the ages of 3 and 21 that can be admi nistered by psychologists or others with direct knowledge of the indivi duals behavior. The AB S-S: 2 assesses nine domains: independent functioning, physical development, economic activity, language development, numbers and time, prevocational/vocational activity, self-direction, responsibility, and socialization. Internal consistency coe fficients range from .82 to .98 for these domains, and 2-week test-retest reliability ranges from .42 to .79 (Lambert et al., 1993). The ABS-S: 2 was used as the s econd component for the diagnosis of MR; children were required to obtain a standard score of 4 or below on at least two domains (based on the non-MR norms). The Childhood Autism Rating Scale. (CARS; Schopler, Reichler, & Renner, 1988). The CARS is a 15-item observational ra ting scale developed to identify children with autism and to distingui sh them from developmentally handicapped children without autism. Internal consistency has been reported to be .94, and 1-year te st-retest reliability has been reported to be .88 (Schopler et al .). In this study, children who received a CARS raw score greater than 30 were excluded.
11 Wonderlic Personnel Test. (WPT; Dodrill, 1981). The WPT is a 50-item paperand-pencil test designed as a screening measur e of adult intellectual abilities. The test score is the number of items answered corr ectly in 12 minutes. In a sample of 120 normal adults, the Wonderlic estimate of intelligence correlated .93 with the Wechsler Adult Intelligence Scale (WAIS) Full Scale IQ score, and the Wonderlic score was within 10 points of the WAIS IQ score for 90% of the subjects (Dodrill). The Wonderlic IQ estimate was used as the cognitive screeni ng measure for the primary caregiver, and individuals obtaining a Standa rd Score less than 75 were screened out of the study. Measures of Child and Parent Functioning Child Behavior Checklist for 1 to 5 Year Olds. (CBCL 1-5; Achenbach & Rescorla, 2000). The CBCL 1-5 is a 99-item parent-rating scale designed to measure the frequency of childrens behavior and emo tional problems in the past 2 months. Only two items from the CBCL 1-5 differ from the earlier version (i.e., CBCL /2-3; Achenbach, 1992), which had concurrent validi ty with the Richman Behavior Checklist (Spiker, Kramer, Constantine, & Bryant 1992). Although designed for children between the ages of 1 and 5, the CBCL was used for all children in this study because of their cognitive delay. This procedure was reco mmended by the test author (Achenbach, personal communication, June 10, 2003). The A ggressive Behavior, Externalizing, and Total Scales of the CBCL 1-5 yield interrater (mother-father) reliability coefficients of .66, .67, and .65 and 1-week test-retest relia bility coefficients of .87, .87, and .90, respectively (Achenbach & Rescorla). Fo r this study, the CBCL 1-5 Aggressive Behavior subscale, along w ith the DISC-IV-P, was used for diagnosis of ODD. The Externalizing and Total Scales were also used as the primar y measures of child behavior
12 outcome for all children. In this sample, internal consistency estimates for the Externalizing and Total Scales were .82 and .95, respectively. Eyberg Child Behavior Inventory. (ECBI; Eyberg & Pincus, 1999). The ECBI is a 36-item parent-rating scale of disruptive be havior. The ECBI Intensity Scale measures the frequency with which disruptive behavior occurs, and the Problem Scale measures how problematic the childs behavior is for the parent. The Intensity and Problem Scales of the ECBI yield interrater (mother-fath er) reliability coefficients of .69 and .61 (Eisenstadt, McElreath, Eyberg, & McNeil, 1994) and test-retest reliability coefficients of .80 and .85 across 12 weeks and .75 and .75 acr oss 10 months, respectively (Funderburk, Eyberg, Rich, and Behar, 2003). The ECBI wa s given on a weekly basis to the primary caregiver as a measure of weekly progress in PCIT and was also used as a measure of treatment outcome. In this sample, internal consistency estimates for the Intensity and Problem Scales were .90 and .91, respectively. Parenting Stress Index-Short Form. (PSI-SF; Abidin, 1995). The short form of the PSI is a 36-item parent self-report inst rument containing three factor-analyticallyderived subscales: Parent Distress, Parent-C hild Dysfunctional Interaction, and Difficult Child. The short form subscales have shown 6-month test-retest reliabilities of .85, .68, and .78 respectively. The PSI a nd the PSI-SF total scores ar e highly correlated with one another (.94; Abidin). On the long form of the PSI, higher scores have been associated with increased severity of conduct-disorder ed behavior (Eyberg, Boggs, & Rodriguez, 1992; Ross, Blanc, McNeil, Eyberg, & Hembree-Kigin, 1998). On the short form, oppositional behavior was a strong predictor of the Difficult Child subscale in a sample of children in Head Start (Reitman, Currier & Stickle, 2002). The PSI-SF subscales were
13 used as outcome measures in this study to a ssess the effects of PCIT on parent stress. In this sample, internal consistency estimate s for the Parent Distress, Parent-Child Dysfunctional Interaction, and Difficu lt Child subscales were .73, .89, and .82, respectively. Sutter-Eyberg Student Behavior Inventory-Revised. (SESBI-R; Eyberg & Pincus, 1999). The SESBI-R is a 38-item teach er-rating scale of disruptive behavior at school. The SESBI-R Intensity Scale measures the frequency with which disruptive behavior occurs, and the Problem Scale measur es how problematic the childs behavior is for the teacher. The two scales of the SESB I-R have shown test-retest correlations of .80 and .73, respectively (Querido & Eyberg, 2003) SESBI-R Intensity scores correlate highly (.74) with the Oppositional Scale of the Conners Teacher Rating Scale-Revised: Long Form (Querido & Eyberg). Both the In tensity and Problem sc ores predict school conduct referrals and suspensions during th e subsequent 1 and 2 years (Schuhmann, 2000). The SESBI-R was used as an outcome m easure to assess generalization of changes in child behavior to the clas sroom. Internal consistency estimates in this sample for the Intensity and Problems Scales were .94 and .96, respectively. Dyadic Parent-Child Interaction Coding System-II. (DPICS-II; Eyberg, Bessmer, Newcomb, Edwards, & Robinson, 1994) The DPICS-II is the revised version of the DPICS (Eyberg & Robinson, 1983), a behavi oral coding system that measures the quality of parent-child social interactions. It provides an observational measure of parent and child behaviors in the labor atory or clinic during three 5-minute standard situations that vary in the degree of parental control required (i.e., child led play, parent-led play, and clean-up). Identical categories are c oded for parent and child and include
14 verbalizations (e.g., command, praise, criti cal statements), vocalizations (e.g., laugh, yell), and physical behaviors (e.g., destructive behavior, positive touch). Several sequences of behavior are also coded, such as compliance to total commands. Adequate reliability and validity have been established (Eyberg et al.), and Cohens kappa estimates ranged from .46 to 1.0 in a recent study (Brestan, Foote, & Eyberg, 2006). DPICS categories analyzed for this st udy were parent behavior description, reflection, praise, criticism, question, and command as well as child compliance to commands. Behavior descriptions are statements about the ch ilds current actions (e.g., Youre building a fence), and praises are verbalizations expressing a positive evaluation of the child. Praise includes both unlabeled praise (e.g., Good Job!) and labeled praise (e.g., I like the way you ar e sitting quietly.). Criticisms are any statements of disapproval of the child or th e childs attributes, ac tivities, products, or choices (e.g., You put it in the wrong place.) Questions are verbal inquiries from the parent to the child and can either request an acknowledgement from the child in response (e.g., Thats the red one, right?) or a verbal response beyond an acknowledgement (e.g., Where is the blue Lego?). Commands are directions from the parent to the child and can be either indirect (e.g., Can you put the red block on top?) or direct (e.g., Put the red block on top.). Child compliance is defined as the percentage of parent commands that are obeyed by the child when there is opportunity to comply. The DPICS-II was used to assess changes in parent-child interactions. Specifically, we examined the frequency of the verbalizati ons parents are taught to use (i.e., behavior descriptions, reflections, and pr aises) and the verbalizations parents are taught not to use (i.e., questions, commands, and criticisms) dur ing the child-led play. These groupings of
