A Pilot Study of Universal Teacher–Child Interaction Training at a Therapeutic Preschool for Young Maltreated Children

Abstract

Young children are disproportionately exposed to maltreatment but are underrepresented in research on effective treatments. Universal Teacher-Child Interaction Training (TCIT-U), developed from Parent-Child Interaction Therapy, may be especially appropriate for maltreated children as they often experience caregiver disruptions which pose challenges to traditional parent-child treatment. Furthermore, research suggests that teachers can play an important role for children who lack positive caregiving experiences. The current study examined the effectiveness of TCIT-U versus treatment-as-usual (TAU) at a therapeutic preschool for youth exposed to maltreatment. Thirty-eight children (2–5 years old) and eight teachers from four classrooms participated in the study. Teacher behaviors were observed and coded at baseline, mid-treatment, post-treatment, and 3-month follow-up. Teachers reported on children’s behavior and social-emotional skills at baseline, post-treatment, and 3-month follow-up. TCIT-U teachers demonstrated substantial increases in positive attending skills (PRIDE [Praise, Reflection, Imitation, Description, and Enjoyment] skills) and decreases in negative talk and questions during intervention phases, and these skills were maintained at follow-up. In addition, children in the TCIT-U classrooms demonstrated a significantly greater increase in overall social-emotional skills by post-treatment than children in the TAU classrooms, and effect sizes were moderate for all child outcomes. Findings provide preliminary support for TCIT-U’s effectiveness in a therapeutic setting for children exposed to maltreatment.

Keywords

Teacher–Child Interaction Training child maltreatment teacher behavior therapeutic day treatment child behavior social-emotional skills

Early exposure to maltreatment, including physical, sexual, and psychological abuse, and neglect, can interfere with normative development (Lieberman, Chu, Van Horn, & Harris, 2011; Wilson, Hansen, & Li, 2011) and have long-term consequences, including later substance abuse, delinquency, school problems, and various psychological disorders (Cook et al., 2005). According to the National Child Abuse and Neglect Data System (NCANDS), young children are the most vulnerable to maltreatment (U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau, 2018). Complex trauma refers to the chronic, simultaneous, or subsequent experiences of multiple traumatic events occurring within the caregiving system that begin in early childhood (Cook et al., 2005). Complex trauma, often including maltreatment and subsequent stressors (e.g., foster care placement), can lead to impairment in multiple developmental areas, including social-emotional development, wherein children’s abilities to understand emotions, modulate behavioral impulses and emotional states, and regulate attention are impaired (Chu & Lieberman, 2010; Cook et al., 2005; Pears, Fisher, Bruce, Kim, & Yoerger, 2010). Caregivers may frequently exhibit negative affect and low sensitivity to the child’s needs, provide limited models and support of emotional regulation, and use harsh or inconsistent discipline, which contribute to such social-emotional deficits (Robinson et al., 2009). Social-emotional deficits resulting from complex trauma play a key role in the development of emotional and behavioral concerns. For example, social-emotional concerns, such as emotion dysregulation, likely worsen behavior problems, harm relationships with peers and teachers, and lead to poor school adjustment, especially in preschool and kindergarten when children are faced with new demands for self-regulation and social engagement (e.g., Denham, Bassett, Way, et al., 2012; Kalvin, Bierman, Gatzke-Kopp, 2016).

Adults have an important role in fostering social-emotional development. Consequently, evidence-based interventions for young children—especially those based in behavioral and attachment theories—typically include strategies for enhancing adult–child interactions. For instance, many interventions targeting externalizing problems of young children are dyadic parent management programs (e.g., Parent–Child Interaction Therapy [PCIT], Eyberg, Boggs, & Algina, 1995; Child–Parent Psychotherapy, Lieberman & Van Horn, 2005), thus requiring a caregiver to actively participate in treatment. Unfortunately, young children who have experienced complex trauma often experience caregiver disruptions or inconsistencies (e.g., foster care placement). These children may not have a consistent caregiver who can provide consistent social emotional support or who can participate in parent–child interventions. Given these challenges to treatment, school-based interventions may be especially appropriate for providing evidence-based mental health services to young children exposed to complex trauma.

The Role of Teacher–Child Relationships

Teachers can provide emotional support, reinforcement, instruction, and modeling to help children develop social-emotional skills (Denham, Bassett, & Zinsser, 2012). Research has demonstrated that warm, responsive teacher–child relationships can decrease child externalizing and internalizing behavior problems (Hamre & Pianta, 2001). Moreover, there is evidence that positive teacher-child relationships can buffer the negative developmental and behavioral outcomes related to poor early caregiving experiences (Buyse, Verschueren, & Doumen, 2011; Sabol & Pianta, 2012). Given the amount of time children spend at school, interventions with promise for improving teacher-child relationships, such as Teacher-Child Interaction Training (TCIT), are an avenue for reaching children who may otherwise experience barriers to accessing quality services.

Furthermore, child behavior problems are a major contributing factor to teacher stress and burnout (Friedman-Krauss, Raver, Morris, & Jones, 2014). Highly stressed teachers tend to have more conflictual teacher–child interactions, less positive classroom environments, and more difficulty effectively managing their classrooms (Dobbs & Arnold, 2009) which perpetuate child behavior problems (Siekkinen et al., 2013). Providing teachers with training in enhancing positive teacher-child relationships and effectively managing child behavior problems can improve both teacher and child well-being. Although serving a critical role in children’s social-emotional development and school readiness, few teachers receive training in child behavior management and/or relationship-building techniques.

Teacher training programs are emerging to address the need for the prevention and intervention of behavior problems and promotion of social-emotional skills in early childhood education settings. For example, Positive Behavior Interventions and Supports (PBIS; Stanton-Chapman, Walker, Voorhees, & Snell, 2016) and the Pyramid Model (Fox, Dunlap, Hemmeter, Joseph, & Strain, 2003) are two models that utilize a tiered approach with varying intervention intensity. Such tiered models begin with universal prevention in which all children in the classroom receive the strategies. Strategies often include establishing rules and routines and increasing positive teacher–child interactions. Secondary prevention strategies target children who are at risk for behavior problems or who have not responded to universal strategies, while tertiary or individualized practices are designed for children with identified challenging behaviors (Hemmeter, Snyder, Fox, & Algina, 2016). These and other universal and secondary prevention programs (e.g., My Teaching Partner and Making the Most of Classroom Interactions, Early, Maxwell, Ponder, & Pan, 2017; BEST in CLASS, Sutherland et al., 2018) typically involve group or individual didactic instruction (e.g., workshops, manuals, web-based materials) and sometimes include individualized consultation (e.g., classroom observations, performance feedback). The present study examined universal prevention program, Teacher–Child Interaction Training (TCIT), which uniquely incorporates live or in-vivo coaching of teachers’ positive communication and behavioral modification strategies.

Teacher–Child Interaction Training

TCIT is a promising intervention aimed at improving teacher–child relationships and effectively managing children’s behaviors (McIntosh, Rizza, & Bliss, 2000). TCIT was developed as a modification of PCIT, an evidence-based treatment with well-established support for decreasing disruptive child behaviors and parental stress and improving parent–child relationships with children ages 2 to 7 years old (Eyberg, Nelson, & Boggs, 2008). Similar to PCIT, TCIT involves two phases: Child-Directed Interaction (CDI) focuses on teacher–child relationship building through teachers’ use of PRIDE skills (Praise, Reflection, Imitation, Description, and Enjoyment), and Teacher-Directed Interaction (TDI) focuses on behavioral strategies appropriate for the classroom (Gershenson, Lyon, & Budd, 2010). Over the past decade, TCIT has evolved from a dyadic teacher–child intervention targeting a specific child to a classroom-wide intervention with the hope of reaching a greater number of children.

