Children and Youth Services Review 46 (2014) 238–247

Contents lists available at ScienceDirect

Children and Youth Services Review

j ourna l homepage: www.e lsev ie r .com/ locate /ch i ldyouth

Evaluation of an existing parenting class within a women's statecorrectional facility and a parenting class modeled from Parent–ChildInteraction Therapy☆

Ashley T. Scudder ⁎, Cheryl B. McNeil 1, Karishma Chengappa, Amanda H. CostelloDepartment of Psychology, West Virginia University, United States

☆ This research was supported by funding from the WCollege Academic Enrichment Fund and the DepartmentResearch Fund. The authors would like to thank theCorrections, Lakin Correctional Center, and the West Virfor their collaboration and involvement in participant eand administrative and facility support throughout the cu⁎ Corresponding author at: Department of Psychiatry,W

Clinic, University of Pittsburgh School of Medicine, 3811506, Pittsburgh, PA 15213, United States. Tel.: +1 412 57

E-mail address: scudderab@upmc.edu (A.T. Scudder).1 Is now at Kennedy Krieger Institute/Johns Hopkins

United States.

http://dx.doi.org/10.1016/j.childyouth.2014.08.0150190-7409/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 April 2013Received in revised form 13 June 2014Accepted 18 August 2014Available online 16 September 2014

Keywords:Maternal incarcerationPrisonCorrectional programmingParenting classesBehavioral parent trainingParent–Child Interaction Therapy

Approximately 336,000 households with minor children are believed to be directly affected by parental imprison-ment, and themajority of inmates in correctional facilities across the United States have been reported to be parentsof minors (Mumola, 2000). Subsequently, parent training provided through correctional programming has the po-tential to impact a large number of American children affected by parental incarceration. However, there are no em-pirically supported “best practices” for parenting programs provided during incarceration and programs often varyacross facilities. The current study examines an existing parenting class offered within a correctional facility and aparenting class modeled from Parent–Child Interaction Therapy (PCIT), a program with an existing evidence basefor improving parent and child outcomes. Eighty-two women incarcerated in a female, state correctional facilitywere randomized to one of the two parentingmodels. Both parentingmodels werematched in dose to the facility'sexisting program, ofweekly, 90-minute, group parenting classes, consisting of 10–15mothers. Seventy-onewomencompleted the parenting classes. Participants in bothmodels were combined for pre-treatment and post-treatmentassessments that included group administration of self-report measures and individual behavior observations ofparenting interaction role-plays. At post-treatment, mothers completing the PCIT-based training demonstratedhigher levels of parenting skills and reported higher levels of treatment satisfaction than mothers completing theexisting facility class. Mothers completing the existing class reported higher levels of parenting knowledge ofchild development than the PCIT-based class. Following both parenting models, similar decreases were found be-tween groups in parenting stress and child abuse potential.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Recent rates suggest that 1 in 20 individuals in the U.S. populationwill serve time in federal or state correctional systems during theirlifetime, and 1 in 198 will serve a sentence of more than one year(West, 2010). The incarcerated female population has grown substan-tially over the past 30 years (Sabol & Couture, 2008; West, 2010). Amajority of female inmates are in child-rearing age and are mothers ofyoung children (Glaze & Maruschak, 2008). At last estimate therewere 65,600 incarcerated mothers with 150,000 children below

est Virginia University Eberlyof Psychology Doctoral StudentWest Virginia Department ofginia Department of Educationnrollment, delivery of classes,rrent project.estern Psychiatric Institute and

O'Hara Street, Bellefield Towers8 9485.

University School of Medicine,

18 years of age (Glaze & Maruschak, 2008). The majority of childrenbeing younger than 10 years, and 22% of children being younger than5 years old (Glaze & Maruschak, 2008).

1.1. Impact of incarceration on mothers & children

Incarceratedmothers frequently report lower family income, paren-tal education, socioeconomic status, poorer health, and higher levels ofparental psychopathology and stress such as depression, anxiety,parenting-related stress, inappropriate and inconsistent discipline, andchild physical abuse compared to non-incarcerated mothers(e.g., Beck, 2000; Glaze & Maruschak, 2008; Green, Miranda,Daroowalia, & Siddique, 2005; Greene, Haney, & Hurtado, 2000; Houck& Loper, 2002; Kjellstrand & Eddy, 2011; Murray, Farrington, & Sekol,2012; Wright, Salisbury, & VanVoorhis, 2007). Maternal incarcerationmay be particularly disruptive as a majority of mothers (i.e., 64.3%)report living with their children and serving as a primary caregiver oftheir children prior to incarceration (Mumola, 2000). The decline inthe level of contact that mothers have with their children during incar-ceration increases the likelihood of significant family disruption andnegative parent and child outcomes (Glaze & Maruschak, 2008;

239A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

Murray et al., 2012). During incarceration mothers commonly reportdaily stressors such as issues related to their parental rights andinvolvement in the child welfare system, a lack of father support, fostercare placement, and parole requirements. Following release from cor-rectional facilities, most mothers report increased stressors anddifficulty in discipline (see Murray et al., 2012 for a detailed review).

Children of incarcerated parents are at risk for a range of adversebehavioral outcomes. During parental incarceration, children oftenexperience unstable childcare arrangements, reduced income, movesto new households and schools, and separation from siblings, peers,and other significant relationships (Murray et al., 2012). Parental incar-ceration is also associatedwith higher risk for disruptive child behaviors(Murray et al., 2012). Estimates of behavior problems among children ofincarceratedmothers followingmaternal incarceration have been foundto range from 52–70% (Block & Potthast, 1998; Woodward, 2003).

1.2. Current prison parenting programming

The U.S. spends billions of dollars annually at the state and federallevels to accommodate prison populations (Pew Center on the States,2008). A portion of this spending is used to provide programmingsuch as educational courses, mental health treatment, and recoveryservices during incarceration. Effective programs are needed and havepotential for broad impact as this is a population of high need and ofhigh levels of lifetime service use across multiple service systems(e.g., mental health, juvenile justice, child welfare, corrections, andearly education). However, there is mixed evidence for the utility ofprogramming in altering long-term outcomes such as recidivism(e.g., Davis, Bozick, Steele, Saunders, & Miles, 2013; Palmer, 1996).Although recidivism rates of future criminal behavior have beenshown to be lower for individuals completing correctional programssuch as parenting programs (Gordon & Weldon, 2003; Palmer, 1996),potential selection biases exist, as participation in correctional program-ming is often determined by a classification of the individual's level of“risk,”which is made at intake and reevaluated overtime by the facility,therefore individuals who complete correctional programming maylikely differ from those not participating in facility programs.

The range of parenting-related services provided in correctionalfacilities is broad (e.g., prison nurseries, child visitation, parenting clas-ses, and therapeutic services), however a class format has been mostcommonly used to provide parent training to incarcerated individuals(Hughes & Harrison-Thompson, 2002). In 2001, a National Institute ofJustice survey identifying innovative programs for women in correc-tional settings found 21 parent education program models across theU.S. A more recent national review of prison parenting programsyielded 28 parenting programs (Eddy et al., 2008). Although femaleand coed facilities are likely to offer parenting programs (i.e., 90.2%and 55%, respectively; Hughes & Harrison-Thompson, 2002), only 27%of incarcerated mothers report attending parenting classes (Glaze &Maruschak, 2008).

There are currently no commonly accepted “best practices” andnational standards to guide correctional facilities regarding curriculumdevelopment, participant selection, delivery practices, or assessmentof parenting programs. Most programs are developed in-house andare not established based on the recent empirical literature (Eddyet al., 2008). Furthermore, there is little systematic knowledge aboutthe impact of facility-based parenting programs on incarcerated parentsand their families. A recent review found only 17 empirical evaluationsof parenting programs involving incarceratedwomen. Of these, few hada comparison or waitlist control group, and only one study includedrandomization to treatment condition. The interventions used variedfrom general parenting discussion groups to multi-componentprograms (e.g., mothers' support group, didactic parenting class, thera-peutic visitation program). In this review, facility-based parentingprograms provided information about general communication skills(100%), parenting techniques (96%), and child development (68%).

