Behavior Modification

http://bmo.sagepub.com

Effectiveness of Teacher-Child Interaction Training (TCIT) in a

Preschool Setting
Aaron R. Lyon, Rachel A. Gershenson, Farahnaz K. Farahmand, Peter J.
Thaxter, Steven Behling and Karen S. Budd
Behav Modif 2009; 33; 855 originally published online Sep 23, 2009;
DOI: 10.1177/0145445509344215

The online version of this article can be found at:

http://bmo.sagepub.com/cgi/content/abstract/33/6/855

Published by:

http://www.sagepublications.com

Additional services and information for Behavior Modification can be found at:

Email Alerts: http://bmo.sagepub.com/cgi/alerts

Subscriptions: http://bmo.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations http://bmo.sagepub.com/cgi/content/refs/33/6/855

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

 Behavior Modification
33(6) 855–884
Effectiveness of Teacher- © 2009 SAGE Publications
DOI: 10.1177/0145445509344215
Child Interaction Training http://bmo.sagepub.com

(TCIT) in a Preschool
Setting

Aaron R. Lyon1, Rachel A.

Gershenson2, Farahnaz K.
Farahmand2, Peter J.Thaxter2,
Steven Behling2, and Karen S. Budd2

Abstract
This research addressed the need for trained child care staff to support
optimal early social-emotional development in urban, low-income, ethnic
minority children. We evaluated effectiveness of Teacher-Child Interaction
Training (TCIT), an approach adapted from Eyberg’s Parent-Child Intera-
ction Therapy (PCIT). TCIT focuses on increasing preschool teachers’
positive attention skills and consistent discipline in order to enhance
children’s psychosocial functioning and prevent mental health problems.
A total of 12 teachers participated in small-group workshop sessions
with in vivo coaching on their use of skills in the classroom. A multiple-
baseline design across four classrooms (3 teachers each) evaluated effects
of training on teacher behaviors during weekly classroom observations.
Findings indicated systematic increases in trained skills during intervention,
and consumer evaluations showed that the training was rated positively.
Our results suggest that TCIT is a promising approach for enhancing
positive teacher-child interactions in a preschool setting and should receive
further investigation.

1
University of Washington
2
DePaul University

Corresponding Author:
Karen S. Budd, DePaul University, Department of Psychology, 2219 N. Kenmore Ave, Chicago,
IL 60614
Email: [email protected]

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

856 Behavior Modiication 33(6)

Keywords
Parent–Child Interaction Therapy, community, teacher training, preschool,
disruptive behavior

Behavioral difficulties (e.g., problems with aggression, disruptiveness, and

noncompliance) beginning as early as 2 years of age are likely to continue
over time (Shaw, Lacourse, & Nagin, 2005) and are associated with such
negative outcomes as criminality and antisocial behavior in adulthood
(Vitelli, 1997). These problems are also known to adversely impact the
settings in which the children regularly interact, such as day care and pre-
school. Behavioral difficulties can damage student-teacher relationships
(Birch & Ladd, 1998) and can potentially lead to teacher burnout (Hastings
& Bham, 2003). Due to the early emergence of functional impairments and
established negative trajectory, prevention and early intervention programs
are important.
The need for effective prevention is heightened by evidence that many
children who ultimately develop psychosocial problems will not receive ade-
quate mental health care (Stephenson, 2000). This is especially true among
low-SES, ethnic minority children and families (Rawal, Romansky, Jenu-
wine, & Lyons, 2004; Snowden, 1999; U.S. Department of Health and
Human Services, 2001). School-based prevention and intervention programs
involving teachers appear promising for addressing early behavioral difficul-
ties, especially for populations less likely to seek traditional mental health
services (Atkins et al., 2006; Breitenstein et al., 2007). Whereas evidence-
based approaches involving parents have been effective with young children
displaying disruptive behavior problems (Brinkmeyer & Eyberg, 2003;
McMahon & Forehand, 2003; Webster-Stratton & Reid, 2003), most children
do not have access to these programs. Some researchers have emphasized the
need to adapt empirically supported treatments to community contexts such
as schools and test their feasibility in everyday practice (Weisz, Sandler,
Durlak, & Anton, 2005). Among the promising approaches, Webster-Stratton’s
evidence-based parent-training intervention has been adapted via mental
health consultation to train teachers and parents in the context of Head Start
(HS) programs (Raver et al., 2008; Williford & Shelton, 2008).
In a meta-analysis, Wilson, Gottfredson, and Najaka (2001) found that
school-based prevention programs were consistently effective when they
were behavioral or cognitive-behavioral in nature. In addition to their effects
on children, teacher-administered programs have the potential to improve the
climate of the classroom by promoting teacher’s mental health and reducing
burnout. Teacher use of school-based behavior management tools has been

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 857

associated with higher levels of teacher self-efficacy beliefs (Rimm-Kaufman

& Sawyer, 2004), a component of many professional burnout prevention pro-
grams (Rabinowitz, Kushnir, & Ribak, 1996). Thus, teacher-implemented
school-based programs have the potential to reach underserved populations
and promote classroom environments that support healthy psychosocial
functioning for both children and teachers.