15 verbalizations are subs equently referred to as the Do and Dont skills, respectively. In addition, we investigated changes in child compliance rates during the combined parent-led play and clean up situations usi ng alpha compliance, which is defined as the percentage of times the child complies wh en there is an opportunity to comply. Coder training was aided by the use of The Workbook (Eyberg, Edwards, Bessmer, & Litwins, 1994), a DPICS codertraining manual. Undergraduate student coders were trained to 80% agreement with a criterion tape prior to coding family interactions for this stud y. Coder training involved week ly meetings and 3 hours of weekly homework/practice for approximatel y 12 weeks for each of the coders. The DPICS coders were uninformed as to the families group status (i.e., IT or WL). However, attempts to mask assessment points (i.e., Time 1 and Time 2) were unsuccessful. As displayed in Table 2-2, interrater and Cohens ka ppa reliabilities for these categories, calculated for one-third of the observations of each situation at each assessment point, ranged from 58% (noncomplia nce) to 91% (information question) and .55 (noncompliance) to .89 (labeled praise and information question), respectively. Table 2-2. Inter-coder Reliability of the Dyad ic Parent-Child Inte raction Coding SystemII (DPICS-II) for Mothers and Children with Comorbid Mental Retardation and Oppositional Defiant Disorder DPICS Category Percent Agreement a Kappa CDI Do Skills Behavior Description 68 .63 Unlabeled Praise 89 .84 Labeled Praise 89 .89 Reflective Statement 67 .66 CDI Dont Skills Indirect Command 67 .65 Direct Command 90 .83 Descriptive/Reflective Question 86 .80 Information Question 91 .89 Criticism 72 .66
16 Table 2-2. Continued DPICS Category Percent Agreement a Kappa Child Behavior Compliance 73 .69 Noncompliance 58 .55 No Opportunity for Compliance 71 .62 Total 77 .73 a Percent Agreement is based on summing the ag reements across participants and dividing by the agreements plus disagr eements across all situations. Measure of Consumer Opinion Therapy Attitude Inventory. (TAI; Eyberg, 1993). The TAI was designed to assess parental satisfaction w ith the process and outcome of therapy. It consists of 10 multiple-choice questions addressing the impact of parent training on such areas as confidence in discipline skills, quality of the parent-child inte raction, the childs behavior, and overall family adjustment. Te st-retest reliability over a 4-month follow-up period and correlations between the TAI and both parent-rating scales and observational measures of treatment change have been demonstrated (Brestan, Jacobs, Rayfield, & Eyberg, 1999). In addition, discriminative valid ity has been shown between outcomes of alternative treatments (Eisenst adt et al., 1993; Eyberg & Matarazzo, 1980). The TAI total score was used in this study to assess mate rnal satisfaction at the end of treatment. Internal consistency estimates for the TAI Total scale was .53 for the 10 families that completed the measure at the Time 2 evaluation. Study Design and Procedure The study was approved by the University of Florida Health Science Center Institutional Review Board and the Director of Research of the Alachua County School Board. We telephoned all of the head teacher s of the special education preschool and kindergarten classrooms in Alac hua County to describe the study in detail and ask for
17 their participation in recruitment of families. Teachers were asked to send home information packets about the study to parent s of children in thei r classroom who would likely meet inclusion criteria a nd benefit from treatment. The information packet included a cover letter describing th e study, brief demographic questionnaire, CBCL, two copies of the informed consent form (one to keep), and a self-addressed, stamped envelope. Parents who rated their child within the c linically significant ra nge on the Aggressive Behavior subscale of the CBCL were telephon ed and scheduled for an evaluation to complete study screening. Parents who rated their child within normal limits were offered information on other resources for parent training and parenting classes in the community. In addition to recruitment through special education classrooms, local health care professionals were provided information flyers to give to potential participant families in their practice. Families responding to the flyers were scheduled for a screening evaluation visit. These families completed the informed consent and all screening measures du ring that evaluation. At the screening evaluation visit, the WPPSI-III was administered to all children, and the CARS was completed based on observations made during the WPPSI-III administration. In addition, the ABS-S: 2, DISC -IV-P, and WPT were administered to the primary caregiver. Families that did not meet study criteria were given feedback on their childs behavior and cognitive and adaptive functioning, and appropriate recommendations or referrals were made. For families that met inclusion criteria for the study, the primary caregiver completed the ECBI and PSI-SF. Behavior al observations of parent-child interactions were also vide otaped during three 5-minute parent-child interaction situations (i.e., child led play, parent-led pl ay, and clean-up) with the primary
18 caregiver and were later code d using the DPICS-II by a tr ained undergraduate research assistant. For purposes of re liability coding, a second underg raduate research assistant coded one of the three situations, randomly selected for each family. Families received $10 for participation in the Time 1 assessment. Following the Time 1 assessment with each family, the SESBI-R was mailed to the childs teacher for completion along with a c onsent form signed by the parent permitting exchange of information with the school, a self-addressed, stamped envelope, and $5 in compensation for the teachers time. Children were blocked for sex and randomly assigned to either the IT or WL group. The mentor generated two lists of random numbers (one for boys, one for girls) and c oncealed these lists th roughout the duration of the study. After each family met criteria, they were assigned the next number on the list (even was WL and odd was IT), and the auth or received the group assignment from the mentor. Families were then contacted by telephone and informed of their group assignment immediately following the Time 1 a ssessment. For families in the IT group, two therapists were assigned to the family and asked to telephone the family to schedule the first PCIT session within a week. Families in the WL group were reminded that they would begin treatment in 4 months, following a second assessment. Four months after the Time 1 assessment, all families were telephoned to schedule the Time 2 assessment. For families in th e IT group, this was their post-treatment assessment [although 4 (40%) families were not completed with treatment at this point], and for families in the WL group, the Time 2 assessment was also their pre-treatment assessment. The Time 2 assessment session included the same parent-child observations and outcome measures (CBCL, ECBI, and PSI-S F) that were completed at the Time 1
19 assessment. An additional measure of consumer satisfaction (i.e., TAI) was included in the Time 2 assessment for the IT group onl y. Families received $15 in compensation for completion of the Time 2 assessment. The SESB I-R was also sent to the childs current teacher in the same manner as it was during the Time 1 assessment. Treatment PCIT sessions were conducted once a w eek and were approximately 1 hour in length. A treatment manual that provided written outlines in checklist form for each session was followed to help ensure treatme nt fidelity (Eyberg & Child Study Lab, 1999). Each family was seen individually by two therap ists. Lead therapists included the author (2 cases) and eight other graduate students in clinical child psychol ogy with prior training and experience as a PCIT therapist with at least two families. The co-therapists were graduate students and interns in clinical ps ychology who had read th e treatment protocol and observed a prior case. All therapists attended weekly group supervision with the author and his mentor. In PCIT, parents are taught skills to es tablish a nurturing a nd secure relationship with their child while increasing their childs prosocial behavior and decreasing negative behavior. Treatment progresses through two di stinct phases. Child-Directed Interaction (CDI) resembles traditional play therapy a nd focuses on strengthening the parent-child attachment, increasing positive parenting, a nd improving child social skills. ParentDirected Interaction (PDI) resembles clini cal behavior therapy and focuses on improving parents ability to set limits and follow through consistently, and on reducing child noncompliance and other negative behavior. The principles and skills of each phase of treatment were first taught to the parents alone in a teaching session, and in subsequent sessions parents were coached in the skills
20 as they played with their child. The parents we re asked to practice th e CDI skills at home during daily 5-minute play sessions. Families continued in treatment until the parents demonstrated mastery of the skills and thei r childs behavior came within standard deviation of the normative mean on the ECBI. Furthermore, the therapists worked actively to keep families in treatment until the completion criteria were met. The average length of treatment for the IT treatment completers ( n = 8) was 12 sessions ( SD = 1.77), which is similar to other PCIT outcome studi es. Fathers were involved in treatment (i.e., attended at least one session) in 4 (50%) of these families. The weekly ECBI Intensity and Problem T scores for these 8 families are show n in Figure 1. The la st observation was carried forward for those families who completed treatment before the 14th week so that each time point represents the same number of families. 40 45 50 55 60 65 70Wee k 1 Wee k 2 Wee k 3 We ek 4 We ek 5 We ek 6 Week 7 Week 8 Week 9 Wee k 1 0 Wee k 1 1 Wee k 1 2 Wee k 1 3 Wee k 1 4Session NumberT-Score ECBI Intensity Scale (n = 8) ECBI Problem Scale (n = 8) Figure 2-1. ECBI Intensity and Pr oblem T-Scores During Treatment During the CDI phase of treatment, the parent s learned to follow the childs lead in play by using the non-directive attending skil ls, called the Do skill s: Praising the child,