Some also provided special parent–child visitation opportunities (57%)and emphasized parenting during incarceration (54%; Eddy et al.,2008). Anger and stress management (25%) and parent–child interac-tion laboratories (7%) were less commonly noted. Very few programsin correctional facilities currently offer a parentingmodel that combinesclassroom instruction with a behavioral component in which motherspractice specific parenting skills (e.g., Block & Potthast, 1998; Sandifer,2008). The reported duration of programs ranged from 1 to 24 weeksand varied from one class per week to full-day trainings. Completeprograms ranged from 5 to 72 h. All studies included the pre- andpost-treatment measurement of parental self-reports on constructsbelieved to affect parenting such as parenting self-esteem, stress,knowledge, or institutional adjustment.Most found at least one positivepre–post effect; however when a comparison or waitlist control groupwas included, treatment group changes were often reported to be sim-ilar to the no-treatment or wait-list control group. No studies includedstructured behavior observation or measurement of demonstratedparenting skills (see Eddy et al., 2008 for a summary of studies examin-ing parenting class outcomes). Focus groups to assess inmate interest inspecific components of parenting programs indicated the highest levelof interest in components specific to parenting while incarcerated(e.g., writing an appropriate and encouraging letter to children ofvarious ages, learning to make a phone call go well, or increasing thequality of parent–child interactions during facility visitation). Inmatesweremost interested in topics of appropriate limit setting anddisciplineas well as family and parent–child relationships during the transition tothe community following incarceration (Eddy et al., 2008).

1.3. Relevant outcomes following facility-based parenting programs

Facility-based group parenting classes have been found to enhancepre- to post-treatment differences for both child and parent stress(Loper & Tuerk, 2006) and increase parental report of non-violentapproaches to child behavior management (e.g., Sandifer, 2008;Showers, 1993). Following a 15-week parenting program, Thompsonand Harm (2000) found a decrease in inappropriate expectations of chil-dren, belief in corporal punishment, and parent–child role reversal on theAdult–Adolescent Parenting Inventory (AAPI-II; Bavolek & Keene, 1999).Surratt (2003) examined three groups of mothers in a substance abuseoffender village (i.e., thosewho had received no treatment, were current-ly in treatment, and who had completed treatment) and also found thatattendance at parenting classes predicted decreased belief in corporalpunishment. Sandifer (2008) found decreased belief in corporal punish-ment, inappropriate expectations of child development, and parent–child role reversal as well as an increased parental empathy following a12-week, group treatment with didactic and interactive components.These findings suggest that further empirical examinations with strongmethodological rigorwill be important in order to understand the impactof facility-based parenting programs on parenting outcomes.

1.4. Parent–Child Interaction Therapy (PCIT)

Parent–Child Interaction Therapy (PCIT) is a long-establishedevidence-based intervention for the treatment of externalizing behaviorproblems among children 2–7 years of age as well as the enhancementof parenting practices in families of children 4–12 years of agewhohavea history of child physical abuse. PCIT is one program developed fromHanf's two-stage model (Reitman & McMahon, 2012), which has beeninfluenced by Baumrind's (1967) theory demonstrating children'sdual needs for parental nurturance and limits to achieve optimal out-comes. PCIT consists of several core features: (a) the parent and childare actively involved together in treatment sessions, (b) interactionsare coded to assess progress and determine treatment planning,(c) parents are coached to assist in reaching a level of mastery of bothplay-therapy and discipline skills, (d) traditional play-therapy skillsare taught to enhance the quality of the parent–child relationship,

240 A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

(e) parental problem-solving skills and use of behavioral principles arecoached to develop direct strategies for management of problembehaviors, (e) PCIT is clinically validated, and (f) changes are madebased on empirical evidence (Eyberg, 2005). Throughout PCIT, parentswork tomaster specific skills which have been demonstrated to unique-ly impact child behavior (e.g., Hart & Risley, 1995; Tempel, Wagner, &McNeil, 2013).

The positive treatment effects of PCIT are demonstrated acrossvarying severity levels of child disruptive behavior (Funderburk et al.,1998). PCIT has also been shown to be effectivewith families presentingwith mild parent psychopathology such as mild depression or anxiety(Eyberg & Robinson, 1982). Treatment effects are also found to general-ize from the clinic to the home (Boggs, 1990), school setting (McNeil,Eyberg, Eisenstadt, Newcomb, & Funderburk, 1991), and to parent–child interactions with untreated siblings (Brestan, Eyberg, Boggs, &Algina, 1997; Eyberg & Robinson, 1982). Efficacy studies demonstrateimportant changes in parents' behaviors such as increased reflectivelistening, physical proximity, and positive verbal reinforcement,as well as decreased parenting stress, use of corporal punishment,and use of high rates of parental commands, criticism, and sarcasm(Chaffin et al., 2004; Schuhmann, Foote, Eyberg, Boggs, & Algina,1998). The efficacy of PCIT is seen in treatment with diverse co-occurring child concerns such as Attention-Deficit/Hyperactivity Disor-der (Nixon, 2001), internalizing disorders (Pincus, Eyberg, & Choate,2005), chronic illness (Bagner, Fernandez, & Eyberg, 2004), develop-mental disabilities (McDiarmid & Bagner, 2005), and general child mal-treatment (Chaffin et al., 2004; Fricker-Elhai, Ruggiero, & Smith, 2005).Furthermore, PCIT treatment gains have been shown to maintain up to6 years following treatment (Hood& Eyberg, 2003). In addition, parentsreport high levels of satisfaction concerning the content and course ofPCIT treatment (e.g., Chaffin et al., 2004; McNeil, Herschell, Gurwitch,& Clemens-Mowrer, 2005; Schuhmann et al., 1998; Timmer, Urquiza,Zebell, & McGrath, 2005), which is thought to be a result of decreasesin parenting-related distress as well as increases in parents' levelsof confidence in their ability to manage child problem behaviors(e.g., Timmer et al., 2005).

PCIT is considered a “best practice” for the treatment of child physi-cal abuse by the Kauffman Foundation (2004) and the National ChildTraumatic Stress Network, and has also been indicated as a “supportedand acceptable” treatment by the National Crime Victims Researchand Treatment Center and the Center for Sexual Assault and TraumaticStress. PCIT is particularly appropriate to address abuse occurring inthe context of child discipline (e.g., Borrego, Urquiza, Rasmussen, &Zebell, 1999; Chaffin et al., 2004). It is common for parents with a histo-ry of child physical abuse to have too high or too low expectations ofchild behavior. In PCIT with child physical abuse populations, coachingtargets the reduction of negative parental behaviors during parent–child interactions and enhancing the appropriateness of parents'expectations of the child based on the child's developmental level.

1.5. PCIT-based group formats

Several PCIT group formats have been developed and studied. Initialfindings suggest that group formats are effective in altering behavioraloutcomes across settings (e.g., Berkovits, O'Brien, Carter, & Eyberg,2010; McNeil et al., 2005; Niec, Hemme, Yopp, & Brestan, 2005).Existing group formats vary in the number of recommended groupparticipants, length and number of sessions, mastery and graduationrequirements, and modes of coaching (e.g., in room versus observationroom). Group PCIT maintains in-session parent practice during both arelationship enhancement phase known as Child-Directed Interaction(CDI) and a behavior management phase known as Parent-DirectedInteraction (PDI). Eyberg et al. (2009) developed a group PCIT manualin which treatment progression is based on 50% of group participantsmeeting mastery skills prior to completion of CDI and PDI treatmentstages. Preliminary observational data of pre-to post-treatment

parent–child interactions showed large effects for increased parentpraise (d = 2.73, N = 11) and reduced parental criticism (d = 1.90,N=11). Furthermore, parent-report measures also reflected treatmentacceptability in group PCIT as mothers and fathers indicated high levelsof treatment satisfaction (M = 44 on a 50-point scale, N = 20).

Berkovits et al. (2010) conducted a comparison of two preventativePCIT-based interventions (i.e., Primary Care PCIT and PCIT AnticipatoryGuidance), involving families of children with subclinical externalizingbehavior scores. In Primary Care PCIT, 17 families were coached andcoded during parent–child interactions during four weekly classesheld in the waiting room of a primary care physician's office. In PCITAnticipatory Guidance, families participated in self-guided treatmentusing PCIT resources (i.e., written descriptions of CDI and PDI skills,practice suggestions for using the skills in the same manner as recom-mended in Primary Care PCIT, and additional parenting “tip sheets”given to parents in the Primary Care PCIT groups). Parent-reportedparental locus of control increased, while parenting stress, and childexternalizing behavior problems decreased at post-treatment. Further-more, targeted treatment gains and satisfaction with the programwere maintained at the 6-month follow-up for parents in both adaptedmodels. McNeil et al. (2005) used a two-day workshop format toenhance parenting skills among foster parents. Parents did not haveat-home practice of skills with their children during theworkshop peri-od, yet reported clinically and statistically significant decreases in childbehavior intensity and problem scores on the Eyberg Child BehaviorInventory (Eyberg & Pincus, 1999) at the one-month post-workshopfollow-up. Additionally, Lee, Wilsie, and Brestan-Knight (2011) devel-oped a pre-parent education model for undergraduate students whichwas based on PCIT skills and found increases in behavior managementknowledge and demonstrated use of praise during role-plays of child-led play interactions. These findings suggest that alternative PCIT-based treatment models may be feasible, effective, and acceptableprevention and early intervention approaches.