An adaptation, universal TCIT (or TCIT-U), instructs teachers to increase their use of PRIDE skills and decrease negative comments and unnecessary commands and questions. The behavior modification strategies include using effective commands, prompts, natural consequences, differential social attention, and a modified time-out for use in a classroom which is determined in collaboration with teachers. Through the course of TCIT-U, teachers are coached in their classrooms across a variety of natural situations (e.g., small group, story time, transitions) to encourage generalization of teachers’ skill use. TCIT-U utilizes a time-limited approach in training (i.e., set number of didactics and coaching) in an effort to efficiently train teachers with varying skill levels. A more detailed review of TCIT-U is provided by Gershenson and colleagues (2010).

Lyon and colleagues (2009) were the first to evaluate TCIT-U’s effect on teacher behavior across four classrooms at an urban child care setting serving low-income, ethnic minority children. Researchers used multiple-baseline analysis to examine the effects of TCIT-U on teacher behavior. Results yielded small to moderate effect sizes (ES) for teacher behavior; child outcome data were not reported. Garbacz, Zychinski, Feuer, Carter, and Budd (2014) conducted another study in the same setting but in different classrooms with younger children (2- to 3-year-old children) than Lyon and colleagues. Their study examined associations between teacher skill change and child outcomes (i.e., behavior problems and social-emotional skills). They found that high levels of teacher skill change were associated with greater social-emotional skills and fewer behavioral problems in children with baseline ratings in the problem range (Garbacz et al., 2014). Fernandez and colleagues (2015) investigated TCIT-U compared with non-TCIT-U in randomly assigned kindergarten and first-grade classrooms in an urban public school. Similar to the other studies discussed here, results documented increased rates of teachers’ positive attention and decreased rates of negative attention to child behavior. There were small to large effects of TCIT-U on teacher behavior compared with the control group. TCIT-U teachers also had significantly lower distress due to child behavior problems than teachers in the control group. Most recently, Budd, Garbacz, and Carter (2016) examined teacher and child outcomes comparing researcher-delivered TCIT-U versus local staff–delivered TCIT-U in a public elementary school. They found similar improvements in positive teacher behaviors and child social-emotional skills and behavior problems in both groups, suggesting TCIT-U can be effectively implemented by local school staff. Overall, preliminary literature on classroom-wide or universal TCIT suggests that teachers in a range of educational settings learn the skills quickly and that the intervention leads to teacher and child behavior change.

Therapeutic Day Treatment

The present study evaluated TCIT-U at a therapeutic day treatment preschool designed for young children with a history of maltreatment. Many children are referred to therapeutic day treatment because their needs are not met by typical preschools, day care programs, or outpatient treatment due to behavioral and/or developmental problems (Tse, 2006). Therapeutic day treatment settings are focused on alleviating behavioral and emotional problems in young children. Although studies document commonalities in program components, research on the effectiveness and efficacy of therapeutic day treatment settings for children exposed to trauma is limited, and there does not appear to be a protocol that guides the mental health service delivery in such classrooms (Kanine, Tunno, Jackson, & O’Connor, 2015). TCIT-U offers an empirically supported intervention targeting the child behavior and social-emotional problems that day treatment aims to address.

Aims and Hypotheses

The current study aimed to expand upon the growing literature of TCIT-U by examining its effects on teacher and child behavior and child social–emotional skills, specifically with children at risk for social-emotional and behavioral problems due to their complex trauma exposure. In support of TCIT-U’s potential as a school-based intervention for children exposed to maltreatment, PCIT—TCIT’s parent counterpart—has been identified as a “best practice” and “supported and acceptable” intervention for child maltreatment (Thomas & Zimmer-Gembeck, 2012). Evidence shows that youth exposed to maltreatment who participate in PCIT demonstrate decreased externalizing and internalizing behavior problems, as well as decreased parental stress (Timmer, Urquiza, & Zebell, 2006; Timmer, Urquiza, Zebell, & McGrath, 2005). Given PCIT’s effectiveness at targeting issues of children exposed to maltreatment, the present study assessed whether TCIT-U has similar success.

In addition, existing literature lacks studies testing TCIT-U relative to other psychosocial interventions. Thus, this study included a comparison group to test whether TCIT-U is effective compared with treatment-as-usual (TAU) in a therapeutic day treatment setting. The following were hypothesized:

Hypothesis 1: Teachers in TCIT-U classrooms would demonstrate increases in positive attending skills and decreases in negative attending by post-treatment and maintain skills at a 3-month follow-up versus TAU teachers.

Hypothesis 2: Children in TCIT-U classrooms would demonstrate greater decreases in disruptive behaviors and increases in social-emotional skills by post-treatment and maintain improvements at follow-up compared with children in TAU classrooms.

Method

Participants and Setting

Our study was conducted at a therapeutic day treatment preschool in an urban Midwestern city. The preschool was non-profit and specifically served children, infancy to 5 years old, with a history of maltreatment and referred for behavioral and/or developmental concerns. The protocol for the study was approved by the university Institutional Review Board, state Children’s Division Department of Social Services, and preschool administration.

Eight teachers (TCIT-U n = 4, TAU n = 4) and 39 children (TCIT-U n = 21, TAU n = 18) from four classrooms were eligible to participate in present study. Two classrooms had 2- to 3-year-old children (toddler classrooms); while two classrooms served 4- to 5-year-old children (preschool classrooms). Approximately 80% of children attending the preschool were in custody of the Children’s Division living in foster or kinship care. All children and teachers present for baseline data collection were eligible to participate in the study. One of the 39 eligible children was excluded from analyses because he discharged from treatment soon after the study began.

Child demographic information was obtained from the preschool’s program evaluation records which included data from intake interviews and questionnaires by caregivers, case workers, and teachers. Across all youth, 90% were reported to have experienced parental substance abuse, 72% neglect, 56% witnessed violence, 33% were exposed to illegal drugs prenatally, and 15% and 5% were victims of physical and sexual abuse, respectively. Maltreatment types were not independent from each other; thus, many children experienced multiple types of maltreatment. Treatment groups differed significantly by age (F = 9.78, p = .003), with the TCIT-U group being older on average (4.03 years, SD = 0.72) than the TAU group (3.25 years, SD = 0.85), and history of parental mental illness (χ² = 4.14, p = .04), with the TCIT-U group experiencing more parental mental illness (71.4%) versus TAU children (38.9%). Table 1 depicts the children’s descriptive information by treatment group.

Table 1.

Child Descriptive Information.

Characteristic	TCIT (n = 21)	TAU (n = 18)	Test statistic^a
Gender			1.61
Female	47.6%	27.8%
Male	52.4%	72.2%
Race/Ethnicity			1.14
Black/African American	66.7%	55.6%
White/Caucasian	14.3%	27.8%
Hispanic/Latino	4.8%	5.6%
Bi-/Multi-racial	14.3%	11.1%
Mean age (SD)	4.03 (0.72)	3.25 (0.85)	9.76**
Maltreatment
Neglect	71.4%	72.2%	.003
Physical abuse	23.8%	5.6%	2.48
Sexual abuse	—	11.1%	2.46
Witness violence	47.6%	66.7%	1.43
Parental substance abuse	90.5%	88.9%	0.03
Parental mental illness	71.4%	38.9%	4.17*
Mean preschool attendance (SD)	48.33 (33.13)	30.57 (23.57)	3.60
Mean time in study (SD)	13.90 (7.75)	15.03 (7.97)	0.19

Note. TCIT = Teacher–Child Interaction Training; TAU = treatment-as-usual; preschool attendance = number of weeks in attendance prior to study; time in study = number of weeks from baseline to follow-up.