21 reflecting the childs statements, and descri bing the child's behavior. They learned to change child behavior by directing the Do skills to the childs appropriate play and consistently ignoring undesirabl e behaviors. Parents were al so taught to avoid using the Dont skills, which are directive/intrusiv e verbalizations a nd include questions, commands, and criticisms. During CDI coachin g sessions, therapists actively coached parents in their use of these skills until parents met criteria for skill mastery, as assessed during a 5-minute observation at the start of each session (i.e., 10 beha vioral descriptions, 10 reflections, and 10 labeled praises; and fewer than three questions, commands, and criticisms). For the current study, CDI was limited to five coaching sessions due to the time constraint of the Time 2 evaluation 4 months after the beginning of treatment. During the PDI phase of treatment, parents learned to direct th e childs behavior when necessary with effective commands and specific consequences for compliance and noncompliance. In PDI coaching sessions, parents worked toward meeting the mastery criteria of the PDI sk ills that serve as an indicator of their consistency. Throughout the PDI phase of treatment, the therapist guided the parents in applyi ng the principles and procedures of CDI and PDI to the childs beha vior at home and in ot her settings. Initially, parents were instructed to pr actice the PDI skills in brief 5to 10-minute practice sessions after the daily CDI play session. Homework a ssignments proceeded gradually to use of the PDI procedure only at times when it is important that the child obey a specific command. Mastery of the PDI skills was demonstrated when 75% of the parents commands were direct and there was 100% correct follow-through on commands. In the last few sessions, parents we re taught variations of the PDI procedure to deal with aggressive behavior and public misbehavior, as they approached mastery of the PCIT
22 skills and assumed increasing re sponsibility for applying the principles creatively to new situations that arise. Of the 15 IT families, 2 families (13%) dropped out of the study before attending their first treatment session, and 5 additiona l families (33%) di scontinued treatment before termination criteria were met. We reported one adverse event for one of the families that discontinued treatment prematurely. In this case, we received a call from the Department of Child and Family Services repor ting that the child had been taken to them with bruises on her body, but the perpetrator of abus e was unclear at the time. The family was still in treatment, but the mother was noncompliant with homework and not demonstrating skill changes in treatment. We referred the mother to a more basic parenting group addressing self-help skills and sa fety issues and asked her to practice her 5-minute CDI homework session four times before scheduling her next PCIT session, but she did not call back. Treatment Integrity To evaluate treatment integrity, all therapy sessions were videotaped. Undergraduate research assistants, who were different assistants th an the DPICS coders (to mask coders of group status), randomly se lected 50% of the session tapes from each family and recorded, on the integrity checklist s included in the treatment manual, the key elements of each session that were covered by the therapist. From this subset of tapes, 50% were again randomly selected and checked independently by a second undergraduate research a ssistant to provide an interobser ver reliability estimate of the treatment integrity data. Accuracy was 97% with the treatment protocol, and percent agreement interrater reliabilit y, calculated on 50% of comple ted checklists (34 sessions), was 97% (range = 79% 100%)
CHAPTER 3 RESULTS We were interested in the efficacy and ge neralization of PCIT with children with co-morbid disruptive behavior and MR and ex amined differences between the IT and WL groups at the Time 2 assessment in (a) pare nting behaviors during observed parent-child interactions; (b) child compliance during obser ved parent-child interactions; (c) child disruptive behaviors rated by parents (d) parent self-report of pare nting stress; and (e) child disruptive behaviors rated by teachers. The primary analyses used to determine treatment effects were 2 X 2 repeated measures analyses of variance (ANOVAs) with time (i.e., Time 1 and Time 2 assessments) as the within subjects variab le and group (i.e., IT and WL groups) as the between subjects variable. For all analyses, alpha was set at .05 because each dependent variable was of independent sc ientific interest. However, we reported exact alpha levels for all outcome variables to permit examinati on of potential experiment-wise error rate. Use of analysis of covariance (ANCOVA), with pre-treatment scores as the covariate, has also been recommended for randomized controlled trials because it is a more statistically powerful analytic method than a repeated measures ANOVA (Rausch, Maxwell, & Kelley, 2003). However, Miller and Chapman (2001) have described misuses of ANCOVA and suggested that it ma y artificially incr ease power because it removes shared variance between preand post-treatment scores. Given the current debate in the literature about the most appropr iate statistical appro ach, we have reported our findings using repeated measures ANOVAs and have included results using 23
24 ANCOVA in Appendix A. Any differences in significant findings between the two statistical approaches have b een noted in the text, however. Although all analyses were conducted with the study completers (n = 22), we also examined results when using only the 8 treatm ent completers in the IT group compared to the 12 study completers in the WL group. In addition, we analyzed the results without including the childre n (1 in the IT group and 3 in the WL group) who changed medication status between the Time 1 and Ti me 2 assessment points. Results for these analyses are included in Appendix B and C, respectively. Figure 3-1 illustrates the participant flow throughout the course of the study. Assessed for eli g ibilit y ( n = 58 ) Excluded ( n = 28) Did not meet inclusion criteria ( n = 26) Refused to participate ( n = 1) Did not attend Time 1 assessment ( n = 1 ) Allocated to intervention (n = 15) Received intervention ( n = 13) Did not attend first session ( n = 2) Completed allocated interventi on ( n = 8) Allocated to waitlist (n = 15) Completed waitlist ( n = 12) Did not complete waitlist ( n = 3) Analyzed ( n = 10) Did not complete Time 2 assessment ( n = 5) Analyzed ( n = 12) Did not complete Time 2 assessment ( n = 3) Did not complete Observations at Time 2 assess men t ( n = 1) Randomized ( n = 30) Figure 3-1. Participant Flow Diagram
25 Observed Parent-Child Interaction One WL family completed the Time 2 assessment by phone and was unable to come in for the observation due to lack of transportation, resulting in a sample size of 21 mother-child dyads for analyses of observa tional data. Results re vealed a significant group by time interaction for both the CDI Do skills, F (1, 19) = 19.53, p < .001, and CDI Dont skills, F (1, 19) = 7.31, p = .014. The very large effect sizes between groups indicate that mothers in th e IT group used significantly more CDI Do skills and fewer CDI Dont skills during the child-led play si tuation at the Time 2 assessment than the mothers in the WL group. For child complian ce, there was also a significant group by time interaction, F (1, 19) = 8.61, p = .009. The very large effect size between groups indicated that children in th e IT group were significantly more compliant to maternal commands during the combined parent-led play and clean up situati ons than children in the WL group. Mean scores and the effect si zes of change between the IT and WL group on the observational measures of parent-child interaction at the Time 1 and Time 2 assessment points are presented in Table 3-1. Table 3-1. Mean Scores for Observational Meas ures of Parent-Child Interaction at Time 1 and 2 Assessments Time 1 Time 2 Measure Group M SD M SD p d a Do skills IT 5.60 2.63 18.10 8.90 < .001 1.43 WL 4.91 2.66 3.64 4.86 Dont skills IT 43.90 14.45 10.00 9.94 .014 1.06 WL 43.18 16.78 28.36 18.29 % Child compliance b IT 63.88 19.22 85.20 9.44 .009 1.10 WL 68.89 19.71 59.72 25.68 Note. IT = immediate treatment ( n = 10), WL = wait-list control ( n = 11). a Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. b Percent compliance was averaged across both parent-directed play and clean-up observations.