The current study investigates whether a PCIT-based parenting classprovides additional benefit when compared to the existing parentingclasses in a women's correctional facility in enhancing demonstratedparenting skills as well as parent-reported knowledge of child develop-ment, parenting stress, child abuse potential, and treatment acceptabil-ity. This study aims to explore the effectiveness of two facility-basedgroup parenting models in enhancing parent-reported and observedparenting outcomes. Additionally, the study provides further under-standing of the feasibility and acceptability of a PCIT-based parentingclass, while adhering to restrictions characteristic of a correctionalfacility.

2. Method

2.1. Participants

All womenwere currently incarcerated at a state correctional facilityhousing approximately 500women at all levels of security classification.The facility was located in the Appalachian region of the United Statesand was the only all-female facility in the state. Participants wereselected from a list of womenwaiting to participate in parenting classes.Womenwere eligible for participation if theywere (a) eighteen years orolder, (b) recommended for parenting programming through the stan-dard Individualized Recovery/Resiliency Plan process conducted by theDepartment of Corrections staff at facility intake, (c) the mother of achild between the ages of 2 to 12 years, (d) not parole eligible beforestudy completion, and (e) had available time during the scheduledparenting classes. Study criteria were also established that individualswould be excluded from sample if they (f) were not fluent Englishspeakers or (g) had participated in the existing parenting program atthe facility within the past year. No women met the exclusion criteria.Eighty-four women were selected for the current study. Two womendid not complete the initial pre-treatment assessment. These women

241A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

were placed back on the parenting waitlist to receive the program at alater date. Of the 82 participants enrolled, 71 completed pre- andpost-treatment assessments and are included in the current study.Data provided by participants at pre-treatment revealed no significantdifferences between study completers and non-completers for parentage, race/ethnicity, or education level. Participants were reported tonot complete the study for three primary reasons: scheduling conflictswith other facility programming such as another class or job,movementto another facility, or parole.

2.2. Procedure

The current study protocol was approved by the West VirginiaUniversity Institutional ReviewBoard. Prior to the pre-treatment assess-ment, consent forms were numbered with a participant identificationnumber. Participants self-selected participant identification numbersbased on their chosen seat at the initial pre-treatment informed consentprocess. Participant identification numbers were randomly assigned toone of two treatment conditions, (a) the PCIT-based class or (b) theexisting facility parenting class; this process was completed using anonline randomizer and coordination with a clerical staff memberuninvolved with the collection of the data or the parenting classes. Allparticipants were informed that they were able to withdraw from theresearch project at any time without penalty.

Participants in both treatment groups were combined for studyassessments oneweek prior to the start of parenting classes and follow-ing the completion of the classes. Assessments at each time pointincluded group administration of self-report measures and individualbehavior observations of two structured parenting role-play interac-tions (i.e., child-directed play and parent-directed play). Research assis-tants were available to assist with individual questions or clarificationrelated to assessment measures. Participants were asked to answeritems related to child behavior based on “one of their own childrenwho was 12 years or younger and demonstrated the most behavioraldifficulties.” This age range was selected for several reasons: 1) concep-tually, PCIT has been established as an EBT for children with disruptivebehavior difficulties ages 2–7 years and families who have a history ofchild physical abuse with children ages 4–12 years old, and 2) theexisting facility structure did not provide parents with a unique curric-ulum based on child age, therefore due to the preliminary nature ofthe study, it was thought to be most feasible and acceptable to divideclasses formothers of children 2–12 and 13–18 years. During individualbehavior observations, a research assistant played the role of a childbetween the ages of 3–6 years following a semi-structured child behav-ior script. The DPICS-III has been extensively studied with families ofchildren ages 3–6 years and has been normed for parent–child interac-tions with children in this age range. The average age of all mothers'children (i.e., not only the reported target child within the specifiedrange) was 6.76 years. Mothers were asked to select their “mostdifficult child within the range”, as mothers often had more than onechild within the selected age range.

During the 5-minute interaction, participants were asked to “inter-act as they typically would with their own child or a child of this age.”All research assistants involved in the coding of the behavioral observa-tions did not attend class sessions and remained uninformed of partici-pants' treatment conditions throughout the course of the study. Bothgroups received a parenting class certificate for participation and noincentives were provided unique to study participation (i.e., researchparticipation was not documented in inmate facility profile to ensurethat participation did not influence sentencing, tangible rewards suchas food or money were not provided for participation). To receive aclass participation certificate, participants were required to attend andparticipate in a majority of classes, which was an existing facility policy.All participants enrolled in the study met the requirements of thisattendance policy.

2.3. Measures

2.3.1. Participant information formA questionnaire was designed for description of parent-reported

demographic (e.g., parent age, education, child age) and parent–childcontact information (e.g., type of contact, frequency of contact). Thisformwas developed to include demographic and contact characteristicspreviously examined in the parenting programs literature (e.g., Tuerk &Loper, 2006).

2.3.2. Participant file reviewA file review using available participant facility files was used to

collect basic demographic information that was not included in the par-ticipant information form. This information included participant intel-lectual and achievement test scores, behavioral characteristics duringincarceration, and facility risk and mental health classifications.

2.3.3. Dyadic Parent–Child Interaction Coding System-III (DPICS-III)The Dyadic Parent–Child Interaction Coding System-III (Eyberg,

Nelson, Duke, & Boggs, 2005) is a behavioral coding system, which hasbeen designed and psychometrically assessed for the measurement ofinteraction quality within play interactions of parents and children3–6 years of age. The DPICS-III consists of behavioral categories tomeasure all parent and child verbalizations, vocalizations (i.e., yell,whine, and laugh), physical behaviors (i.e., positive and negativetouch), and responses (e.g., compliance to command). The psychomet-ric properties for the DPICS have been studied extensively (see Eyberget al., 2005 for an overview). Reliability and validity studies of theDPICS during live coding situations have demonstrated adequate results(for a table of kappas for individual parent and child categories duringlive coding see Eyberg et al., 2005).

2.3.4. Adult–Adolescent Parenting Inventory—Second Edition (AAPI-II)The AAPI-II (Bavolek & Keene, 1999) is a 40-item assessment of

parenting and child-rearing attitudes of adult and adolescent parentand pre-parent populations. The AAPI-II uses a continuous scale to clas-sify individuals' parenting-related knowledge to provide an index of riskfor practicing behaviors known to be attributable to child abuse andneglect. Participants respond to statements using a 5-point Likert-typescale. The Inappropriate Expectations of Children scale consists ofseven items related to parents' understandings of child growth anddevelopment, which specifically examines the degree towhich parentalexpectations exceed developmental capabilities of children. Low scoressuggest demandingness and limited emotional support of their chil-dren. The Parental Lack of Empathy toward Children's Needs scale ismade up of 10 items that examine parental understanding and valueof children's needs. Low scores suggest limited general empathy andvalue found in the parental role. Low scores also suggest that parentingdemands might lessen the worth of having children (Bavolek & Keene,1999). The Strong Parental Belief in the Use of Corporal Punishmentscale consists of 11 items that assess attitudes toward various methodsof discipline. Low scores indicate preference of corporal punishmentstrategies such as spanking children. All subscales demonstrate highinternal reliability (Cronbach's alpha = .89–.96).

2.3.5. Parenting Stress Index, Third Edition (PSI)The PSI (Abidin, 1995) is a 120-item assessment of stressors partic-

ular to parenting. This measure yields data on potential sources of stressrelated to family through the parent's perception of his or her function-ing in the parental role, issues of the parent–child relationship, childdevelopment issues, life stress, and degree of support from the otherparent. The PSI yields 17 scores including seven Child Domain scores,eight Parent Domain scores, and a Total Stress score, plus an optionalLife Stress score. Participants respond to statements using a 5-pointLikert scale ranging from 1 (Strongly agree) to 5 (Strongly disagree).Scores on the PSI remain stable over time, as indicated by correlation

242 A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

coefficients of .91 across a 1- to 3-month interval (Abidin) to .70 over a1-year interval (Hamilton, 1980 as cited in Abidin).

2.3.6. Child Abuse Potential Inventory (CAP)The CAP (Milner, 1986) is a 160-item assessment developed to

estimate risk for committing child physical abuse, which includes anabuse potential scale and subscales. The Abuse scale is a 77-item mea-sure that combines six subscales: rigidity; distress; unhappiness; andproblemswith child, family, and others. Items are rated as ‘agree’ or ‘dis-agree.’ The Abuse scale has been reported to have high internal consis-tency (KR-20 = .92 to .95), test–retest reliability (.83 at 1 month,.75 at 3 months), and good discriminant and predictive validitybetween abusive and non-abusive adults (i.e., Milner, 1986). Normativevalue for theAbuse scale is 91with a signal-detection cutoff score of 166(Milner, 1986).