Adapting PCIT to the Classroom

In the current study, we assessed effectiveness of our preventive adaptation
of an existing manualized parent-training intervention, Parent-Child Interac-
tion Therapy (PCIT), to the preschool setting. PCIT draws from attachment,
social learning, and developmental theories in the treatment of young chil-
dren’s (aged 2-7 years) externalizing behavior (Brinkmeyer & Eyberg, 2003).
The program proceeds in two successive phases, the child-directed interac-
tion (CDI) and the parent-directed interaction (PDI), each of which involves
a combination of parent didactics, live coaching and feedback, and daily
“special time” practice at home. During CDI, parents learn positive attention
and nondirective play therapy skills to establish and strengthen a warm rela-
tionship with their child (see Table 1). Simultaneously, parents are coached to
reduce questions, commands, and critical statements. In PDI, parents are
taught to incorporate positively stated effective commands and discipline/
behavioral management strategies into their interactions. PCIT has received
considerable empirical support for its efficacy in the treatment of external-
izing problems in preschool-aged children (Eyberg, Nelson, & Boggs, 2008;
Thomas & Zimmer-Gembeck, 2007), but has not typically been implemented
as a prevention strategy.
In response to the success of PCIT with parents and the practical advan-
tages of intervening in school settings, a few school-based applications of
PCIT have been developed to train teachers to implement PCIT skills in the
classroom. McIntosh, Rizza, and Bliss (2000) presented a case study of an
individual child and single teacher using an adaptation they termed “Teacher-
Child Interaction Therapy.” Behavioral observations across sessions indicated
increased teacher use of skills and decreased child disruptive behavior and
noncompliance relative to baseline. Although lacking in experimental con-
trol, this study suggested that training resulted in positive changes in the
controlled training setting, but it did not examine changes in the classroom.
In another study, Filcheck, McNeil, Greco, and Bernard (2004) compared
teacher-implemented PCIT to a whole-classroom token economy approach
to behavior management in a single HS classroom using an uncontrolled,

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

858 Behavior Modiication 33(6)

Table 1. PRIDE Skills in Parent-Child Interaction Therapy

Skill Rationale Examples
PRAISE appropriate Increases specified “Terrific counting”
behavior behavior “I’m proud of you for being
Contributes to warm polite”
interaction
REFLECT appropriate Demonstrates listening/ Child: “I made a star”
speech understanding Adult: “Yes, you made a
Reinforces speech star”
IMITATE appropriate Gives positive attention to Child: (building a tower)
behavior/play good behavior Adult: (begins to stack
Promotes cooperation blocks)
DESCRIBE appropriate Communicates interest “I see you drew a smiley
behavior in and reinforces child’s face”
prosocial play “We are building a house”
Holds child’s attention
Be ENTHUSIASTIC Increases warmth of Voice is playful with lots of
interaction inflection
Keeps child interested Frequent smiles

case study design. Results suggested that the token economy and PCIT
were both effective in increasing the teacher’s use of praise, but that PCIT
was more effective in reducing the teacher criticisms and child noncompli-
ance beyond the gains from the token economy. The work by Filcheck and
colleagues is encouraging but is limited due to the lack of experimental con-
trol. Furthermore, observations of teacher behavior occurred only during a
single activity (circle time), which leaves unclear whether improvements
generalized to less structured activities (e.g., free play).
Tiano and McNeil (2006) compared implementation of PCIT in four rural
HS classrooms to three no-treatment control classrooms. Training sessions
were followed by live coaching of teachers in the classrooms. Results
showed that teachers in the intervention group used more labeled praise than
controls after treatment but evidenced no differences in unlabeled praise or
criticism. Experimental control was weakened by the small number of class-
rooms, data collection only on primary teachers, and the limited number of
observations. As acknowledged by the authors, the results from the pro-
gram’s implementation with a rural, White, HS population “may not apply
to urban schools, various ethnicities, or high-income populations” (Tiano &
McNeil, 2006, p. 229).
The current project expands prior adaptations of PCIT to the classroom in
a number of ways. First, our program focuses on a predominantly ethnic

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 859

minority, urban sample of children and an ethnically diverse group of teach-

ers. Second, we increased the number of training sessions and PCIT skills
emphasized. Third, we incorporated multiple observations across a variety of
everyday classroom situations to track changes in teachers’ use of skills over
time. Fourth, we employed a two-stage multiple baseline design across class-
rooms to experimentally evaluate the effects of intervention on teacher
behavior. Finally, our program included a consultative collaboration compo-
nent, which has been identified as an important element for engaging
participants in underserved communities (Washington, 2004). By adapting
an empirically based behavioral parent training program with an in vivo
coaching component for use with teachers, we tested an innovative approach
to early school-based intervention.