26 Results using ANCOVA for these observational measur es led to the same findings as the interaction effects from the repeated measures ANOVA. Parent Report Results revealed a significant group by time interaction for both the CBCL Externalizing Scale, F (1, 20) = 8.08, p = .010, and Total Scale, F (1, 20) = 9.07, p = .007. There were very large effect sizes between groups, indica ting that mothers in the IT group reported substantially fewer behavior problems at the Time 2 assessment than mothers in the WL group. Similarly, there was a significant group by time interaction and a very large effect size for the ECBI Intensity Scale, F (1, 20) = 11.71, p = .003. Although the group by time interaction for the ECBI Problem Scale did not meet statistical significance, F (1, 20) = 3.68, p = .069, differences between the groups yielded a medium effect size. Results using ANCOVA for the CBCL and ECBI led to the same findings as the interaction effects from the repeated measures ANOVA. For parenting stress, the group by time interaction was not significant for the Parental Distress subscale, F (1, 20) = .15, p = .703, and differences between the IT and WL groups resulted in a small effect size. Although the group by tim e interaction did not meet statistical significance for either the Parent-Child Dysf unctional Interaction, F (1, 20) = 2.99, p = .099, or the Difficult Child subscales, F (1, 20) = 3.55, p = .074, the differences between the groups on these two scal es yielded medium to large effect sizes. When using ANCOVA, there was a significant difference between the IT and WL groups on the Difficult Child subscale, F (1, 19) = 4.80, p = .041, but results were not significant on the Parental Distress, F (1, 19) = .19, p = .671, or Parent-Child Dysfunctional Interaction, F (1, 19) = 2.59, p = .124, subscales. Table 3-2 shows the Time 1 and Time 2
27 mean scores for the 22 study completers and th e effect sizes of change between the IT and WL groups on each outcome measure. Table 3-2. Mean Scores for Mothers on Pa rent Report Measures at Time 1 and 2 Assessments Time 1 Time 2 Measure a Group M SD M SD p d b Child Behavior Checklist Externalizing IT 34.60 7.73 19.60 10.72 .010 1.05 WL 36.25 6.25 31.25 8.56 Total IT 89.70 29.45 51.90 27.87 .007 1.12 WL 95.17 16.41 83.83 20.44 Eyberg Child Behavior Inventory Intensity IT 156.40 34.30 94.60 26.22 .003 1.27 WL 170.92 19.47 148.17 30.33 Problem IT 21.40 6.11 10.10 9.42 .069 .50 WL 18.67 7.98 14.67 8.74 Parenting Stress Index-Short Form Parent Distress IT 30.60 4.70 30.00 5.72 .703 .18 WL 30.17 7.80 28.67 8.47 Parent-Child Dysfunctional Interaction IT 31.10 10.10 26.20 8.93 .099 .38 WL 29.33 6.75 29.58 8.82 Difficult Child IT 42.60 8.40 32.60 8.87 .074 .77 WL 43.67 7.79 38.98 6.80 Note: IT = immediate treatment ( n = 10), WL = wait-list control ( n = 12). a Scores for all measures ar e reported as raw scores. b Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. Generalization of Treatment E ffects to the School Setting Repeated Measures ANOVA yielded nonsigni ficant group by time interactions for both the Intensity Scale, F (1, 12) = .20, p = .662, and Problem Scale, F (1, 12) = .01, p = .943, of the SESBI-R. Differences between the IT and WL groups for the SESBI-R Intensity and Problem Scales yielded small effect sizes. Results using ANCOVA for the SESBI-R led to the same findings as the group by time interactions from the repeated measures ANOVA. Mean scores and the effect sizes of change be tween the IT and WL
28 group on the SESBI-R for the 14 children in scho ol at the time of the Time 1 and Time 2 assessments are presented in Table 3-3. Table 3-3. Mean SESBI-R Scores for Teachers at Time 1 and 2 Assessments Time 1 Time 2 Measure a Group M SD M SD p d b SESBI-R Intensity IT 142.25 30.73 126.38 19.46 .662 .25 WL 140.17 38.77 134.83 48.88 SESBI-R Problem IT 10.88 10.63 9.00 7.80 .943 .08 WL 12.00 9.27 9.67 9.99 Note: IT = immediate treatment ( n = 8), WL = wait-list control ( n = 6), SESBI-R = Sutter-Eyberg Student Behavior Inventory Revised. a Scores on all measures are reported in raw scores. b Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. Consumer Satisfaction The TAI was administered only to the IT group at the Time 2 assessment, so no group comparisons were conducted. Across all 10 study completers in the IT group, TAI scores ranged from a 42 to 50, with a mean of 46.40 ( SD = 2.88). These findings suggest these families were highly satisfied with PCIT. There was no significant difference between the 8 families that successfully completed treatment ( M = 46.88; SD = 2.75) and the 2 families that did not successfully complete treatment ( M = 44.50; SD = 3.54), t (8) = 1.05, p = .324. Even though the TAI was mailed w ith a self-addressed, stamped envelope to the 5 IT families who dropped out of treatment before completing the Time 2 assessment (i.e., IT study dropouts), none of these families sent it back. Thus, we were unable to make comparisons between th e 5 IT study dropouts and the 10 IT study completers. Intent-To-Treat Analyses In addition to the comparisons between groups for families who completed the Time 2 assessment, outcome was assessed using an intent-to-treat an alysis of the CBCL
29 Externalizing Scale, the primary outcome measure. We used a single imputation approach to estimate the dropout families scor es based on available data. Mothers in the IT group completed the ECBI on a weekly ba sis prior to each treatment session as a measure of treatment progress, so their last score before drop ping out is the best estimate of the childs functioning at the Time 2 assessment. In estimating the CBCL Externalizing scor es for families that dropped out before completing the Time 2 assessment, we first conducted a linear regression for all study completers with the last observed ECBI Inte nsity raw score as the independent variable. We then used the regression equation from this model, Y = .23X 1.98, to predict the Time 2 CBCL Externalizing scores for th e dropout families based on their last ECBI score. For the 2 IT families that dropped out of the study before attending their first treatment session and the 3 WL families who did not complete the Time 2 assessment, the Time 1 ECBI Intensity raw score was used because that score was their last ECBI score. Using this method, results of the repe ated measures ANOVA showed a significant group by time interaction, F (1, 28) = 5.09, p = .032, d = .54. Clinical Significance Clinically significant change on outcome measures with established cutoff scores was determined using the twofold criteria suggested by Jacobson, Roberts, Berns, and McGlinchey (1999): (a) the magn itude of change had to be statistically reliable and (b) families had to be in a range that rendered them indistinguishable from well-functioning individuals. To determine the reliability of the magnitude of change, a reliable change index (RCI; Jacobson, Follette, & Re venstorf, 1984) was used to ensure that the degree of change exceeded the margin of measuremen t error. The formula for calculating the RCI involved dividing the magnitude of change between the Time 1 and Time 2 assessment
30 scores by the standard erro r of the difference score. RCIs greater than 1.96 were considered to be sufficient in magnitude (Jacobson et al., 1999). For the second criterion, in which families had to be indistinguishable from well functioning individuals, families had to be at or above the published cuto ff value for each measure at the Time 1 assessment and below the cutoff value at the Time 2 assessment. Rates of clinically significant change we re compared for study completers in the IT and WL groups. On the CBCL Externalizing S cale, clinically significant gains at the Time 2 assessment were found for 7 (70%) ch ildren in the IT group and 2 (17%) children in the WL group. On the CBCL Total Scale, c linically significant ga ins were found for 4 (40%) children in the IT gr oup and 2 (17%) children in th e WL group. A similar pattern was seen on the ECBI Intensity Scale, where 5 (50%) children in the IT group showed clinically significant improveme nts at the Time 2 assessment compared to 1 (8%) in the WL group, and on the ECBI Problem Scale, where 6 (60%) mothers in the IT group showed clinically significant improvements at the Time 2 assessment compared to 4 (33%) in the WL group. For the PSI-SF Parental Distress subscal e, no mothers in the IT group and 1 (8%) mother in the WL group show ed clinically significant im provements at the Time 2 assessment. As shown in Table 3-4, howeve r, few mothers in the IT (20%) and WL (50%) groups had scores in the clinically significant rang e at the Time 1 assessment, leaving little room for clinically significant improvements. For the PSI-SF Parent-Child Dysfunctional Interaction subs cale, only 2 (20%) mothers in the IT group and 1 (8%) mother in the WL group show ed clinically significant im provements at the Time 2 assessment. However, on the PSI-SF Difficult Child subscale, 5 (50%) mothers in the IT