Test Statistic = F test for continuous variables and Pearson chi-square test for categorical variables.

p < .05. **p < .01.

Teachers were 62.5% female (n = 5) with a mean age of 40.38 years (SD = 13.73). Half of the teachers identified as Black/African American, while others identified as White/Caucasian (n = 2) or Bi/Multi-Racial (n = 2). Most teachers had college degrees (n = 6, 75%), and more than half (n = 5) had degrees in early childhood development or education. Teachers averaged about 8 years (SD = 11.58) at the therapeutic preschool and 7 years (SD = 12.06) teaching in their current classroom. On average, teachers in the TAU group had significantly more years of total teaching experience (F = 7.47, p = .03) than teachers in the TCIT-U group (25.90 and 9.25 years, respectively).

Procedures

TCIT-U protocol

TCIT-U was implemented over 12 weeks and included group didactic and individualized coaching sessions. The four TCIT-U teachers (two from a toddler classroom and two from a preschool classroom) participated in eight 1.5-hr weekly group didactic sessions, including four CDI sessions and four TDI sessions. The trainer, a doctoral student in clinical child psychology who was trained by TCIT-U developers, also met individually with TCIT-U teachers for 20-min individualized coaching sessions in their classrooms 1 to 2 times per week. The current study followed the protocol as described in Gershenson et al. (2010) and used in other TCIT-U studies (Garbacz et al., 2014; Lyon et al., 2009). During the CDI phase, teachers learned to Praise specific appropriate behavior (e.g., “Thank you for waiting your turn”); Reflect appropriate speech by repeating, paraphrasing, or elaborating on children’s words; Describe appropriate behavior as it occurs (e.g., “You are picking up the toys”); and show Enjoyment during interactions to increase positive classroom climate. These behaviors are often targeted in early childhood special education, however, can have different terminology. For instance, behavioral descriptions have been referred to as narration or mapping (Kaiser & Roberts, 2013; Lane et al., 2016); reflections have been called language expansions (Barton & Wolery, 2007; Kaiser & Roberts, 2013); and labeled and unlabeled praises have also been termed as behavior-specific and general praise, respectively (Barton & Wolery, 2007; Royer, Lane, Dunlap, & Ennis, 2016).

Teachers also learned how to reduce unnecessary (i.e., “rapid-fire”) questions and commands, avoid negative talk or criticisms, and ignore mild inappropriate behavior, such as whining and fidgeting. During the TDI phase, teachers learned how to use specific behavior management strategies, such as the use of effective commands and consistent follow-up strategies including Sit-and-Watch, a modified time-out procedure appropriate for classroom use that is used for aggressive and unsafe behaviors (Gershenson et al., 2010).

TAU

During the study period, teachers in the TAU classrooms continued to use their typical strategies for managing child disruptive behavior and social-emotional competence. The preschool utilized a program called Conscious Discipline (Bailey, 2000), a classroom method of teaching self-regulation to children. Classrooms with children 3 years and older also used Al’s Pals (Geller, 1999), a curriculum for teaching social skills through use of lessons with puppets and music. Teachers previously participated in online training and/or workshops for these methods. Although TCIT-U and Conscious Discipline both aim to improve adult–child relationships by increasing warm, sensitive interactions and environments, Conscious Discipline discourages the use of rewards and consequences. Instead, it suggests a positive school climate will lead children to develop the internal motivation to use appropriate behavior (Bailey, 2000; Hoffman, Hutchinson, & Reiss, 2009). In addition, Conscious Discipline is delivered through workshops, while TCIT-U also includes individualized coaching. There is no published peer-reviewed research on the effects of Conscious Discipline on child outcomes.

Coder training process

Graduate- and undergraduate-level coders were trained in the observational system by the TCIT-U trainer. Coder training included review of the Dyadic Parent–Child Interaction Coding System manual (DPICS-IV; Eyberg, Nelson, Ginn, Bhuiyan, & Boggs, 2013) and modified TCIT-U coding procedures, completion of homework assignments and quizzes to evaluate proper identification of teacher verbalizations, and completion of practice coding from videotaped and live teacher–child interactions. To maintain reliability while coding, coders met weekly to review any possible coding issues. For training purposes, percent agreement was obtained by comparing two coders’ total agreements and disagreements across eight categories per 5-min video-recorded observations. The formula to compute percent agreement per observation was (total number of agreements across categories / [total agreements + total disagreements]) × 100 (Eyberg et al., 2013). Across all practice videos and coders during training, mean percent agreement was borderline acceptable: baseline = 79%; mid-treatment = 75%; post-treatment = 80%; follow-up = 80%. However, before a given coder was allowed to code for the study, they had to establish 80% agreement, averaged across their practice videos, with the TCIT-U trainer and other coders. Eighty percent is the recommendation for clinical readiness from the PCIT manual (Eyberg et al., 2013).

Research procedures

The four participating classrooms were assigned to TCIT-U or TAU so that each group included one toddler classroom and one preschool classroom. The present study is a quasi-experimental or nonrandomized design because one classroom declined TCIT-U training due to availability limitations, however, agreed to participate in the TAU condition. Child outcome data via teacher-report were collected at baseline (2 weeks prior to CDI), post-treatment (end of TDI phase), and follow-up (3 months after post-treatment).

Teacher behavior data were collected via video-recordings of teacher–child interactions in their classrooms. Each teacher was video-recorded 4 times within a 2-week period at baseline, mid-treatment (after the CDI phase), post-treatment, and 3-month follow-up. If a video was of poor audio quality, the teacher was recorded again until four sessions were obtained. A central goal of TCIT-U is for teachers to use skills with multiple children, rather than just one-on-one, and across situations. Therefore, teachers were coded during interactions with at least two children. Fifty-two percent of observations included teacher–child interactions with five or more children (i.e., large group) while 47.5% occurred in small groups, or two to four children. Group size did not differ between study groups (χ² = .16, p = .69): TCIT-U (31 large groups, 26 small groups) and TAU (32 large groups, 31 small groups). In addition, observations included a range of activities: 34% center activities (e.g., arts-and-crafts, sensory work, learning activities), 24% circle time (e.g., morning circle, group lessons), 17.5% free play, 3% playground, 3% songs, 6% reading books, and 11% during transitions (e.g., cleaning up, getting in line). No significant differences in type of activities were observed between TCIT-U and TAU groups (χ² = 9.73, p = .29).