2.3.7. Therapy Attitude Inventory (TAI)The TAI (Eyberg & Johnson, 1974) consists of 10 items, on a 5-point

Likert-type scale, designed to assess parental satisfaction with therapyfollowing the completion of parent training (i.e., total score rangesfrom 10 to 50). Higher total scores indicate higher levels of satisfactionregarding treatment and improvement of previously identified childbehavior problems. Specific items require parents to rate variouscomponents of the treatment including confidence in implementing in-tervention components and general impressions of treatment. The TAIhas demonstrated strong internal consistency (Cronbach's alpha =.91) and test–retest reliability (r = .85; Brestan, Jacobs, Rayfield, &Eyberg, 1999).

2.4. Interventions

2.4.1. PCIT-based classThe PCIT-based class was adapted to use role-plays and in-room

coaching of parenting skills. To match the existing facility class andthe standard practice of correctional parenting classes, the PCIT-basedclass was conducted using a group format and was limited to sevenclasses. The PCIT-based program incorporated the group PCIT protocol(Eyberg et al., 2009) and active behavioral rehearsal (i.e., role-play)strategies of coding and coaching similar to practices commonly usedin PCIT-related workshops or trainings of community-based mentalhealth clinicians (Scudder & Herschell, in press). Behavioral rehearsal(i.e., a trainee practices specific behaviors during a simulated interactionwith another individual) is considered an effective and potentially cost-effective and feasible solution of knowledge transfer for the training ofvarious behavioral skills (e.g., Beidas, Cross, & Dorsey, 2014; Crosset al., 2011). Each PCIT-based class had 12–15 participants assignedand was led by an instructor with a Masters degree in Psychology andan undergraduate student assistant. Classes were conducted in 90-minute time periods. The first 10 min of each class was used to takeattendance, discuss questions related to previously discussed materialor related to out-of-class practice, review participants' skills summarysheets, and provide an overview of the day's agenda to the class. Thenext 70 min of each class consisted of in-class DPICS-III coding andcoaching of participants in the use of the PCIT mastery skills duringparenting role-plays. The final 10 min of each class was used to discussone of the optional topics from the standard PCIT protocol (e.g., Parentsare Models, Finding Supports) and to assign daily out-of-class practice.During in-class coaching and coding practice, participants were dividedinto small groups (e.g., four groups of three participants, three groups offour participants). They also participated in whole group practice, inwhich most group participants coded as one class participant role-played a child and one role-played the parent while the instructorcoached. Within each group, participants rotated between the role ofparent, child, or coder. The instructor rotated throughout the groupsto provide group participants an opportunity for in-class coding andcoaching of skills and feedback. Each participant also engaged in group

coaching activities focused on skills needed to progress toward CDIskill mastery. Each participant was coached by the class instructor andparticipated in active coding or coaching activities during the remainingcoding-coaching period (i.e., when not being actively coached by theclass instructor). Participants were also coded each class by a researchassistant for 2.5–5 min using the DPICS-III to assess and receive addi-tional feedback on their progress toward mastery of the PCIT masteryskills. Participants were provided with coding sheets and were taughtto code other class participants' use of the PCIT mastery skills as theyparticipated in the coding-coaching practice. Each class concludedwith a review of participants' charted skills progress followed by a sum-mary of the day's class. As a component of the program, participantswere asked to practice the PCIT skills that they learned in-class, for5–10 min a day, with another group participant or with their childduring phone contact or facility visitation. Participants had access totoys appropriate for PCIT (e.g., Legos, dollhouse, farm set with animals)in the facility visitation room used for family visits. Colored pencils andblank paper were also provided to participants. Participants wereencouraged to practice their PCITmastery skills on a daily basis, monitortheir practice using a PCIT homework sheet, and return their homeworksheet to class each week (see Eyberg et al., 2009 to review referencedmaterials).

Several adaptations were used to accommodate the demands andlimitations of the correctional environment (see Eyberg, 2005 for guide-lines related to tailoring, adapting, or modifying treatment). First,although mastery of the CDI and PDI mastery skills remained empha-sized throughout the PCIT-based classes, it was not required thatgroup participants met mastery criteria prior to progression to PDIor before graduation of the program. Participants received a pre-determined number and duration of classes, and all participantsprogressed in treatment at the same time, regardless of skills mastery(e.g., Niec et al., 2005). Second,mothers received the interventionwith-out their children, and the majority (73.9%) did not have visits withtheir children throughout the time period that the parenting classeswere held. Third, out-of-class practice activities consisted largely ofbehavioral rehearsal role-play practice of parenting skills while engag-ing in a coloring activity with a selected class partner. Out-of-class prac-tice may also have been completed in the form of over-the-phone orletter writing use of PCIT mastery skills with their child, as manymothers wished to practice their skills with their own children andoften did not have direct visitation with their children throughout thecourse of the study. Mothers practiced reflecting their children and pro-viding labeled praises, while also reducing commands and negative talkduring phone and written contact. Mothers were less able to practiceproviding behavioral descriptions or reducing questions throughphone or written correspondence.

2.4.2. Existing facility parenting classThe existing facility parenting program was adapted by the facility

from the Partnerships in Parenting manual (Bartholomew, Knight,Chatham, & Simpson, 2002), which was designed for eight, weekly,120-minute meetings. Each class had 12–15 participants assigned andwas led by an instructor with a Masters degree and an inmate assistantwho had previously completed the class. Classes were divided into sev-eral topic categories (e.g., self-esteem, communication, discipline versuspunishment, family origins, family rules and responsibilities, domesticviolence and children, child abuse, and going home). Role-play was anoptional piece of the curriculum to illustrate specific concepts. Classparticipants wrote a three-page homework assignment about “How toDeal with Forgiveness” and “Who am I?” Class discussionswere supple-mented with videos and instructor-led presentations. The existing facil-ity program was adapted by the Department of Education and thefacility programs staff to meet the needs of mothers and the facilitybased on prior experience and use of the curriculum. During the first20 min of each class, participants completed introductory activitiesand discussed previous class topics. In the next 60 min, the instructor

243A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

led participants in a discussion regarding the daily topic in which vari-ous methods to apply each topic to participants' lives were discussed.In the last 10 min of each class, the instructor led a wrap-up.

3. Pre-randomization data

Participant demographic and child contact data were assessed (seeTable 1). Participants had a mean age of 30.83 years (SD = 5.88).Approximately 93% were reported to be white, non-Hispanic, 3.7%were African American, and 3.7% were biracial, which was reflective ofthe state's overall female population (i.e., 94% white, non-Hispanic;U.S. Census Bureau, 2014).Mostmotherswere single and nevermarried(36.6%) or were divorced or separated (36.6%). Themajority of mothers(63.4%) had not graduated from high school and were unemployed orreceiving disability benefits at the timeof incarceration (56.1%). Accord-ing to the U.S. Census Bureau's poverty line criteria for a family of five in2011 (i.e., $26, 434; U.S. Census Bureau, 2014), 72.9% of participantswere identified living in households below the poverty line.

Participants had an average of 2.30 children (SD = 1.31), and theaverage child age of all participants' children (i.e., not only the reportedtarget children) was 6.76 years (SD = 4.37). The majority of partici-pants reported being responsible for the care of at least one of theirchildren (71.0%) and seeing their children every day prior to incarcera-tion (71.0%). Prior to their parenting class, mothers reported that theytalked about their child every day (64.6%), wrote letters to their childat least once a week (57.3%), and talked with their child on the phone

Table 1Demographic information as totals and percentages across group and time.

Group Overalln (%)

PCITn (%)

TAUn (%)

Total participants 82 (100%)Pretreatment 82 (100%) 42 (51.2%) 40 (48.7%)Post-treatment 69 (84.1%) 39 (51.3%) 30 (49.1%)

EthnicityWhite, non-Hispanic 76 (92.6%) 41 (97.6%) 34 (85.0%)African-American 3 (3.7%) 1 (2.4%) 2 (5.0%)Biracial 3 (3.7%) 0 (0%) 3 (7.5%)

Marital statusSingle, never married 30 (36.6%) 15 (59.5%) 15 (37.5%)Divorced/separated 30 (36.6%) 17 (40.5%) 13 (32.5%)Married 9 (11.0%) 5 (11.9%) 4 (10.0%)Unmarried partners 7 (8.5%) 4 (9.5%) 3 (7.14%)Widowed 4 (4.9%) 0 (0%) 4 (10.0%)Unknown 2 (2.4%) 1 (2.4%) 1 (2.5%)

Education1st–8th grade 4 (4.9%) 2 (4.8%) 2 (5.0%)9th–11th grades 48 (58.5%) 28 (66.7%) 20 (50.0%)12th grade 19 (23.2%) 9 (21.4%) 10 (25.0%)Vocational school or some college 8 (9.8%) 1 (2.4%) 7 (17.5%)Unreported 3 (3.7%) 2 (4.8%) 1 (2.5%)