31 group showed clinically signifi cant improvements at the Time 2 assessment compared to only 1 (8%) mother in the WL group. Rates of clinically signifi cant generalization to school were also examined for the children in the IT and WL groups who we re in school during the Time 1 and 2 assessments ( n = 14). For the SESBI-R Intensity Scale, 2 (25%) children in the IT group showed clinically significant improvements in di sruptive behavior at school in contrast to 1 child (17%) in the WL group. For the SESBI-R Problem Scale, 1 (13%) child in the IT group showed clinically signifi cant improvements at the Time 2 assessment compared to 1 (17%) child in the WL group. As shown in Table 3-4, most children were not rated in the clinically significant range on the SESBI-R Intensity or Pr oblem Scales at the Time 1 assessment, which may explain the lower rate s of clinically signi ficant improvements on this instrument. For the study co mpleters, the number of famili es in each group that were above the cutoff value (i.e., in the clinically significant range) at the Time 1 and 2 assessments and that displayed clinically si gnificant improvements are presented in Table 3-4. Table 3-4. Number of Families in the Clinically Significant Range at the Time 1 and 2 Assessments and Clinically Significant Change Significant at Time 1 Significant at Time 2 Significant Change a Measure Group n Frequency (%) Frequency (%) Frequency (%) Child Behavior Checklist Externalizing IT 10 10 (100) 3 (30) 7 (70) WL 12 11 (92) 9 (75) 2 (17) Total IT 10 7 (70) 3 (30) 4 (40) WL 12 12 (100) 10 (83) 2 (17) Eyberg Child Behavior Inventory Intensity IT 10 7 (70) 2 (20) 5 (50) WL 12 11 (92) 9 (75) 1 (8) Problem IT 10 8 (80) 3 (30) 6 (60) WL 12 10 (83) 5 (42) 4 (33)
32 Table 3-4. Continued Significant at Time 1 Significant at Time 2 Significant Change a Measure Group n Frequency (%) Frequency (%) Frequency (%) Parental Distress IT 10 2 (20) 4 (40) 0 (0) WL 12 6 (50) 5 (42) 1 (8) Parent-Child Dysfunctional Interaction IT 10 7 (70) 4 (40) 2 (20) WL 12 8 (67) 8 (67) 1 (8) Difficult Child IT 10 9 (90) 5 (50) 5 (50) WL 12 10 (83) 10 (83) 1 (8) Sutter-Eyberg Student Behavior Inventory Intensity IT 8 4 (50) 2 (25) 2 (25) WL 6 3 (50) 4 (67) 1 (17) Problem IT 8 2 (25) 1 (13) 1 (13) WL 6 2 (33) 2 (33) 1 (17) Note: IT = immediate treatment, WL = wait-list control. a Significant change was reporte d based on the twofold criteria (Jacobson et al., 1984).
CHAPTER 4 DISCUSSION Findings from this study provide empiri cal support for the use of PCIT with children who have co-morbid ODD and MR. As expected, mothers in the IT group effectively changed their interactional style w ith their children and, in turn, the children showed statistically and clinically signifi cant improvements in their behavior. After treatment, mothers were observed to interact more positively with their children. They successfully learned to attend positively to th eir childs appropriat e behaviors and avoid using directive/intrusive verbal izations when instructed to follow their childs lead in play. When required to direct their childs be havior, mothers in the IT group were more successful in gaining their childs compliance, and the change in rates of child compliance was similar to previous research (Eisenstadt et al., 1993; Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998). Children in this study were s ubstantially more compliant at pre-treatment than children in previous PCIT efficacy trials with nondelayed disruptive children, however. This disp arity in pre-treatment levels may suggest differences in the presentation of ODD in ch ildren with and without MR and highlights the need for further research. Child disruptive behavior also significan tly improved based on the parent-report measures. Mothers in the IT group perceive d significant improvements in the childs behavior at home. Furthermore, most of the mothers reports of their childs disruptive behavior following treatment were within nor mal limits, indicating that the changes were clinically as well as statistically significan t. Two of the 3 mothers who reported their 33
34 childs behavior within the clinically signi ficant range at the Time 2 assessment were from the two families that completed the second evaluation (i.e., study completers) but then dropped out of treatment unilaterally before reaching treatment completion goals. The two families had not attended fewer sessions than other families at the 4-month assessment point. Therefore, these results sugge st that mothers repor ts of their childs behavior 4 months after be ginning treatment may be an important indicator of the likelihood of successful completion with continuing treatment. Parents who perceive relatively few changes in their childs behavior after participating in a treatment program for 4 months may become discouraged and discontinue treatment prematurely due to decreased motivation. Mothers reports of parenting stress due to dysfunctional parent-child interactions or disruptive child behavior did not signifi cantly change following PCIT. These findings are consistent with Roberts et al. (2006) a nd different than results with families of normally developing children with disruptive behavior and may be due to the mothers lower levels of parenting stress at pre-treatm ent. However, the moderate to large effect sizes between groups suggests th at these results may be due to an inadequate sample size to detect differences that may be present. In fact, using an ANCOVA with pretreatment scores as covariates, rather than a re peated measures ANOVA, yielded significant findings for parenting stress from disruptive be havior. Maternal dist ress unrelated to the child showed only a small treatment effect, lik ely due to the relatively little maternal distress in this popula tion before treatment. The absence of generalization of gains to the school setting was unexpected. This finding may be due to several methodological lim itations in addressing this question. The
35 timing of the study resulted in approximatel y one-third of the ch ildren (36%) among the study completers not being in school at th e time of the first or second assessments, resulting in a particularly small sample si ze for these analyses. Second, children were not selected for school problems initially, and only half the children had clinically significant problem behaviors at school, with even fewer teachers finding their behavior problematic. Thus, the sample was not ade quate to address the question of school generalization. Because the school behavior of children with MR is so important for their learning, however, it will be important to re plicate this study with children presenting early disruptive classroom beha viors. It will also be impor tant to include observational measures of classroom behaviors to obtain information that may be less vulnerable to positive bias in special education. This study has additional limitations that s hould be acknowledged. First, we did not collect follow-up data and are unable to demons trate the long-term effectiveness of PCIT for children with co-morbid disruptive behavior and MR. Previous research on PCIT with normally developing children with disruptive behavior has demonstrated durable treatment gains (Boggs et al., 2004; Eyberg et al., 2001; Schuhmann et al., 1998), with maintenance of child behavioral improvement s lasting up to 6 years following treatment completion (Hood & Eyberg, 2003). Several factors may influence the maintenance of treatment gains, such as the frequency of parent practice of the newly acquired skills (Schuhmann et al., 1998). For parents of children with MR, these factors may be particularly salient. For example, it is possible that parents would be less likely to co ntinue using PCIT skil ls because they learn new and different home practice skills required in rehabilitative services for their child
36 (e.g., speech therapy). However, mothers of children in this study were highly satisfied with the content and process of PCIT, wh ich may be important for maintenance of treatment gains. Follow-up dropouts have reported lower consumer satisfaction following treatment than those who complete treatment (Luk et al., 2001). Thus, it is important to examine the long-term effectiv eness of PCIT with childre n with co-morbid disruptive behavior and MR in future research. A second potential limitation of this study is the influence of differential attrition. Although not statistically significant, the study dropout rate in the IT group (33%) was higher than in the WL group (20%), and two additional families dropped out of treatment after the Time 2 assessment. We conducted an in tent-to-treat analysis to address this issue statistically, however, and the single imputation procedure su ggested that the significant differences we obtained between groups were not likely a function of attrition. Still, premature termination is a substantial problem in child therapy ge nerally and prevention of treatment attrition must be studied directly in future research with this and other populations. Finally, the generalizability of our findings is limited. The data, including both observational and parent-report measures, were collected only from the primary caregiver, which was the mother in all familie s. Even though fathers were encouraged to participate in treatment, our results cannot be generalized to childrens behaviors with their fathers, fathers perceptions of change s in child behavior, or fathers own stress. Mothers and fathers display differences in thei r ratings of and intera ctions with children with disruptive behavior (Calzada, Rich, Eyberg, & Querido, 2004), and results in this study may have differed for fathers. Families th at participated in this study received
37 treatment at no charge and may differ from families that pay for services. In addition, managed care may limit the number of sessions covered, which might not be adequate for treatment gains similar to those found in this study. Despite these limitations, our results provide import ant information about the efficacy of PCIT for children with co-morbid disruptive behavior and MR. The high rates of disruptive behavior in child ren with MR often go untreated and have been neglected in the evidence-based treatment literature. This study provides encouraging evidence for the application of an evidence-based treatment for disruptive be havior with children with developmental disabilities, and further resear ch in this area is needed. Although children with autism spectrum disorders were not in cluded in this study, their high rates of disruptive behavior (Howlin, 1998) underscore the importance of examining PCIT with this population as well.