Measures

Teacher behavior

Positive teacher behaviors targeted in TCIT-U were assessed via the DPICS-IV (Eyberg et al., 2013), modified for use with teachers in classroom settings (Gershenson et al., 2010). Coders tallied the frequency of eight TCIT-U verbalizations during 5-min video-recorded observations. Behavioral Descriptions are nonevaluative statements in which the teacher describes what a child is doing while he or she is performing the behavior (e.g., “You are building with the blocks”). Reflections are statements in which the teacher repeats, paraphrases, or elaborates upon a child’s verbalization (e.g., child says, “It’s a space ship” and teacher reflects, “You built a space ship”). Praise is a verbalization that expresses “favorable judgement of an attribute, product, or behavior of a child” (Eyberg et al., 2013, p. 55). Two types of praise were coded. Labeled Praises say specifically what the teacher approves of (e.g., “Thank you for putting the blocks away”), while Unlabeled Praises are vague or less specific (e.g., “Good job”). Negative Talk is a verbal expression of disapproval for a child’s behavior, attributes, or products or tells a child what not to do (e.g., “Stop running”). Questions are inquiries that request a verbal response from a child (e.g., “What color is the block?”). Commands are statements in which a teacher directs the behavior of a child. Two types of commands were coded. Direct Commands are statements that direct a child to perform a specific vocal or motor behavior (e.g., “Please put the blocks in the bucket”), while Indirect Commands are suggestions or unclear requests for a child to perform a vocal or motor behavior (e.g., “Can you put the toys away?”) (Eyberg et al., 2013). Sit-and-Watch procedures were not coded in the present study.

A total of 120 video-recorded observations of the teachers (32 baseline, 32 mid-treatment, 30 post-treatment, and 26 follow-up) were coded by a designated “master” coder and a combination of four other coders. The master coder was an advanced undergraduate research assistant who was greater than 80% reliable with the TCIT-U trainer during the training phase (i.e., practice videos). Although efforts were made to blind coders to teachers’ condition (e.g., coders were not explicitly told teachers’ group assignment), given the small size of the study, it is likely that coders were aware of study condition. Sixty-eight percent of the observations were double-coded. Intraclass correlations (ICCs) from a one-way random effects model were calculated for a robust measure of interrater consistency for each coded category and to allow a measure of agreement across multiple coders. The ranges of ICCs across time-points demonstrated good reliability: Labeled Praise (ICC = .95-.98); Unlabeled Praise (ICC = .96-.97); Reflections (ICC = .89-.95); Behavioral Descriptions (ICC = .95-.97); Direct Commands (ICC = .95-.96); Indirect Commands (ICC = .86-.91); Questions (ICC = .97-.98); and Negative Talk (ICC = .86-.89).

Child behavior and social-emotional competence

Teachers’ perceptions of changes in child behavior problems and social-emotional skills were assessed with the Devereux Early Childhood Assessment–Clinical Form (DECA-C; LeBuffe & Naglieri, 2003). The DECA-C is a strengths-based standardized and norm-referenced 62-item behavior rating scale designed for children ages 2 to 5 years. Children were rated by one of their teachers on a Likert-type scale ranging from 0 (never) to 4 (very frequently) to document how often within the past 4 weeks a child exhibited various competencies and problem behaviors. Child social-emotional competence was measured via the Total Protective Factors (TPF) scale. The TPF includes subscales of Initiative (i.e., child’s ability to use independent thought and action to meet own needs), Self-Control (i.e., child’s ability to experience and appropriately express a range of feelings), and Attachment (i.e., extent of mutual, strong relationships between a child and significant adults). A TPF score of 40 and below indicates a concern while a score of 60 and above indicates a strength. Child behavior problems were measured via the Behavioral Concerns (BC) scale. The BC includes subscales for Withdrawal/Depression (i.e., emotional or social withdrawal from reciprocal interactions with peers or adults), Emotional Control Problems (i.e., difficulty modifying overt expression of negative emotions), Attention Problems (i.e., child’s ability to focus while ignoring other stimuli), and Aggression (i.e., hostile or destructive acts toward others or things). A BC scale of 60 and above indicates a potential problem area. In the present study, internal consistency of each scale was good across all time-points: TPF (α = .93-.94) and BC (α = .91-.94).

Training fidelity

Standard fidelity checklists do not presently exist for TCIT-U implementation. Fidelity checklists were designed for this project from the training outline and materials created by TCIT-U developers. In the current study, the TCIT-U trainer completed a checklist after each group training session. Didactic training checklist items included explanations of specific skills taught, video examples of skills or practice coding, role-play or practice of skills, and review of homework. Individual coaching checklist items included coding teacher for 5 min, coaching teacher for 10 min, providing brief verbal feedback on skill use, and completing written feedback.

TAU

The TAU programs used in the preschool also do not have fidelity measures. In an effort to measure TAU, teachers completed a brief survey about the amount of training received in Conscious Discipline and Al’s Pals at baseline and reported the frequency of lessons and activities for each method done in the classroom in the last week at baseline, mid-treatment, post-treatment, and follow-up.

Results

Data were analyzed using SPSS 22 software. Teacher behavior was evaluated via visual inspection of graphed data for changes in the mean frequency of TCIT-U skills per phase and between TCIT-U and TAU groups. For child outcomes, paired t tests were conducted to determine significant changes in DECA-C scores in each study group separately. Then, repeated measures analyses of variance (rANOVA) with a between-subjects factor were performed to determine whether any changes in the dependent variable over time (Time) were due to treatment condition (Group). Covariates included classroom, to account for nesting of children in classrooms, and number of weeks children participated in the study, to account for child participant attrition. For a measure of the strength and direction of the differences in child behavior between TCIT-U and TAU groups, ES were computed using Hedges’s g (Hedges & Olkin, 1985). To account for baseline differences, ES were adjusted by subtracting the baseline-ES from the post-ES (Durlak, 2009). Levene’s tests of homogeneity of variance for DECA-C across time-points were non-significant, and there were no significant baseline differences in DECA-C scores between the groups. Skewness, kurtosis, and z score estimates were within appropriate limits for each DECA-C variable across time-points.

Attrition

Of the 38 children with posttreatment data, attrition in the present study was 32% (n = 12), to be expected because of discharge from treatment or moving classrooms prior to post-treatment data collection. Data were collected for these children upon their discharge or move because they were present for at least the first phase of the study (CDI or TAU equivalent). Attrition analysis was conducted to determine possible reasons for child attrition. Separate variance t tests and chi-square tests found that children who discharged before reaching post-treatment data collection had attended the preschool significantly longer than children who completed posttreatment data collection (t = −2.7, p = .01; completed post = 31.0 weeks; missing post = 58.5 weeks). To account for this attrition, pre–post analyses were completed utilizing intent-to-treat (ITT) analysis by way of the last observation carried forward (LOCF) approach. Children’s data at discharge (i.e., children who completed at least CDI) were carried forward to their post-treatment scores. Seven of the 20 children in the TCIT-U group and five of the 18 children in the TAU group had posttreatment scores using this LOCF method. Conclusions made about the effectiveness of TCIT-U in the present study should be made with caution and are likely conservative estimates. By 3-month follow-up data collection, child attrition increased to 71% of the original sample (n = 27). Analyses using follow-up data were conducted on a subset of children who completed all phases of the study (N = 11). Follow-up results should be evaluated with extreme caution given this level of attrition and sample size.

One TCIT-U teacher resigned from the agency during the course of the study. The teacher completed all phases of TCIT-U. Due to scheduling conflicts, this teacher’s post-treatment observations were conducted in one 20-min session that was coded in four consecutive 5-min segments. No observations were completed at follow-up for this teacher. Figure 1 depicts child and teacher participant flow.

Figure 1.

Participant flow through the study.

Teacher Behavior

Graphs depict the observational data of teacher skills by treatment group for each coding category across time-points. This graphical presentation allows for examination of teacher behavior within treatment groups across phases of the study, as well as comparison between treatment groups. Visual inspection of these graphs included evaluating changes in the mean frequency of selected TCIT-U skills (PRIDE and Avoid skills) from phase to phase for the expected direction and level.