Employment statusUnemployed/disability 46 (56.1%) 27 (64.3%) 19 (47.5%)Employed full time 19 (23.2%) 5 (11.9%) 14 (35.0%)Employed full time 10 (12.2%) 8 (19.0%) 2 (5.0%)Unknown 7 (8.5%) 2 (4.8%) 5 (12.5%)

Annual household income at incarcerationUnknown 37 (45.1%) 21 (50.0%) 16 (40.0%)Under $15,000 21 (25.6%) 11 (26.2%) 10 (25.0%)$15,001–$25,000 10 (12.2%) 5 (11.9%) 5 (12.5%)$25,001–$40,000 9 (11.0%) 5 (11.9%) 4 (10.0%)Over $40,000 5 (6.1%) 0 (0%) 5 (12.5%)

Total childrenOne–two 56 (68.3%) 29 (69.0%) 27 (67.5%)Three–four 20 (24.4%) 10 (23.8%) 10 (25.0%)Five or more 6 (7.3%) 3 (7.1%) 3 (7.5%)

Child now living withGrandmother 46 (47.9%) 25 (50.0%) 21 (45.7%)Father 25 (26.0%) 13 (26.0%) 12 (26.1%)Another family member 10 (10.4%) 2 (4.0%) 8 (17.4%)Non-family member, chosen by courts 8 (8.3%) 6 (12.0%) 2 (4.3%)Other 7 (7.3%) 4 (8.0%) 3 (6.5%)

at least once a week (48.8%). Few (13.2%) reported ever receiving avisit at the facility from their children. The primary caregivers ofmothers' children during maternal incarceration were most commonlygrandmothers (47.9%); however children were also cared for by fathers(26.0%), other family members (10.4%), non-family member chosen bythe courts (8.3%), and other arrangements (7.3%). Approximately 16.0%of mothers indicated that the primary caregiver of their children hadchanged more than once during their incarceration.

4. Participant file review

Following study completion, afile reviewwas completed to examineadditional demographic variables. File review informationwas collectedby an assigned facility staff member with facility permission for specificfile access. Fifty-six files (68.3%) were available for review. Files unableto be reviewed were primarily reported to be due to participantmovement to new facilities or parole. Participants' scores on the BetaIII, a brief test of non-verbal cognitive abilities, suggestedmean intellec-tual functioning in the average range (i.e., M = 96.33, SD = 11.65).Examination of achievement scores based on the Test of AchievementBasic Education suggests grade equivalent functioning between 7thand 9th grade levels in mathematics (M = 7.95, SD = 3.29), reading(M = 9.15, SD= 3.81), and language skills (M = 8.86, SD = 3.53). Atfacility entry, most participants had a facility risk categorization ofcommunity or low risk (55.36%) and a mental health score of low risk(53.57%). Prior to participation in the current parentingprogram, partic-ipants had completed three facility programs (SD = 2.68). Conductcharacteristics such as the number of prior offenses (M = 10.39,SD = 11.87), days previously incarcerated (M = 102.70, SD =391.54), and days previously spent in segregation due to facility viola-tions (M = 46.94, SD= 131.06) varied greatly among participants.

Testing the adequacy of the randomization procedures, pre-treat-ment between-groupdifferenceswere examined for participant charac-teristics (i.e., annual household income, frequency of contact with childprior to incarceration, frequency of phone contact in past six weeks,frequency of visitation contact in past six weeks, mother educationlevel, and mother age). The selected participant characteristics andpre-treatment scores on all outcome variables were also examinedbetween participants that completed treatment and those who didnot complete treatment; participants who completed treatment wereolder (M = 31.31, SD= 4.69) than those not completing (M = 27.73,SD = 4.69) their respective parenting program, t (15.49) = −2.27,p=.04. Initial independent sample t-tests comparing pre-treatment dif-ferences between mothers in the PCIT-based class and existing facilityclass indicated no significant differences across relevant outcomevariables.

5. Results

5.1. Parenting skill between-group change from pre-test to post-test

Mixed between-within subjects ANOVAs were conducted tocompare the effectiveness of the two interventions in increasing dem-onstrated parenting skills in mothers across time (see Tables 2 & 3). Inchild-led play role-play situations following the parenting classes,mothers in the PCIT-based class demonstrated significantly higherlevels of positive attention (i.e., Labeled Praises, Reflections, BehavioralDescriptions) than mothers in the existing facility parenting class, F (1,58)= 38.96, p= .00, d= 1.67. The PCIT-based class also demonstratedsignificantly lower levels of negative attention (i.e., Questions, Criti-cisms, and Commands) than mothers in the existing facility parentingclass, F (1, 58) = 17.02, p = .00, d = .83. In parent-led play role-playsituations following the parenting classes, mothers in the PCIT-basedclass demonstrated a significantly higher percentage of effective com-mands than mothers in the facility-based class, F (1, 57) = 3.92, p =.00, d= .54.Mothers in the PCIT-based class also demonstrated a higher

Table 2Means and standard deviations in parent skill, knowledge, abuse potential, stress, and satisfaction between treatments.

PCIT-based TAU

Pre Post Pre Post

Measure M (SD) M (SD) M (SD) M (SD)

DPICSPositive attention “do” skills .88 (.84) 12.38 (9.16) .62 (.75) .81 (1.36)Negative attention “don't” skills 16.21 (8.97) 9.91 (8.41) 13.31 (5.91) 19.12(14.16)Effective command sequences .00 (.00) .09 (.19) .00 (.00) .01 (.07)Compliance-contingent praise .01 (.04) .28 (.37) .02 (.04) .00 (.00)

AAPI-IIInappropriate expectations 20.63 (4.16) 20.95 (3.07) 20.55 (2.99) 22.74 (3.88)Empathetic regard 40.03 (4.49) 44.00 (3.84) 40.61 (3.88) 43.94 (5.37)Discipline 41.75 (6.43) 43.94 (5.37) 43.88 (8.82) 42.81 (8.15)

PSITotal parent stressa 229.50 (28.76) 219.50 (32.09) 222.61 (36.27) 216.16 (47.60)

CAPTotal abuse potentiala 146.08 (87.59) 119.35 (79.19) 140.16 (94.80) 130.29 (93.41)

Note.a Significant main effect for time.

244 A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

percentage of positive attending (i.e., Labeled Praises) following childcompliance compared to mothers in the existing facility parentingclass, F (1, 57) = 14.70, p = .00, d = 1.02.

5.2. Parenting knowledge between-group change from pre-test to post-test

The AAPI-II Inappropriate Expectations subscale was used to exam-ine knowledge of child development. A mixed between-within subjectsANOVA was conducted to compare the effectiveness of the two inter-ventions in increasing child development knowledge (see Tables 2 &3). Following the parenting classes, mothers in the existing facilityclass demonstrated higher levels of knowledge of child developmentthanmothers in the PCIT-based class, F (1, 69)= 7.54, p= .05, d= .51.

5.3. Parenting stress between-group change from pre-test to post-test

A mixed between-within subjects ANOVA was conducted tocompare the effectiveness of the two interventions in reducingparticipants' total stress (see Tables 2 & 3). Following the parenting clas-ses, mothers in both classes demonstrated significant decreases in totalstress. Mothers in the PCIT-based class and the existing facility class didnot differ, F (1, 69) = .179, p = .67, d = .08.

Table 3Contrast in parent skill, knowledge, abuse potential, stress, and satisfaction betweentreatments.

Measure df F p d

DPICSPositive attention “do” skills 1, 58 38.96 .00⁎ 1.67Negative attention “don't” skills 1, 58 17.02 .00⁎ .83Effective command sequences 1, 57 3.92 .05⁎ .54Compliance-contingent praise 1, 57 14.70 .00⁎ 1.02

AAPI-IIInappropriate expectations 1, 69 7.54 .05⁎ .51Empathetic regard 1, 69 .311 .58 .01Discipline 1, 69 2.33 .13 .13

PSITotal parent stress 1, 69 .179 .67 .08

CAPTotal abuse potentiala 1, 69 .923 .34 .13

TAITreatment satisfaction 1, 69 2.07 .04⁎ .50

Note.a Significant main effect for time.⁎ p b 05.⁎⁎ p b .01.

5.4. Child abuse potential between-group change from pre-test to post-test

The CAP Abuse score, AAPI-II Lack of Empathetic Regard subscaleand the AAPI-II Belief in Corporal Punishment subscale were used toexamine abuse potential. The CAP Abuse score was examined withand without elevated lie scores included. Findings were not alteredwith the removal of the elevated lie scores; therefore data from partici-pants with elevated lie scores were included. Mixed between-withinsubjects ANOVAs were conducted to compare the effectiveness of thetwo interventions in reducing mothers' potential for abuse. Followingthe parenting classes, both groups reported significant decreases inabuse potential (see Tables 2 & 3). Mothers in the PCIT-based classand the existing facility class did not differ, F (1, 69) = .923, p = .34,d = .13. Related to the AAPI-II, no significant changes were found ineither group of mothers in lack of empathetic regard, F (1, 69) = .311,p = .58, d = .01 or belief in corporal punishment, F (1, 69) = 2.33,p = .13, d = .13.