APPENDIX A ANALYSES USING ANCOVA Table A-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments Using ANCOVA Time 1 Time 2 Measure Group M SD M SD p d a Do skills IT 5.60 2.63 18.10 8.90 < .001 1.43 WL 4.91 2.66 3.64 4.86 Dont skills IT 43.90 14.45 10.00 9.94 .006 1.06 WL 43.18 16.78 28.36 18.29 % Child compliance b IT 63.88 19.22 85.20 9.44 .006 1.10 WL 68.89 19.71 59.72 25.68 Note. IT = immediate treatment ( n = 10), WL = wait-list control ( n = 11). a Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. b Percent compliance was averaged across both parent-directed play and clean-up observations. Table A-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments Using ANCOVA Time 1 Time 2 Measure a Group M SD M SD p d b Child Behavior Checklist Externalizing IT 34.60 7.73 19.60 10.72 .009 1.05 WL 36.25 6.25 31.25 8.56 Total IT 89.70 29.45 51.90 27.87 .003 1.12 WL 95.17 16.41 83.83 20.44 Eyberg Child Behavior Inventory Intensity IT 156.40 34.30 94.60 26.22 .002 1.27 WL 170.92 19.47 148.17 30.33 Problem IT 21.40 6.11 10.10 9.42 .118 .50 WL 18.67 7.98 14.67 8.74 Parenting Stress Index-Short Form Parent Distress IT 30.60 4.70 30.00 5.72 .671 .18 WL 30.17 7.80 28.67 8.47 Parent-Child Dysfunctional Interaction IT 31.10 10.10 26.20 8.93 .124 .38 WL 29.33 6.75 29.58 8.82 38
39 Table A-2. Continued Time 1 Time 2 Measure a Group M SD M SD p d b Difficult Child IT 42.60 8.40 32.60 8.87 .041 .77 WL 43.67 7.79 38.98 6.80 Note: IT = immediate treatment ( n = 10), WL = wait-list control ( n = 12). a Scores for all measures ar e reported as raw scores. b Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. Table A-3. Mean SESBI-R Scores for Teach ers at Time 1 and 2 Assessments Using ANCOVA Time 1 Time 2 Measure a Group M SD M SD p d b SESBI-R Intensity IT 142.25 30.73 126.38 19.46 .653 .25 WL 140.17 38.77 134.83 48.88 SESBI-R Problem IT 10.88 10.63 9.00 7.80 .931 .08 WL 12.00 9.27 9.67 9.99 Note: IT = immediate treatment ( n = 8), WL = wait-list control ( n = 6), SESBI-R = Sutter-Eyberg Student Behavior Inventory Revised. a Scores on all measures are reported in raw scores. b Cohens d = effect size of change between IT and WL groups at the Time 2 assessment.
APPENDIX B ANALYSES WITH TREATMENT COMPLETERS Table B-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments With Treatment Completers Time 1 Time 2 Measure Group M SD M SD p d a Do skills IT 4.75 2.12 18.13 9.37 < .001 1.44 WL 4.91 2.66 3.64 4.86 Dont skills IT 45.13 12.93 7.75 7.98 .008 1.15 WL 43.18 16.78 28.36 18.29 % Child compliance b IT 65.14 21.58 83.85 10.07 .027 1.02 WL 68.89 19.71 59.72 25.68 Note. IT = immediate treatment ( n = 8), WL = wait-list control ( n = 11). a Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. b Percent compliance was averaged across both parent-directed play and clean-up observations. Table B-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments With Treatment Completers Time 1 Time 2 Measure a Group M SD M SD p d b Child Behavior Checklist Externalizing IT 34.00 7.19 15.88 7.92 .007 1.37 WL 36.25 6.25 31.25 8.56 Total IT 89.50 26.30 47.38 29.50 .004 1.27 WL 95.17 16.41 83.83 20.44 Eyberg Child Behavior Inventory Intensity IT 153.13 25.17 85.25 30.21 .001 1.46 WL 170.92 19.47 148.17 30.33 Problem IT 21.00 4.31 7.25 7.46 .022 .84 WL 18.67 7.98 14.67 8.74 Parenting Stress Index-Short Form Parent Distress IT 30.88 5.11 29.50 6.28 .960 .11 WL 30.17 7.80 28.67 8.47 Parent-Child Dysfunctional Interaction IT 29.63 7.39 25.25 9.33 1.45 .48 WL 29.33 6.75 29.58 8.82 40
41 Table B-2. Continued Time 1 Time 2 Measure a Group M SD M SD p d b Difficult Child IT 42.63 6.78 31.75 9.79 .032 .83 WL 43.67 7.79 38.98 6.80 Note: IT = immediate treatment ( n = 8), WL = wait-list control ( n = 12). a Scores for all measures ar e reported as raw scores. b Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. Table B-3. Mean SESBI-R Scores for Teach ers at Time 1 and 2 Assessments With Treatment Completers Time 1 Time 2 Measure a Group M SD M SD p d b SESBI-R Intensity IT 148.29 27.60 127.86 20.53 .553 .20 WL 140.17 38.77 134.83 48.88 SESBI-R Problem IT 11.86 11.08 8.27 8.32 .886 .13 WL 12.00 9.27 9.67 9.99 Note: IT = immediate treatment ( n = 7), WL = wait-list control ( n = 6), SESBI-R = Sutter-Eyberg Student Behavior Inventory Revised. a Scores on all measures are reported in raw scores. b Cohens d = effect size of change between IT and WL groups at the Time 2 assessment.