PRIDE skills

During baseline observations, TCIT-U teachers, on average, exhibited limited use of PRIDE skills and demonstrated immediate improvements following the CDI phase (mid-observations). TCIT-U teachers’ mean frequency of PRIDE skills were then maintained at post (TDI phase) and 3-month follow-up. On average, TCIT-U teachers’ mean use of PRIDE skills (Labeled Praises, Behavior Descriptions, and Reflections) at mid, post, and follow-up ranged from 3 to 19.5 times greater than their baseline means. TAU teachers similarly demonstrated a low frequency of PRIDE skills at baseline. In contrast to TCIT-U teachers, TAU teachers showed little to no improvements over time. See Table 2 and Figure 2 for comparisons of TCIT-U and TAU teachers’ PRIDE skills.

Table 2.

Teacher Behavior Coding: PRIDE Skills.

Category	Time	TCIT	TAU	Effect sizes
Category	Time	M (SD)	M (SD)	g (95% CI)	Adj g
LP	Pre	0.31 (0.46)	0.13 (0.33)	0.44 [–0.26, 1.14]
	Mid	6.06 (3.49)	0.50 (1.00)	2.11 [1.25, 2.98]^a	1.67
	Post	5.64 (2.97)	0.75 (0.97)	2.22 [1.31, 3.13]^a	1.78
	FU	5.82 (4.32)	0.73 (0.85)	1.72 [0.81, 2.63]^a	1.28
UP	Pre	2.44 (3.90)	4.06 (3.51)	−0.43 [–1.13, 0.28]
	Mid	3.69 (2.62)	3.50 (2.96)	0.07 [–0.63, 0.76]	0.49
	Post	4.43 (3.29)	4.25 (4.09)	0.05 [–0.67, 0.76]	0.47
	FU	3.73 (3.25)	2.67 (1.99)	0.40 [–0.39, 1.18]	0.82
RF	Pre	1.44 (1.77)	1.75 (2.11)	−0.16 [–.085, 0.54]
	Mid	4.13 (5.01)	1.81 (1.47)	0.61 [–0.10, 1.32]	0.77
	Post	5.14 (4.94)	3.44 (3.53)	0.39 [–0.33, 1.11]	0.54
	FU	5.27 (3.72)	2.13 (1.75)	1.11 [0.27, 1.94]^a	1.26
BD	Pre	1.00 (1.41)	1.06 (1.30)	−0.04 [–0.74, 0.65]
	Mid	6.44 (4.57)	0.56 (1.06)	1.73 [0.92, 2.54]^a	1.77
	Post	4.71 (4.85)	1.69 (1.72)	0.83 [0.08, 1.58]^a	0.87
	FU	8.55 (7.28)	2.53 (4.05)	1.04 [0.21, 1.86]^a	1.08

Note. PRIDE = Praise, Reflection, Imitation, Description, and Enjoyment; TCIT = Teacher–Child Interaction Training; TAU = treatment-as-usual; g = Hedges’s g; CI = confidence interval; Adj g = g adjusted for pre-treatment g (Durlak, 2009); LP = labeled praises; FU = follow-up; UP = unlabeled praises; RF = reflections; BD = behavior descriptions.

95% CI indicates statistical significance.

Figure 2.

TCIT-U and TAU teachers’ count of PRIDE skills per observation at pre, mid, post, and follow-up.

Avoid skills

During baseline observations, TCIT-U teachers exhibited a high frequency of Avoid skills (Questions, Commands, and Negative Talk). However, TCIT-U teachers demonstrated substantial decreases in frequency by the end of the CDI phase (mid) and maintained a low frequency of Avoid skills through post and 3-month follow-up. On average, Avoid skills decreased by a range of 22% to 81% from baseline to mid, post, and follow-up. The largest decreases were observed between baseline and mid, or the end of the CDI phase, which is when teachers learn strategies for increasing positive attention while decreasing their use of questions and commands. TAU teachers similarly demonstrated a high frequency of Avoid skills at baseline. In contrast to TCIT-U teachers, TAU teachers showed little to no improvements over time. See Table 3 and Figure 3 for comparison of TCIT teachers’ Avoid skills.

Table 3.

Teacher Behavior Coding: “Avoid or Don’t” Skills.

Category	Time	TCIT	TAU	Effect sizes
Category	Time	M (SD)	M (SD)	g (95% CI)	Adj g
DC	Pre	17.69 (8.68)	13.88 (6.73)	0.48 [–0.22, 1.18]
	Mid	8.63 (6.33)	16.88 (7.56)	−1.15 [–1.90, –0.41]^a	−1.63
	Post	10.36 (6.34)	14.19 (5.02)	−0.66 [–1.39, 0.08]	−1.14
	FU	12.18 (8.87)	12.27 (12.18)	−0.01 [–0.79, 0.77]	−0.49
IC	Pre	7.81 (5.70)	6.69 (3.29)	0.23 [–0.46, 0.93]
	Mid	6.50 (3.87)	8.13 (4.97)	−0.36 [–1.06, 0.34]	−0.59
	Post	7.86 (6.14)	9.69 (5.80)	−0.30 [–1.02, 0.42]	−0.53
	FU	7.00 (4.71)	9.80 (4.83)	−0.57 [–1.36, 0.23]	−0.80
QU	Pre	12.88 (9.98)	11.94 (5.75)	0.11 [–0.58, 0.81]
	Mid	6.06 (5.17)	13.31 (6.87)	−1.16 [–1.91, –0.41]^a	−1.27
	Post	6.29 (4.74)	13.50 (5.91)	−1.30 [–2.09, –0.51]^a	−1.41
	FU	7.27 (2.96)	12.67 (6.68)	−0.96 [–1.76, –0.14]^a	−1.07
NTA	Pre	3.56 (3.04)	4.50 (3.39)	−0.28 [–0.98, 0.41]
	Mid	0.69 (1.04)	3.75 (3.29)	−1.22 [–1.98, –0.47]^a	−0.94
	Post	0.93 (1.03)	3.69 (2.84)	−1.22 [–2.01, –0.44]^a	−0.94
	FU	1.18 (1.27)	2.20 (2.40)	−0.49 [–1.28, 0.30]	−0.21

Note. TCIT = Teacher–Child Interaction Training; TAU = treatment-as-usual; g = Hedges’s g; CI = confidence interval; Adj g = g adjusted for pre-treatment g (Durlak, 2009); DC = direct commands; IC = indirect commands; QU = questions; NTA = negative talk.

95% CI indicates statistical significance.

Figure 3.

TCIT-U and TAU teachers’ count of Avoid skills per observation at pre, mid, post, and follow-up phases.