5.5. Treatment acceptability between-group differences at post-test

A mixed between-within subjects ANOVA was conducted tocompare treatment acceptability following the two interventions (seeTables 2 &3). At post-treatment,mothers in the PCIT-based class report-ed significantly higher levels of treatment acceptability thanmothers inthe existing facility class, F (1, 69) = 2.07, p = .04, d = .50.

6. Discussion

Parenting classes provided during maternal incarceration toenhance parenting skills and attitudes have the potential for broad pub-lic health impact. This study examines the feasibility, effectiveness, andacceptability of two parenting classes conducted through correctionalprogramming, which were matched to the existing parenting programat a state correctional facility. The 10.5 h offered in both brief, groupmodels provided a much lower dose of intervention than evidence-based parent-training programs offered in outpatient mental healthsettings. Themedium to large effects found for demonstrated parentingskills and intervention acceptability using behavioral rehearsal duringrole-play are similar to parent outcomes of prior prevention and earlyintervention studies examining brief formats of PCIT (e.g., Berkovitset al., 2010; Boggs et al., 2004;McNeil et al., 2005). However, further ex-amination will be needed to understand whether mothers' use of par-enting skills will generalize to parent–child interactions with theirchildren.

245A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

To adjust to the correctional environment and to match the facility-based program, the PCIT-based classwas substantially adapted from thestandard PCIT protocol. Particularly, the PCIT-based class did notprovide mothers with an opportunity to practice new skills with theirchildren during parenting sessions which is a parent- training compo-nent that is consistently associated with larger effects in parentbehavior outcomes (Kaminski, Valle, Filene, & Boyle, 2008) and is con-sidered a critical component of PCIT (Eyberg, 2005). The PCIT-basedclass structure emphasized behavioral rehearsal and practice of rela-tionship building communication skills aswell as consistent responding(i.e., reflective listening, attending to appropriate behavior), which arealso components of parent training associated with large effects inparent outcomes (Kaminski et al., 2008). The existing facility parentingprogram taught communication skills and responding during didacticdiscussions, yet also included components consistently associatedwith smaller effects such as teaching parents problem solving andteaching parents to promote children's cognitive, academic, or socialskills (Kaminski et al., 2008).

This study appears to be the first to examine structured observationsof demonstrated parenting skills through behavioral rehearsal duringmaternal incarceration. Mothers enrolled in the PCIT-based classdemonstratedmore positive parenting skills and less negative attentionduring 5-minute child-led play role-play interactions compared tomothers completing the existing parenting class. Although mothers inthe PCIT-based class demonstrated enhanced skills, the positive attend-ing skillswere at a level (M=12) that is less than half the level requiredfor standard PCIT (i.e., 30) and negative attending (i.e., 16 at pre-treat-ment to 9 at post-treatment) was three times the target level duringstandard PCIT. During parent-led role-play interactions at post-treat-ment, participants in the PCIT-based class demonstrated 10% effectivecommands and 30% appropriate responding to “child” compliance.This rate was much lower than the 75% effectively stated commandsand responding required during parent-directed interactions in stan-dard PCIT services. As noted by Cohen (1988), interpretation of d indi-cates .21 = small effect size, .56 = medium effect size, and .8 = largeeffect size. Overall, these increases in parent skills represent mediumto large effects for enhancing mothers' demonstrated parenting behav-iors. Mothers in the existing facility parenting class did not demonstratechange in parenting skills following class completion; which is consis-tent with other studies that suggest minimal skill change following di-dactic parenting interventions (e.g., Kaminski et al., 2008).

Mothers' inappropriate expectations of child development wereshown to decrease more following the existing facility class than thePCIT-based class. These findings are consistent with prior studies offacility-based parenting programs which report significant pre–postchange in child development knowledge based on the AAPI-II Inappro-priate Expectations subscale (e.g., Sandifer, 2008). Teaching parentsabout child development is not a component of parent-trainingprograms associated with program effect sizes for parent behavior orchild externalizing behaviors butmay sometimes be included in compo-nents such as teaching parents to interact positively and respondeffectively to their children (Kaminski et al., 2008).

Similar decreases in parent stress were found between mothers inboth programs. Previous PCIT outcome studies (e.g., Schuhmann et al.,1998; Timmer et al., 2006) and facility-based parenting studies(e.g., Houck & Loper, 2002) have demonstrated decreases in parentstress from pre-treatment to post-treatment that were comparable tothe current study findings. Mothers in this study reported total parent-ing stress to be at a lower level than is reported in some prior PCITstudies (e.g., Capage, Bennett, & McNeil, 2001; Schuhmann et al.,1998). As suggested by Houck and Loper (2002), lower total stressscores for incarcerated mothers on the PSI may be a product of parents'difficulty in responding to some child behavior itemsdue to limited con-tact during incarceration. Similarly, participants in the current studymay have had less opportunity to address the stressors that arecaptured in the PSI compared to parents in community samples due to

the smaller dose of intervention and limited contact with their children.Smaller reductions in parenting stress scores compared to prior exami-nations of standard PCIT may also suggest that simply having parentingstrategies may not have as large an impact on mothers' parenting-related stress as interventions which actively involve the mothers'children.

Additionally, mothers in the PCIT-based class reported similardecreases in risk for child abuse potential as mothers in the existingfacility parenting class at post-treatment. Average CAP abuse scoresfor both classes fell above the normative mean of 91, yet below thesignal-detection cut off of 166. This is consistent with previous researchinvolving familieswith a history of child abusewhich suggests that fam-ilies who receive standard PCIT treatment report pre–post decreases inpotential for abuse (Borrego et al., 1999; Timmer et al., 2005, 2006).Strong test–retest reliability on the CAP suggests that similarbetween-group decreases following treatment are not likely accountedfor by testing effects. The current study did not find changes for eithergroup on the AAPI-II Lack of Empathetic Regard or Belief in CorporalPunishment subscales, yet other studies examining facility-based par-enting programs have found pre–post differences (e.g., Sandifer, 2008;Thompson & Harm, 2000). It is possible that the current treatment for-mat and dose were not sufficient to facilitate changes; further examina-tion of the differences in changes across participants might highlightother contributing factors.

Treatment acceptability was higher for mothers following the PCIT-based class compared to the existing facility class, however incarceratedmothers reported lower levels of acceptability for both the PCIT-basedclass (M= 38.4) and the existing parenting class (M= 35.4) comparedto treatment acceptability in PCIT studies (M = 40–43; Berkovits et al.,2010; Boggs et al., 2004).Mothers' overall satisfactionmay also be influ-enced by the correctional environment, nature of the facility recom-mendation to complete parent programming, group-based format,class size, or the lack of involvement of their children in the program.As treatment acceptability on the TAI has previously been associatedwith treatment gains (Brestan et al., 1999), limited improvement insome areas of parent–child functioning may also account for lowerlevels of satisfaction.

During the current parenting classes, mothers varied greatly in theirparenting attitudes and skills prior to PCIT as well as their class engage-ment and motivation. Incarcerated mothers may vary in their need forparenting classes based on differences such as their individualparenting-related characteristics, sentence length, and contact withtheir children. A thorough needs assessment for the population andscreening of each mother's specific parenting needs may add to theutility of this type of programming. In particular, the family's reunifica-tion andpermanency planning, the age of themother's children, and thepresence or absence of parenting challenges with their children couldbe useful to know prior to involvement in parenting programming.

Studies examining parent–child contact during incarceration havefound that the type and quality of these interactionsmay greatly predictparent and child behavioral outcomes (Poehlmann, Dallaire, Loper, &Shear, 2010). In families with children between the ages of 2.5 and7.5 years, Poehlmann (2005) found higher levels of parent–childinteraction without intervention to be associated with greater difficul-ties. If facility-based parenting classes are intended to assist withcontact, further examination of tailoring or adapting treatment to thecorrectional setting and population may be an area of future work. Asonly 13% of mothers reported visitation with their children duringincarceration, additional parent training topics such as alternative com-munication strategies and effective ways to manage parenting-relatedstress while not in direct contact with their children (e.g., use of parent-ing skills during phone conversation or letterwriting)might be relevantand should be carefully assessed.While some such as Eddy et al. (2008)describe the utility of a parenting support group to balance inmates'needed time to talk and process outside of the components of instruc-tion and skill building, there is some evidence that providing PCIT

246 A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

alone may be more effective for decreasing re-abuse rates than provid-ing families with multiple services at once (Chaffin et al., 2004).