APPENDIX C ANALYSES WITH MEDICALLY STABLE CHILDREN Table C-1. Mean Scores for Observational Measures of Parent-Child Interaction at Time 1 and 2 Assessments With Medically Stable Children Time 1 Time 2 Measure Group M SD M SD p d a Do skills IT 5.78 2.73 18.00 9.43 .002 1.44 WL 4.50 2.00 2.75 3.96 Dont skills IT 46.44 12.73 11.00 10.00 .036 1.06 WL 39.75 17.03 22.88 9.49 % Child compliance b IT 62.65 19.96 86.11 9.54 .016 1.25 WL 63.79 20.35 54.09 27.92 Note. IT = immediate treatment ( n = 9), WL = wait-list control ( n = 8). a Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. b Percent compliance was averaged across both parent-directed play and clean-up observations. Table C-2. Mean Scores for Mothers on Parent Report Measures at Time 1 and 2 Assessments With Medically Stable Children Time 1 Time 2 Measure a Group M SD M SD p d b Child Behavior Checklist Externalizing IT 33.78 7.73 18.11 10.22 .003 1.28 WL 35.89 6.86 32.89 7.42 Total IT 85.33 27.59 47.11 24.82 .003 1.35 WL 95.33 17.44 88.44 20.72 Eyberg Child Behavior Inventory Intensity IT 152.11 33.41 90.00 35.18 .003 1.32 WL 168.22 21.75 148.78 31.86 Problem IT 20.78 6.14 9.00 9.29 .018 .59 WL 16.56 8.03 14.11 7.72 Parenting Stress Index-Short Form Parent Distress IT 30.22 4.82 29.56 5.88 .816 .12 WL 30.00 8.86 28.67 9.54 Parent-Child Dysfunctional Interaction IT 29.78 9.76 24.78 8.18 .116 .53 WL 29.11 7.29 29.78 10.28 42
43 Table C-2. Continued Time 1 Time 2 Measure a Group M SD M SD p d b Difficult Child IT 41.44 8.02 31.33 8.40 .052 .98 WL 43.89 7.22 40.00 7.37 Note: IT = immediate treatment ( n = 9), WL = wait-list control ( n = 9). a Scores for all measures ar e reported as raw scores. b Cohens d = effect size of change between IT a nd WL groups at the Time 2 assessment. Table C-3. Mean SESBI-R Scores for Teach ers at Time 1 and 2 Assessments With Medically Stable Children Time 1 Time 2 Measure a Group M SD M SD p d b SESBI-R Intensity IT 142.43 33.19 126.29 21.02 .071 1.44 WL 148.50 22.96 163.50 12.40 SESBI-R Problem IT 11.43 11.36 10.29 7.46 .811 .03 WL 13.25 8.77 10.50 11.00 Note: IT = immediate treatment ( n = 7), WL = wait-list control ( n = 4), SESBI-R = Sutter-Eyberg Student Behavior Inventory Revised. a Scores on all measures are reported in raw scores. b Cohens d = effect size of change between IT and WL groups at the Time 2 assessment.
LIST OF REFERENCES Abidin, R. (1995). Parenting Stress-Manual (3 rd ed.). Odessa, FL: Psychological Assessment Resources, Inc. Achenbach, T. M. (1992). Manual for the Child Behavior Checklist/2-3 and 1992 Profile. Burlington, VT: University of Verm ont Department of Psychiatry. Achenbach, T. M. & Rescorla, L. A. (2000). Manual of the ASEBA Preschool Forms & Profiles. Burlington, VT: University of Verm ont, Research Center for Children, Youth, & Families. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4 th ed., text revision). Wa shington, DC: Author. Bagner, D. M., & Eyberg, S. M. (2003). Fath er involvement in pa rent training: When does it matter? Journal of Clinical Child and Adolescent Psychology, 32 599-605. Benson, B. A. (1985). Behavior disorders and me ntal retardation: Associations with age, sex, and level of functioning in an outpatient clinic sample. Applied Research in Mental Retardation, 6, 79-85. Benson, B. A., & Aman, M. G. (1999). Disrupt ive behavior disorders in children with mental retardation. In H. C. Quay & A. E. Hogan (Eds.), Handbook of disruptive behavior disorders (pp.559-578). New York, NY: Kluwer Academic/Plenum Publishers. Boggs, S. R., Eyberg, S. M., Edwards, D. L., Rayfield, A., Jacobs, J., Bagner, D., et al. (2004). Outcomes of parent-child interaction therapy: A comparison of treatment completers and study dropouts one to three years later. Child and Family Behavior Therapy, 26, 1-22. Brestan, E. V., Foote, R. C., & Eyberg, S. M. (2006). Dyadic pare nt-child interaction coding system II (DPICS II): Reliability and validity with fa ther-child dyads. Manuscript submitted for publication. Brestan, E. V., Jacobs, J. R., Rayfield, A. D., & Eyberg, S. M. (1999). A consumer satisfaction measure for parent -child treatments and its re lationship to measures of child behavior change. Behavior Therapy, 30, 17-30. 44
45 Calzada, E. J., Eyberg, S. M., Rich, B., & Querido, J. G. (2004). Parenting disruptive preschoolers: Experiences of mothers and fathers. Journal of Abnormal Child Psychology, 32, 203-213. Dodrill, C. B. (1981). An economic method fo r the evaluation of ge neral intelligence in adults. Journal of Consulting and Clinical Psychology, 49 668-673. Durand, V. M. (2001). Future directions fo r children and adoles cents with mental retardation. Behavior Therapy, 32, 633-650. Eisenstadt, T. H., Eyberg, S. M., McNeil, C. B., Newcomb, K., & Funderburk, B. (1993). Parent-child interaction therapy with behavior problem children: Relative effectiveness of two stages and overall treatment outcome. Journal of Clinical Child Psychology, 22, 42-51. Eisenstadt, T. H., McElreath, L. S., Eyberg, S. M., & McNe il, C. B. (1994). Interparent agreement on the Eyberg Child Behavior Inventory. Child and Family Behavior Therapy, 16, 21-28. Eyberg, S. M. (1993). Consumer satisfaction measures for assessing parent training programs. In L. VandeCreek, S. Knapp, & T. L. Jackson (Eds .), Innovations in clinical practice: A sourcebook (Vol. 12). Sarasota, FL : Professional Resource Press. Eyberg, S. M., Bessmer, J., Newcomb, K., Ed wards, D., & Robinson, E. (1994). Dyadic Parent-child interaction coding sy stem second edition: A manual. Social and Behavioral Sciences Documents (Ms. No. 2897). San Rafael, CA: Select Press. Eyberg, S. M., Boggs, S. R., & Rodriguez, C. M. (1992). Relationshi ps between maternal parenting stress and child disruptive behavior. Child and Family Behavior Therapy, 14, 1-9. Eyberg, S. M., & Child Study Lab (1999). Parent-child interacti on therapy: Procedures manual. Unpublished manuscript, University of Florida, Gainesville. Eyberg, S. M., Edwards, D. Bessmer, J., & Litwins, N. (1994). The workbook: A coder training manual for the dyadic parent-c hild interaction coding system II. Social and behavioral sciences documents (Ms. No. 2898). San Rafael, CA: Select Press. Eyberg, S. M., Funderburk, B., Hembree-Kigi n, T., McNeil, C., Querido, J., & Hood, K. (2001). Long-term effectiveness of parent-c hild interaction therapy: A two-year follow-up. Child and Family Behavior Therapy, 23 1-20. Eyberg, S. M., & Matarazzo, R. (1980). Trai ning parents as therapists: A comparison between individual parent-c hild interaction training and parent group didactic training. Journal of Clinical Psychology, 36, 492-499.