Child Behavior and Social-Emotional Competence

Baseline to post

First, paired-sample t tests comparing baseline with post-treatment scores for the TCIT-U group (n = 20) indicated a significant increase in TPF—mean difference = 5.70 (5.93); 95% CI = [2.92, 8.48]; t = 4.30, p < .001—and a significant decrease in overall BC—mean difference = −4.45 (7.95); 95% CI = [–8.17, –0.73]; t = −2.50, p = .02. Conversely, t test comparisons of baseline to post-treatment scores for the TAU group (n = 18) showed a small yet significant increase for TPF—mean difference = −3.00 (5.79); 95% CI = [0.12, 5.88]; t = 2.20, p = .04—but no change in BC—mean difference = −1.72 (6.28); 95% CI = [–4.84, 1.40]; t = −1.16, p = .26. Second, results of rANOVA with ITT analyses for post-treatment data (N = 38) found that Group had a statistically significant effect on TPF—F(1, 34) = 4.76, p = .04—such that children in the TCIT-U group demonstrated a significantly greater increase in overall social-emotional skills by post-treatment than children in the TAU group. Group did not have a statistically significant effect on BC—F(1, 34) = 3.64, p = .07. Adjusted ES showed medium effects at post-treatment (TPF = .58 and BC = −.51). See Table 4 for results of pre–post rANOVA.

Table 4.

Pre–Post Repeated-Measure ANOVA for Child Outcomes (N = 38).

Effects	Mean squares	F(1, 34)	p
Total Protective Factors
Group (error)^a	14.12 (46.71)	0.30	.59
Time	177.50	11.61	.002
Time × Classroom	55.449	3.63	.07
Time × Time in Study	49.19	3.22	.08
Time × Group	72.69	4.76	.04
Error	519.74
Behavioral Concerns
Group (error)^a	4.37 (69.58)	0.06	.80
Time	135.46	5.56	.024
Time × Classroom	86.42	3.55	.07
Time × Time in Study	25.71	1.06	.31
Time × Group	88.69	3.64	.07
Error	24.35

Note. Time in study = number of weeks in study from pre- to post-treatment.

Test of between-subjects effect of Group and error.

Baseline to post to follow-up

For the subset of children who completed 3-month follow-up data collection (N = 11), paired-sample t tests of baseline to follow-up scores for the TCIT-U group (n = 5) showed a significant increase in TPF (t = 3.06, p = .04), suggesting that improvements in children’s overall social-emotional skills were maintained 3 months after treatment was completed. No significant change was found for the TAU group. Results from rANOVA indicated that Group had a statistically significant effect on BC across levels of Time—F(2, 14) = 10.59, p = .002. Post hoc analyses with Bonferroni adjustment showed a significant Time × Group interaction between baseline and post-treatment—F(1, 7) = 11.24, p = .01—such that children in the TCIT-U group demonstrated a decrease while children in the TAU group showed an increase in BC. Group did not have statistically significant effect on TPF—F(2, 14) = 3.18, p = .07—across Time. Adjusted ES showed large effects at post-treatment (TPF = 1.00 and BC = −.88) and follow-up (TPF = .77 and BC = −.87). However, given the small sample size at follow-up, these results of statistical tests should be interpreted with caution.

TCIT Fidelity

For the present study, fidelity item checklists guided the TCIT-U trainer through group didactic and individual training sessions. Across the eight group didactic sessions, the training achieved 80% to 100% fidelity to the checklists. At times, there was not enough time for video examples or role-play of skills. Fidelity for individual coaching checklists was 75% to 100%. During the course of coaching, teachers requested that written feedback be provided as a summary at the end of each phase rather than following each coaching session.

TAU

Teachers reported their hours and type of training received in Conscious Discipline and Al’s Pals curriculum at baseline. All of the teachers had received training in Conscious Discipline and five of the teachers received training in Al’s Pals. TCIT-U and TAU teachers completed similar number of hours of training for Conscious Discipline (TCIT-U: M = 11.25, SD = 7.81; TAU: M = 11.50, SD = 4.12) and Al’s Pals (TCIT-U: M = 8.25, SD = 7.93; TAU: M = 7.50, SD = 8.70). At the beginning of each phase (baseline, mid, post, and follow-up), each classroom reported how many lessons of Conscious Discipline and Al’s Pals they completed in the past week. There were no significant differences between TCIT-U and TAU classrooms in terms of lesson frequency. On average, teachers led children through Conscious Discipline’s calming strategies 4.33 (TCIT-U) and 4.83 (TAU) times per week and provided Al’s Pals lessons 1.00 (TCIT-U) and 1.67 times per week (TAU).

Discussion

In this implementation of TCIT-U, teachers were trained in skills to improve teacher–child relationships, modify child behavior problems, and facilitate the development of child social-emotional skills. The present project replicated the TCIT-U protocol as described in recent studies (e.g., Garbacz et al., 2014; Gershenson et al., 2010; Lyon et al., 2009) and utilized similar measures to allow for between-study comparisons. Our study aimed at contributing to the growing empirical support of TCIT-U by testing its effectiveness compared with TAU in a therapeutic preschool setting and with a sample of children at risk for behavioral and social-emotional problems due to maltreatment exposure. Overall, findings indicate teachers participating in TCIT-U increased their use of positive attending skills and decreased their use of demands and negative attending during the intervention and 3 months following the intervention period. Results also provide preliminary support, consistent with prior studies, for the effects of TCIT-U on improving child social-emotional skills.

Teacher Behavior Change

Consistent with the hypotheses, all teachers exhibited low PRIDE skills and high Avoid skills at baseline. TCIT-U teachers subsequently demonstrated the intended change in mean frequency and direction of PRIDE and Avoid skills, while TAU teachers demonstrated little to no change over time. Measuring TAU teachers’ use of TCIT-U skills also provided an assessment of treatment contamination that has been an issue in previous TCIT studies involving a control group (Tiano & McNeil, 2006).

Findings are consistent with past research on TCIT in that TCIT-U teachers demonstrated the greatest changes in the use of Labeled Praises, Behavior Descriptions, and Negative Talk. Also, teachers’ use of commands was the least likely to change, similar to past studies (e.g., Fernandez et al., 2015; Lyon et al., 2009). The current study aimed to document whether teachers’ skill acquisition translated to use in their natural teacher–child interactions by gathering observational data outside of coaching sessions. Teacher skill use in this study is consistent with other TCIT studies that coded skills during coaching sessions (e.g., Garbacz et al., 2014; McIntosh et al., 2000), suggesting TCIT-U skills can generalize to a variety of typical classroom situations. The finding that TCIT-U teachers’ questions decreased while overall Commands remained relatively stable is likely because teachers are encouraged to rephrase some questions to commands for clarity during lessons. Future research should examine other factors that could influence teacher skill acquisition, such as observation setting or situation (e.g., free play, transitions), teacher characteristics (e.g., age, years teaching, culture), and participation (e.g., homework, attendance) with a larger sample of teachers. In addition, documenting the Sit-and-Watch process and effects in more detail will promote further development of and evidence for this component of TCIT-U.

Child Behavior Change

Baseline to post-treatment results of within-group tests supported the hypotheses such that children in the TCIT-U group demonstrated significant decreases in behavior problems and increases in overall social-emotional skills, while children in the TAU group demonstrated few behavioral improvements based on teacher-report. There was also evidence of a subset of TCIT-U children maintaining improvements in social-emotional skills 3 months after training was completed. Subsequent tests of the interaction of treatment group on change over time found support for the hypothesis that TCIT-U children demonstrated greater improvement in overall social-emotional skills than TAU children from baseline to post-treatment. These findings are consistent with previous studies of TCIT-U that found improvements in social-emotional skills (Budd et al., 2016; Garbacz et al., 2014). Although there was a significant difference in overall behavior problems between TCIT-U and TAU groups in the subset of children who completed follow-up data collection, such findings should be considered with caution given the level of attrition and small sample at follow-up. Finally, the present study found medium to large effects of TCIT-U compared with TAU at post-treatment and follow-up for child outcome variables which are consistent with prior research (e.g., Fernandez et al., 2015; Garbacz et al., 2014).