Families of incarcerated mothers are often involved with multipleservice systems (e.g., corrections, juvenile justice, child welfare, andbehavioral health). The timing, context, and format of prevention andearly intervention should be considered in relation to family outcomesand cost-effectiveness. For example, parent training through correction-al programmingmay remove some barriers to the service experience ashighlighted by the 13% attrition rate in this sample in comparison to the40–60% attrition rate of child outpatient therapy (Kazdin, Holland, &Crowley, 1997). Additional intervention contexts such as child contactprograms and community-based parenting programs following paroleshould also be considered for this population as mothers involved inthese programs likely have more direct contact with their children.Future studiesmight also examine the use of individual family PCIT dur-ing visitation or reunification services because this standard PCIT formatallows for all components associated with large effects in parent andchild outcomes to be incorporated and has demonstrated long-termmaintenance of outcomes.

6.1. Limitations

Several limitations of the current study are important to note. First,data of child functioning were not included, and therefore no informa-tion is available to understand the severity of child behavioral difficul-ties of the participants' children. The mothers' children were notrequired to have behavior problems and parents were not required tohave a history of harsh parenting or child physical abuse, which resultedin varied parenting needs across participants and may have impactedtreatment effects. Future studies should examine participant character-istics associated with larger effects. Second, treatment and assessmentinteractions relied on role-play situations and behavioral rehearsalwith adults. Assessingmother–child interactionsusing theDPICS codingsystemwould provide information regarding the generalization of skillsduring interactionswithmothers' own children. Third, treatment integ-rity in the existing facility parenting class was instructor-reported andwas not elaborate. As a result, it is difficult to compare the existing facil-ity class to correctional parenting classes occurring across the country.Fourth, the current comparison group design could not delineate thecause of similar levels of change in parenting stress and risk for childabuse between groups. The addition of a wait-list control group wouldadd clarity to future findings as it would provide more understandingof change overtime with and without intervention. Fifth, only short-term outcomes were assessed in the current study. Long-term follow-up would provide a better understanding of whether parenting skilland attitude changes are maintained over time. Similarly, examinationof whether participation in parenting classes is associated with reduc-tions in criminal recidivism rates or re-abuse rates would have broadimplications.

6.2. Summary

These findings provide some initial evidence of the transportabilityof behavioral parent training programs into correctional settings.While the PCIT-based class was considered to be acceptable and effec-tive for improving demonstrated parenting skills during parentingrole-plays, participants in the facility-based class reported largerincreases in child development knowledge. Other parent outcomessuch as parenting stress and risk for child abuse were similar betweengroups. Future studies will be necessary to understand whether PCIT-based services provided in correctional settings as a supplementalcomponent or as standalone prevention or early interventionapproaches are effective for enhancing long-term parent and childoutcomes.

References

Abidin, R. (1995). Parenting stress index: Professional manual (3rd ed.). Lutz, FL: Psycho-logical Assessment Resources, Inc.

Bagner, D.M., Fernandez, M.A., & Eyberg, S. M. (2004). Parent–Child Interaction Therapyand chronic illness: A case study. Journal of Clinical Psychology in Medical Settings,11, 1–6.

Bartholomew, N. G., Knight, D. K., Chatham, L. R., & Simpson, D.D. (2002). Partners inparenting: A DATAR/First Choice treatment manual. Fort Worth, Texas: Texas ChristianUniversity.

Baumrind, D. (1967). Child care practices anteceding three patterns of preschool behav-ior. Genetic Psychology Monographs, 75(1), 43–88.

Bavolek, S. J., & Keene, R. G. (1999). Adult-Adolescent Parenting Inventory-AAPI-2: Adminis-tration and development handbook. Park City, UT: Family Development Resources, Inc.

Beck, A. J. (2000). Prisoners in 1999. (Report no. 183476). Washington, DC: U.S. Depart-ment of Justice. Bureau of Justice Statistics.

Beidas, R. S., Cross, W., & Dorsey, S. (2014). Show me, don't tell me: Behavioral rehearsalas a training and analogue fidelity tool. Cognitive and Behavioral Practice, 21, 1–11.

Berkovits, M.D., O'Brien, K. A., Carter, G. C., & Eyberg, S. M. (2010). Early identification andintervention for behavior problems in primary care: A comparison of two abbreviat-ed versions of Parent–Child Interaction Therapy. Behavior Therapy, 41, 375–387.

Block, K. J., & Potthast, M. J. (1998). Girl scouts beyond bars: Facilitating parent–childcontact in correctional settings. Child Welfare League of America, 27, 561–578.

Boggs, S. R. (1990, August). Generalization of treatment to the home setting: Direct obser-vation analysis. In. S. M. Eyberg (chair), Parent–Child Interaction Therapy (PCIT) out-come studies: Multiple method assessment. Symposium presented at the meeting ofthe American Psychological Association, Boston.

Boggs, S. R., Eyberg, S. M., Edwards, D. I., Rayfield, A., Jacobs, J., Bagner, D., & Hood, K. K.(2004). Outcomes of Parent−Child Interaction Therapy. A comparison of treatmentcompleters and study dropouts one to three years later. Child & Family BehaviorTherapy, 26, 1–22.

Borrego, J., Jr., Urquiza, A. J., Rasmussen, R. A., & Zebell, N. (1999). Parent–Child InteractionTherapy with a family at high risk for physical abuse. Child Maltreatment, 4, 331–342.

Brestan, E. V., Eyberg, S. M., Boggs, S. R., & Algina, J. (1997). Parent–Child InteractionTherapy: Parent perceptions of untreated siblings. Child & Family Behavior Therapy,19, 13–28.

Brestan, E., Jacobs, J., Rayfield, A., & Eyberg, S. M. (1999). A consumer satisfaction measurefor parent–child treatments and its relationship to measures of child behaviorchange. Behavior Therapy, 30, 17–30.

Capage, L. C., Bennett, G. M., & McNeil, C. B. (2001). A comparison between AfricanAmerican and Caucasian children referred for treatment of disruptive behaviordisorders. Child & Family Behavior Therapy, 23, 1–14, http://dx.doi.org/10.1300/J019v23n01_01.

Chaffin, M., Silovsky, J. F., Funderburk, B., Valle, L. A., Brestan, E. V., Balachova, T., et al.(2004). Parent–Child Interaction Therapy with physically abusive parents: Efficacyfor reducing future abuse reports. Journal of Consulting and Clinical Psychology, 72,500–510.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). LawrenceErlbaum Associates.

Cross, W. F., Seaburn, D., Gibbs, D., Schmeelk-cone, K., White, A.M., & Caine, E. D. (2011).Does practice make perfect? A randomized control trial of behavioral rehearsal onsuicide prevention gatekeeper skills. Journal of Primary Prevention, 32(3–4),195–211, http://dx.doi.org/10.1007/s10935-011-0250-z.

Davis, L. M., Bozick, R., Steele, J. L., Saunders, J., & Miles, J. M. V. (2013). Evaluating theeffectiveness of correctional education: A meta-analysis of programs that provide educa-tion to incarcerated adults. Santa Monica, CA: RAND Corporation (http://www.rand.org/pubs/research_reports/RR266).

Eddy, J., Martinez, C., Schiffmann, T., Newton, R., Olin, L., Leve, L., et al. (2008). Develop-ment of a multisystemic parent management training intervention for incarceratedparents, their children and families. Clinical Psychologist, 12, 86–98.

Eyberg, S. M. (2005). Tailoring and adapting Parent–Child Interaction Therapy to newpopulations. Education and Treatment of Children, 28, 197–201.

Eyberg, S. M., & Johnson, S. M. (1974). Multiple assessment of behavior modification withfamilies: Effects of contingency contradicting and order of treated problems. Journalof Consulting and Clinical Psychology, 42, 594–606.

Eyberg, S. M., Lewis, C. A., Butler, A.M., Chase, R., Fussell, R., Nelson, M., et al. (2009).Man-ual for group Parent–Child Interaction Therapy. University of Florida.

Eyberg, S. M., Nelson, M. M., Duke, M., & Boggs, S. R. (2005).Manual for the dyadic parent–child interaction coding system (3rd ed.). University of Florida.

Eyberg, S., & Pincus, D. (1999). Eyberg Child Behavior Inventory & Sutter-Eyberg StudentBehavior Inventory—Revised: Professional manual. Odessa, FL: Psychological Assess-ment Resources.

Eyberg, S. M., & Robinson, E. (1982). Parent–child interaction training: Effects on familyfunctioning. Journal of Clinical Child Psychology, 11, 130–137.

Fricker-Elhai, A., Ruggiero, K., & Smith, D. (2005). Parent–Child Interaction Therapy withtwomaltreated siblings in foster care. Clinical Case Studies, 4, 13–39, http://dx.doi.org/10.1177/1534650103259671.

Funderburk, B. W., Eyberg, S. M., Newcomb, K., McNeil, C. B., Hembree-Kigin, T., & Capage,L. (1998). Parent–Child Interaction Therapy with behavior problem children: Mainte-nance oftreatment effects in the school setting. Child & Family Behavior Therapy, 20,17–38.