46 Eyberg, S. M., Nelson, M. M., & Boggs, S. R. (in press). Empirically supported treatments for Oppositional and Conduct Disorders. Journal of Clinical Child and Adolescent Psychology. Eyberg, S. M., & Pincus, D. (1999). Eyberg Child Behavior Inventory and Sutter-Eyberg Student Behavior Inventory: Professional Manual. Odessa, FL: Psychological Assessment Resources. Eyberg, S. M., & Robinson, E. A. (1983). C onduct problem behavior: Standardization of a behavior rating scal e with adolescents. Journal of Clinical Child Psychology, 12 347-357. Frick, P. J. (1998). Conduct disorders and severe antisocial behavior. New York: Plenum. Funderburk, B. W., Eyberg, S. M., Newcom b, K., McNeil, C. B., Hembree-Kigin, T. & Capage, L. (1998). Parent-child inter action therapy with behavior problem children: Maintenance of treatment effects in the school setting. Child and Family Behavior Therapy, 20, 17-38. Funderburk, B. W., Eyberg, S. M, Rich, B. A., & Behar, L. (2003). Further psychometric evaluation of the Eyberg and Behar rati ng scales for parents and teachers of preschoolers. Early Education and Development, 14 67-81. Handen, B. L. (1998). Mental retardation. In E. J. Mash & R. A. Barkley (Eds.), Treatment of childhood disorders (2 nd ed., pp. 369-415). New York, NY: The Guilford Press. Harris, S. L., Alessandri, M., & Gill, M. J. (1991). Training parents of developmentally disabled children. In J. L. Matson & J. A. Mulick (Eds.), Handbook of mental retardation (2 nd ed., pp. 373-381). New York, NY: Pergamon Press. Hollingshead, A. B. (1975). Four factor inde x of social status. Unpublished manuscript. Yale University, New Haven, Ct. Hood, K. K., & Eyberg, S. M. (2003). Outcomes of parent-child in teraction therapy: Maintenance three to six years after treatment. Journal of Clinical Child and Adolescent Therapy, 32 419-429. Howlin, P. (1998). Practitioner review: Psychol ogical and educational treatments for autism. Journal of Child Psychology and Psychiatry, 39 307-322. Jacobson, J. W. (1982). Problem behavior and psychiatric impairment within a developmentally disabled populat ion: I. Behavior frequency. Applied Research in Mental Retardation, 3, 121-139.
47 Jacobson, N., Follette, W., & Revenstorf, D. (1984). Psychotherapy outcome research: Methods for reporting variability and evaluating clinical significance. Behavior Therapy, 15, 336-352. Jacobson, N., Roberts, L., Berns, S., & McG linchey, J. (1999). Methods for defining and determining the clinical significance of tr eatment effects: Description, application, and alternatives. Journal of Consulting and Clinical Psychology, 67 300-307. Jensen, P. S. Watanabe, H. D., Richters, J. E ., Roper, M., Hibbs, E. D., Salzberg, A. D., et al. (1996). Scales, diag noses, and child psychopathology: II. Comparing the CBCL and the DISC against external validators. Journal of Abnormal Child Psychopathology, 24 151-168. Kazdin, A. E. (1996). Dropping out of child psychotherapy: Issues for research and implications for practice. Clinical Child Psychology and Psychiatry, 1 133-156. Kazdin, A. E. (2002). The state of child and adolescent psychotherapy research. Child and Adolescent Mental Health, 7 53-59. Kobe, F. H., & Mulick, J. A. (1995). Mental retardation. In R. T. Ammerman & M. Hersen (Eds.), Handbook of child behavior therapy in the pediatric setting (pp. 153-180). New York, NY: Wiley. Lambert, N., Nihira, K., & Leland, H. (1993). Manual for the Adaptive Behavior Scale School: Second Edition Austin, TX: ProEd. Lambert, N. M., Windmiller, M., Co le, L., & Figueroa, R. A. (1975). Standardization of a public school version of the AAM D Adaptive Behavior Scale. Mental Retardation, 13, 3-7. Loeber, R., Burke, J. D., Lahey, B. B., Winters, B. A., & Zera, B. A. (2000). Oppositional defiant and conduct disorder: A review of the past 10 years, part 1. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 14681484. Luk, E., Staiger, P., Mathai, J., Wong, L., Bi rleson, P., & Adler, R. (2001). Evaluation of outcome in child and adolescent mental he alth services: Children with persistent conduct problems. Clinical Child Psychology and Psychiatry, 6 109-124. McNeil, C., Eyberg, S., Eisenstadt, T., Ne wcomb, K., & Funderburk, B. (1991). Parentchild interaction therapy with behavi or problem children Generalization of treatment effects to the school setting. Journal of Clinical Child Psychology, 20 140-151. Miller, G. A., & Chapman, J. P. (2001). Mi sunderstanding analysis of covariance. Journal of Abnormal Psychology, 110 40-48.
48 Querido, J. G., & Eyberg, S. M. (2003). Ps ychometric properties of the Sutter-Eyberg Student Behavior Inventory-Revised with preschool children. Behavior Therapy, 34, 1-15. Rausch, J. R., Maxwell, S. E., & Kelley, K. (2003). Analytic methods for questions pertaining to a randomized pretes t, posttest, follow-up design. Journal of Clinical Child and Adolescent Psychology, 32 467-486. Reid J. B. (1993). Prevention of conduct disord er before and after sc hool entry: Relating interventions to developmental findings. Development and Psychopathology, 5, 243-262. Reitman, D., Currier, R. O., & Stickle, T. R. (2002). A critical evalua tion of the Parenting Stress Index Short Form (PSISF) in a Head Start population. Journal of Clinical Child and Adolescent Psychology, 2002 384-392. Roberts, C., Mazzucchelli, T., Studman, L., & Sanders, M. (2006). Behavioral family intervention for children with developmenta l disabilities and behavioral problems. Journal of Clinical Child and Adolescent Psychology, 35 180-193. Roberts, C., Mazzucchelli, T., Taylor, K., & Reid, R. (2003). Early intervention for behaviour problems in young children w ith developmental disabilities. International Journal of Disability, Development, and Education, 50 276-292. Ross, C. N., Blanc, H. M., McNeil, C. B., Eyberg, S. M., & Hembree-Kigin, T. L. (1998). Parenting stress in mothers of young child ren with oppositional defiant disorder and other severe behavior problems. Child Study Journal, 28, 93-110. Sanders, M. R. (1999). Triple P-Positive Parenting Program: Towards an empirically validated multilevel parenting and family support strategy for the prevention of behavior and emotional problems in children. Clinical Child and Family Psychology Review, 2 71-90. Schopler, E., Reichler, R. J., & Renner, B. R. (1988). The Childhood Autism Rating Scale (CARS). Los Angeles, CA: Western Psychological Services. Schuhmann, E. M. (2000). Predictive validity of the Su tter-Eyberg Student Behavior Inventory-Revised and the Teacher's Report Form. Dissertation Abstracts International: Section B: The Sciences and Engineering, 60 (9B), 4909. Schuhmann, E. M., Foote, R. C., Eyberg, S. M., Boggs, S. R., & Algina, J. (1998). Efficacy of parent-child in teraction therapy: Interim report of a randomized trial with short-term maintenance. Journal of Clinical Child Psychology, 27 34-45.
49 Shaffer, D., Fisher, P., Lucas, C. P., Dul can, M. K., & Schwab-Stone, M. E. (2000). NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 28-38. Spiker, D., Kramer, H., Consta ntine, N., & Bryant, D. (1992). Reliability a nd validity of behavior problem checklists as measures of stable traits in low birth weight, premature preschoolers. Child Development, 63 1481-1496. Walters, A. S., & Blane, K. K. (2000). Ment al retardation. In C. H. Zeanah (Ed.), Handbook of infant mental health (2 nd ed., pp. 271-281). New York, NY: Guilford. Wechsler, D. (2002). Wechsler Preschool and Primary Scale of Intelligence-Third Edition San Antonio, TX; The Psychological Corporation. Wierzbicki, M., & Pekarik, G. (1993). A meta-analysis of psychotherapy dropout. Professional Psychology: Re search and Practice, 24 190-195.
BIOGRAPHICAL SKETCH Daniel Marc Bagner was born in East Williston, New York, on February 21 st 1979 to Jeffrey and Bernice Bagner. He has one si ster, Jessica, who is 3 years older and with whom he remains very close. Daniel gr aduated from Wheatley High School in 1997 and received his Bachelor of Arts degree in psychology and philosophy from Washington University in St. Louis in May of 2001. In July of 2001, Daniel Bagner enrolled in the doctoral program in the Department of Clini cal and Health Psychology at the University of Florida. While at the University of Florida, he has concentrated his research in the field of clinical child psychology and receiv ed a National Research Service Award to conduct his dissertation. Dani el will be attending Br own Medical School for his internship in July of 2006. His future career goals include mast ering research and clinical skills in aspiration of working in an academic setting. 50