Contrary to the hypotheses, TCIT-U children did not demonstrate statistically significant change in overall behavior problems from baseline to post-treatment compared with TAU children. One possible reason for these nonsignificant findings is the level of attrition in the present study. Thirty-two percent of children did not receive the whole intervention, namely, the TDI phase which is when teachers are taught behavioral strategies to address specific behavior problems. Thus, it may be that the present results are an underestimation of the treatment effects, which is often the case in ITT research (Salim, Mackinnon, Christensen, & Griffiths, 2008). Although the direction of mean differences and medium to large ES comparing treatment groups suggest patterns consistent with hypotheses, future research with a larger sample and greater study control is needed.

Limitations and Future Directions

Although this study provides several important contributions to the TCIT-U literature, it has notable limitations. First, more advanced statistical analyses to account for nested data structure, such as through multilevel modeling, were not possible given the present study’s sample size. In an attempt to account for nesting, classroom was included as a covariate in the current analyses, and missing data were partly remediated by using ITT. Future randomized trials with larger samples and more classrooms will allow for better control of treatment effects and more advanced analyses of possible mediators and moderators, thereby identifying the unique effects of TCIT-U.

Second, child outcomes data were measured through teacher report which can lead to reporter bias. Previous TCIT studies that measured child outcomes via standardized observational tools found that TCIT was related to decreases in inappropriate child behavior (Filcheck, McNeil, Greco, & Bernard, 2004; Tiano & McNeil, 2006). Future research will be strengthened by masking treatment conditions for teacher participants, gathering child outcome data from multiple informants, and incorporating other methods of measuring child outcomes, such as observational or direct assessment of behavior and social-emotional skills (Carter, Briggs-Gowan, & Davis, 2004).

Third, due to the therapeutic nature of the preschool, children could enroll and terminate services at any time; thus, attrition was substantial. The reasons for children discharging or moving classrooms included successful completion of treatment goals (which is an indication of positive progress for those children), moving home placements (and thus moving preschools), and the agency’s need to move children into older classrooms to make spaces in younger classrooms available. Interpretations of these results, especially at follow-up, should proceed with caution. Further investigations are recommended to determine the long-term effects of TCIT-U on teacher and child behavior. Such information will advance understanding of the implementation effectiveness and sustainability of TCIT-U.

Finally, information about children’s complex trauma history was limited to aggregate analyses of medical records maintained by the therapeutic preschool. Thus, variables associated with maltreatment exposure (e.g., severity, frequency, etc.) could not be used to understand child outcomes of the intervention. For TCIT-U to reach a level similar to PCIT as a supported intervention for child maltreatment, extensive research is required to assess TCIT-U’s effectiveness with children exposed to maltreatment and complex trauma.

Conclusion

Despite these limitations, our study provides encouraging preliminary support for TCIT-U’s effectiveness in improving teacher and child behavior with children at risk of behavioral and social-emotional problems due to their maltreatment history. Young children who have experienced maltreatment or complex trauma need consistent responsive caregivers (parents, teachers, and/or foster parents) and access to appropriate mental health supports (Chu & Lieberman, 2010). Our intent was to evaluate the promise of TCIT-U for addressing child maltreatment or complex trauma. Growing research on TCIT has shown it to be a promising program for training teachers in strategies for creating positive classroom environments and warm, responsive relationships with children. The growing evidence for the influence of positive teacher–child relationships on improvements in child behavior reinforce the importance of teachers and classrooms as a means of providing mental health interventions to youth. Thus, TCIT-U may provide a critical avenue to help protect and promote the well-being of vulnerable children through school-based mental health intervention.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The writing of this manuscript was supported in part by the National Institute of Child Health and Human Development (NICHD) under Grant No. F31 HD088020-01A1 awarded to Lindsay Huffhines and the National Science Foundation (NSF) Graduate Research Fellowship Program under Grant No. 2016210612 awarded to Katie J. Stone. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NICHD or NSF.

References

Bailey

(2000). Conscious discipline: 7 basic skills for brain smart classroom management. Oviedo, FL: Loving Guidance.

Barton

E. E.

Wolery

(2007). Evaluation of e-mail feedback on the verbal behaviors of pre-service teachers. Journal of Early Intervention, 30, 55–72.

Budd

K. S.

Garbacz

L. L.

Carter

J. S.

(2016). Collaborating with public school partners to implement Teacher–Child Interaction Training (TCIT) as universal prevention. School Mental Health, 8, 207–221.

Buyse

Verschueren

Doumen

(2011). Preschoolers’ attachment to mother and risk for adjustment problems in kindergarten: Can teachers make a difference? Social Development, 20, 33–50.

Carter

A. S.

Briggs-Gowan

Davis

N. O.

(2004). Assessment of young children’s social-emotional development and psychopathology: Recent advances and recommendations for practice. Journal of Child Psychology and Psychiatry, 45, 109–134.

Chu

A. T.

Lieberman

A. F.

(2010). Clinical implications of traumatic stress from birth to age five. Annual Review of Clinical Psychology, 6, 469–494.

Cook

Spinazzola

Ford

Lanktree

Blaustein

Cloitre

. . . van der Kolk

(2005). Complex trauma in children and adolescents. Psychiatric Annals, 35, 390–398.

Denham

S. A.

Bassett

H. H.

Way

Mincic

Zinsser

Graling

(2012). Preschoolers’ emotion knowledge: Self-regulatory foundations, and predictions of early school success. Cognition & Emotion, 26, 667–679.

Denham

S. A.

Bassett

H. H.

Zinsser

(2012). Early childhood teachers as socializers of young children’s emotional competence. Early Childhood Education Journal, 40, 137–143.

10.

Dobbs

Arnold

D. H.

(2009). Relationship between preschool teachers’ reports of children’s behavior and their behavior toward those children. School Psychology Quarterly, 24, 95–105.

11.

Durlak

J. A.

(2009). How to select, calculate and interpret effect sizes. Journal of Pediatric Psychology, 34, 917–928.

12.

Early

D. M.

Maxwell

K. L.

Ponder

B. D.

Pan

(2017). Improving teacher-child interactions: A randomized controlled trial of making the most of classroom interactions and my teaching partner professional development models. Early Childhood Research Quarterly, 38, 57–70.

13.

Eyberg

S. M.

Boggs

Algina

(1995). Parent-child interaction therapy: A psychosocial model for the treatment of young children with conduct problem behavior and their families. Psychopharmacology Bulletin, 31, 83–91.

14.

Eyberg

S. M.

Nelson

M. M.

Boggs

S. R.

(2008). Evidence-based psychosocial treatments for children and adolescents with disruptive behavior. Journal of Clinical Child & Adolescent Psychology, 37, 215-237.

15.

Eyberg

S. M.

Nelson

M. M.

Ginn

N. C.

Bhuiyan

Boggs

S. R.

(2013). Dyadic parent-child interaction coding system: Comprehensive manual for research and training (4th ed.). Gainesville, FL: PCIT International.

16.

Fernandez

M. A.

Adelstein

J. S.

Miller

S. P.

Areizaga

M. J.

Gold

D. C.

Sanchez

A. L.

. . . Gudiño

O. G.

(2015). Teacher-Child Interaction Training: A pilot study with random assignment. Behavior Therapy, 46, 463–477.

17.