Glaze, L. E., & Maruschak, L. M. (2008). Parents in prison and their minor children. Bureau ofJustice Statistics Special Report, NCJ 22984.

Gordon, H. R., & Weldon, B. (2003). The impact of career and technical educationprograms on adult offenders: Learning behind bars. The Journal of CorrectionalEducation, 54, 200–209.

247A.T. Scudder et al. / Children and Youth Services Review 46 (2014) 238–247

Green, B.L., Miranda, J., Daroowalia, A., & Siddique, J. (2005). Trauma exposure, mentalhealth functioning, and program needs of women in jail. Crime & Delinquency, 51,133–151.

Greene, S., Haney, C., & Hurtado, A. (2000). Cycles of pain: Risk factors in the lives ofincarcerated mothers and their children. The Prison Journal, 80, 3–23.

Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of youngAmerican children. Baltimore, MD: Paul H. Brookes.

Hood, K., & Eyberg, S. M. (2003). Outcomes of Parent–Child Interaction Therapy: Mothers'reports on maintenance three to six years after treatment. Journal of Clinical Child andAdolescent Psychology, 32, 419–429.

Houck, K., & Loper, A. (2002). The relationship of parenting stress to adjustment amongmothers in prison. American Journal of Orthopsychiatry, 72, 548–558, http://dx.doi.org/10.1037/0002-9432.72.4.548.

Hughes, M. J., & Harrison-Thompson, J. (2002). Prison parenting programs: A national sur-vey. The Social Policy Journal, 1, 57–74.

Kaminski, J. E., Valle, L. A., Filene, J. H., & Boyle, C. L. (2008). A meta-analytic review ofcomponents associated with parent training program effectiveness. Journal ofAbnormal Child Psychology, 36, 567–589.

Kauffman Foundation (2004). Closing the quality chasm in child abuse treatment: Identi-fying and disseminating best practices. In N.C.V.R.T. Center (Ed.), (Charleston, SC).

Kazdin, A. E., Holland, L., & Crowley, M. (1997). Family experience of barriers to treatmentand premature termination from child therapy. Journal of Consulting and ClinicalPsychology, 65, 453–463.

Kjellstrand, J. M., & Eddy, J. (2011). Parental incarceration during childhood, family con-text, and youth problem behavior across adolescence. Journal of OffenderRehabilitation, 50, 18–36.

Lee, E. L.,Wilsie, C. C., & Brestan-Knight, E. (2011). Using Parent–Child Interaction Therapyto develop a pre-parent education module. Children and Youth Services Review, 33,1254–1261.

Loper, A. B., & Tuerk, E. H. (2006). Parenting programs for incarcerated parents: Currentresearch and future direction. Criminal Justice Policy Review, 17, 407–427.

McDiarmid, M.D., & Bagner, D.M. (2005). Parent–Child Interaction Therapy for childrenwith disruptive behavior and developmental disabilities. Education and Treatment ofChildren, 28, 130–141.

McNeil, C. B., Eyberg, S. M., Eisenstadt, T., Newcomb, K., & Funderburk, B. (1991). Parent–Child Interaction Therapy with behavior problem children: Generalization of treat-ment effects to the school setting. Journal of Clinical Child Psychology, 20, 140–151.

McNeil, C. B., Herschell, A.D., Gurwitch, R. H., & Clemens-Mowrer, L. (2005). Training fos-ter parents in Parent–Child Interaction Therapy. Education and Treatment of Children,28, 182–196.

Milner, J. S. (1986). The Child Abuse Potential Inventory: Manual (2nd ed.). Webster, NC:Psytec.

Mumola, C. J. (2000). Incarcerated parents and their children.Washington, D.C.: US Depart-ment of Justice Bureau of Justice Statistics.

Murray, J., Farrington, D. P., & Sekol, I. (2012). Children's antisocial behavior, mentalhealth, drug use, and educational performance after parental incarceration: A sys-tematic review and meta-analysis. Psychological Bulletin, 138(2), 175–210, http://dx.doi.org/10.1037/a0026407.

Niec, L. N., Hemme, J. M., Yopp, J., & Brestan, E. V. (2005). Parent–Child InteractionTherapy: The rewards and challenges of a group format. Cognitive and BehavioralPractice, 12, 113–125.

Nixon, R. D.V. (2001). Changes in hyperactivity and temperament in behaviorally dis-turbed preschoolers after Parent–Child Interaction Therapy (PCIT). Behavior Change,18, 168–176.

Palmer, T. (1996). Programmatic and nonprogrammatic aspects of successful interven-tion. In A. Harland (Ed.), Choosing options that work: Defining the demand and evaluat-ing the supply (pp. 131–182). Thousand Oaks, CA: Sage.

Pew Center on the States (2008). One in 100: Behind bars in America 2008. Retrieved 11August 2011, from. www.pewcenteronthestates.org/

Pincus, D. B., Eyberg, S. M., & Choate, M. L. (2005). Adapting Parent–Child InteractionTherapy for young children with separation anxiety disorder. Education andTreatment of Children, 28, 163–181.

Poehlmann, J. (2005). Incarcerated mothers' contact with children, perceived family rela-tionships, and depressive symptoms. Journal of Family Psychology, 19(3), 350–357,http://dx.doi.org/10.1037/0893-3200.19.3.350.

Poehlmann, J., Dallaire, D., Loper, A. B., & Shear, L. D. (2010). Children's contact with theirincarcerated parents: Research findings and recommendations. American Psycholo-gist, 65, 575–598, http://dx.doi.org/10.1037/a0020279.

Reitman, D., &McMahon, R. J. (2012). Constance “Connie”Hanf (1917–2002): Thementorand the model. Cognitive and Behavioral Practice, http://dx.doi.org/10.1016/j.cbpra.2012.02.005.

Sabol, W. J., & Couture, H. (2008). Prison inmates at midyear 2007-statistical tables. Wash-ington, D.C.: US Department of Justice Bureau of Justice Statistics (NCJ 221944).

Sandifer, J. L. (2008). Evaluating the efficacy of a parenting program for incarceratedmothers. The Prison Journal, 88, 423–445.

Schuhmann, E. M., Foote, R., Eyberg, S. M., Boggs, S., & Algina, J. (1998). Parent–ChildInteraction Therapy: Interim report of a randomized trial with short-term mainte-nance. Journal of Clinical Child Psychology, 27, 34–45.

Showers, J. (1993). Assessing and remedying parenting knowledge among women in-mates. Journal of Offender Rehabilitation, 20, 35–46, http://dx.doi.org/10.1300/J076v20n01.

Surratt, H. (2003). Parenting attitudes of drug-involved women inmates. The PrisonJournal, 83, 206–220, http://dx.doi.org/10.1177/0032885503083002006.

Tempel, A. B., & Herschell, A.D. (2014s). Building an evidence-base for the training ofevidence based treatments: Use of an expert-informed approach. ProfessionalPsychology Research and Practice (in press).

Tempel, A. B., Wagner, S. M., & McNeil, C. B. (2013). Behavioral parent training skills andchild behavior: The utility of behavioral descriptions and reflections. Child and FamilyBehavior Therapy, 35(1), 25–40.

Thompson, P., & Harm, N. (2000). Parenting from prison: Helping children and mothers.Issues in Comprehensive Pediatric Nursing, 23, 61–81, http://dx.doi.org/10.1080/01460860050121402.

Timmer, S. G., Urquiza, A. J., Herschell, A.D., McGrath, J. M., Zebell, N. M., Porter, A. L., et al.(2006). Parent–Child Interaction Therapy: Application of an empirically supportedtreatment to maltreated children in foster care. Child Welfare League of America,919–939.

Timmer, S. G., Urquiza, A. J., Zebell, N. M., & McGrath, J. M. (2005). Parent–Child Interac-tion Therapy: Application to maltreating parent–child dyads. Child Abuse & Neglect,29, 825–842.

Tuerk, E. H., & Loper, A. B. (2006). Contact between incarcerated mothers and their chil-dren: Assessing parenting stress. Journal of Offender Rehabilitation, 43, 23–43.

U.S. Census Bureau (2014, February 7). State & County Quickfacts: West Virginia.Retrieved February 7, 2014, from. http://quickfacts.census.gov

West, H. C. (2010). Prison inmates at midyear 2009—statistical tables.Washington, D.C.: USDepartment of Justice Bureau of Justice Statistics (NCJ 230113).

Woodward, R. (2003). Families of prisoners: Literature review on issues and difficulties.Wright, E. M., Salisbury, E. M., & VanVoorhis, P. (2007). Predicting the prison misconducts

of women offenders: The importance of gender-responsive needs. Journal ofContemporary Criminal Justice, 23, 310–340.