Filcheck

H. A.

McNeil

C. B.

Greco

L. A.

Bernard

R. S.

(2004). Using a whole-class token economy and coaching of teacher skills in a preschool classroom to manage disruptive behavior. Psychology in the Schools, 41, 351–361.

18.

Fox

Dunlap

Hemmeter

M. L.

Joseph

G. E.

Strain

P. S.

(2003). The teaching pyramid: A model for supporting social competence and preventing challenging behavior in young children. Young Children, 58, 48–52.

19.

Friedman-Krauss

Raver

C. C.

Morris

P. A.

Jones

S. M.

(2014). The role of classroom-level child behavior problems in predicting preschool teacher stress and classroom emotional climate. Early Education and Development, 25, 530–552.

20.

Garbacz

L. L.

Zychinski

K. E.

Feuer

R. M.

Carter

J. S.

Budd

K. S.

(2014). Effects of Teacher-Child Interaction Training on teacher ratings of behavior change. Psychology in the Schools, 51, 850–865.

21.

Geller

(1999). Al’s pals: Kids making healthy choices. Richmond, VA: Wingspan.

22.

Gershenson

R. A.

Lyon

A. R.

Budd

K. S.

(2010). Promoting positive interactions in the classroom: Adapting parent-child interaction therapy as a universal prevention program. Education & Treatment of Children, 33, 261–287.

23.

Hamre

B. K.

Pianta

R. C.

(2001). Early teacher-child relationships and the trajectory of children’s school outcomes through eighth grade. Child Development, 72, 625–638.

24.

Hedges

L. V.

Olkin

(1985). Statistical methods for meta-analysis. London, England: Academic Press.

25.

Hemmeter

M. L.

Snyder

P. A.

Fox

Algina

(2016). Evaluating the implementation of the pyramid model for promoting social-emotional competence in early childhood classrooms. Topics in Early Childhood Special Education, 36, 133–146.

26.

Hoffman

L. L.

Hutchinson

C. J.

Reiss

(2009). On improving school climate: Reducing reliance on rewards and punishment. International Journal of Whole Schooling, 5(1), 13–24.

27.

Kaiser

A. P.

Roberts

M. Y.

(2013). Parents as communication partners: An evidence-based strategy for improving parent support for language and communication in everyday settings. Perspectives on Language Learning and Education, 20, 96–111.

28.

Kalvin

C. B.

Bierman

K. L.

Gatzke-Kopp

L. M.

(2016). Emotional reactivity, behavior problems, and social adjustment at school entry in a high-risk sample. Journal of Abnormal Child Psychology, 44, 1527–1541.

29.

Kanine

R. M.

Tunno

A. M.

Jackson

O’Connor

(2015). Therapeutic day treatment for young maltreated children: A systematic literature review. Journal of Child and Adolescent Trauma, 8, 187–199.

30.

Lane

J. D.

Ledford

J. R.

Shepley

Mataras

T. K.

Ayres

K. M.

Davis

A. B.

(2016). A brief coaching intervention for teaching naturalistic strategies to parents. Journal of Early Intervention, 38, 135–150.

31.

LeBuffe

P. A.

Naglieri

J. A.

(2003). Devereux early childhood assessment clinical form (DECA-C-C): A measure of behaviors related to risk and resilience in preschool children. Lewisville, NC: Kaplan Press.

32.

Lieberman

A. F.

Chu

Van Horn

Harris

W. W.

(2011). Trauma in early childhood: Empirical evidence and clinical implications. Development and Psychopathology, 23, 397–410.

33.

Lieberman

A. F.

Van Horn

(2005). Don’t hit my mommy!: A manual for child-parent psychotherapy with young witnesses of family violence. Washington, DC: Zero to Three.

34.

Lyon

A. R.

Gershenson

R. A.

Farahmand

F. K.

Thaxter

P. J.

Behling

Budd

K. S.

(2009). Effectiveness of teacher-child interaction training (TCIT) in a preschool setting. Behavior Modification, 33, 855–884.

35.

McIntosh

D. E.

Rizza

M. G.

Bliss

(2000). Implementing empirically supported interventions: Teacher-child interaction therapy. Psychology in the Schools, 37, 453–462.

36.

Pears

K. C.

Fisher

P. A.

Bruce

Kim

H. K.

Yoerger

(2010). Early elementary school adjustment of maltreated children in foster care: The roles of inhibitory control and caregiver involvement. Child Development, 81, 1550–1564.

37.

Robinson

L. R.

Morris

A. S.

Heller

S. S.

Scheeringa

M. S.

Boris

N. W.

Smyke

A. T.

(2009). Relations between emotion regulation, parenting, and psychopathology in young maltreated children in out of home care. Journal of Child and Family Studies, 18, 421–434.

38.

Royer

D. J.

Lane

K. L.

Dunlap

K. D.

Ennis

R. P.

(2016). A systematic review of teacher-delivered behavior-specific praise on K–12 student performance. Remedial and Special Education. Advance online publication. doi:10.1177/0741932517751054

39.

Sabol

T. J.

Pianta

R. C.

(2012). Recent trends in research on teacher-child relationships. Attachment & Human Development, 14, 213–231.

40.

Salim

Mackinnon

Christensen

Griffiths

(2008). Comparison of data analysis strategies for intent-to-treat analysis in pre-test-post-test designs with substantial dropout rates. Psychiatry Research, 160, 335–345.

41.

Siekkinen

Pakarinen

Lerkkanen

Poikkeus

Salminen

Poskiparta

Nurmi

(2013). Social competence among 6-year-old children and classroom instructional support and teacher stress. Early Education and Development, 24, 877–897.

42.

Stanton-Chapman

Walker

V. L.

Voorhees

M. D.

Snell

M. E.

(2016). The evaluation of a three-tier model of positive behavior interventions and supports for preschoolers in Head Start. Remedial and Special Education, 37, 333–344.

43.

Sutherland

K. S.

Conroy

M. A.

Algina

Ladwig

Jessee

Gyure

(2018). Reducing child problem behaviors and improving teacher-child interactions and relationships: A randomized controlled trial of BEST in CLASS. Early Childhood Research Quarterly, 42, 31–43.

44.

Thomas

Zimmer-Gembeck

M. J.

(2012). Parent–child interaction therapy: An evidence-based treatment for child maltreatment. Child Maltreatment, 17, 253–266.

45.

Tiano

J. D.

McNeil

C. B.

(2006). Training Head Start teachers in behavior management using parent-child interaction therapy: A preliminary investigation. Journal of Early and Intensive Behavior Intervention, 3, 220–233.

46.

Timmer

S. G.

Urquiza

A. J.

Zebell

(2006). Challenging foster caregiver-maltreated child relationships: The effectiveness of parent-child interaction therapy. Children and Youth Services Review, 28, 1–19.

47.

Timmer

S. G.

Urquiza

A. J.

Zebell

N. M.

McGrath

J. M.

(2005). Parent-child interaction therapy: Application to maltreating parent-child dyads. Child Abuse & Neglect, 29, 825–842.

48.

Tse

(2006). Research on day treatment programs for preschoolers with disruptive behavior disorders. Psychiatric Services, 57, 477–486.

49.

U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. (2018). Child maltreatment 2016. Retrieved from https://www.acf.hhs.gov/cb/research-data-technology/statistics-research/child-maltreatment

50.

Wilson

K. R.

Hansen

D. J.

(2011). The traumatic stress response in child maltreatment and resultant neuropsychological effects. Aggression and Violent Behavior, 16, 87–97.