Teacher-Child Interaction Training

We termed our adaptation of PCIT to the preschool context Teacher-Child
Interaction Training (TCIT) rather than “Teacher-Child Interaction Ther-
apy,” which has been used previously (McIntosh et al., 2000). This was done
to emphasize that the target children were not necessarily clinically referred,
and that the teachers were not mental health therapists. Our goal was to
develop a program that improved functioning and positive relationships in
the classroom setting for all children. The separate phases of CDI and PDI
were maintained, although PDI was renamed teacher-directed interaction
(TDI). TCIT incorporates the core elements of PCIT while making adapta-
tions in order to enhance its appropriateness for the preschool setting and use
by classroom teachers. Table 2 lists the key modifications and the rationale
for each. More detail on these modifications and the process for developing
the adapted model is provided in the method section and in Gershenson,
Lyon, and Budd (2009).
The current research was designed to contribute to the small research base
of PCIT applications in preschool classrooms by pilot testing the TCIT model
in a community setting serving at-risk urban children. Using a two-stage
multiple-baseline design across four classroom teams of teachers, we evalu-
ated the effects of intervention in changing teacher behaviors during
interactions with children in the classroom. We hypothesized that teachers
would demonstrate systematic increases in use of positive attention skills and
decreases in criticisms following training in CDI. Furthermore, we hypothe-
sized that teachers would implement a structured discipline procedure
following training in TDI. As a secondary question, we examined the accept-
ability of TCIT to teachers.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

860 Behavior Modiication 33(6)

Table 2. TCIT Adaptations From PCIT

PCIT TCIT
Single parent/family Group training
PCIT typically occurs with In order to have an efficient impact, TCIT
individual families, although uses a group training format with two
group applications have been classrooms of three teachers each per
reported.a group.
Single child Multiple children
Parents participate in PCIT with Since teachers rarely interact with only
a single child for the duration one student, teachers begin practicing
of the program. skills with individual children and quickly
progress to applying them with groups of
students.
Standardized setting Natural setting
PCIT therapy uses a TCIT is designed to minimize classroom
standardized set of toys and disruption and maximize generalizability.
instructions, during which Teachers are observed and coached in
parent behavior is observed their typical classroom settings while
and coached. Generalization interacting with students.
beyond the therapy room
becomes a goal in the second
phase (PDI).
Full range of skills Skill subset
In PCIT, parents learn a TCIT focuses on a subset of PCIT skills
range of CDI and PDI skills most applicable to the classroom. In CDI,
appropriate to parent-child teachers are taught all the PRIDE skills and
interactions. to avoid criticisms; however, teachers are
encouraged to reduce but not eliminate
commands and questions, due to their
appropriateness in teaching. In TDI, they
are taught to use effective commands and a
range of methods for encouraging follow-
through with commands (e.g., prompts,
natural consequences, differential social
attention), plus a modified form of timeout
called Sit and Watch (see Table 4 and
discussion of training session content).
Data-driven Time-limited
Parents continue in PCIT until Given differing teacher skills and progress
they reach mastery criteria rates, performance-based completion
on their use of skills during criteria are impractical. Instead, TCIT
a standardized observation consists of eight didactic training sessions
period. and in-room coaching.

(continued)

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 861

Table 2. (continued)
PCIT TCIT
Electronic “bug” In-room coaching
PCIT therapists coach parents Most classrooms lack observation areas,
from behind a one-way and teachers move around the entire
mirror using an electronic classroom frequently. Therefore, TCIT
radio device. trainers shadow teachers in their
classrooms and provide feedback as they
engage students.
Verbal feedback Verbal and written feedback
Parents are provided with To accommodate the multiple demands
verbal feedback following of preschool classrooms, teachers are
coaching sessions. provided with both verbal feedback and a
concrete written feedback sheet following
in-class coaching.
Standardized discipline procedure Collaborative procedure design
Consequences for misbehavior Existing classroom rules, discipline
result in the implementation procedures, and teacher expectations for
of an established protocol appropriate behavior vary considerably.
involving As a result, TCIT teachers establish a
a warning, a time out chair, classroom-specific Sit and Watch plan from
and a time out backup (e.g., a general framework to create a discipline
time out room). program consistent with their classroom
needs.

a. Applications of PCIT with groups of families have been developed by McNeil, Herschell,
Gurwitch, and Clemens-Mowrer (2005) and by Niec, Hemme,Yopp, and Brestan (2005).

Method
Setting
The TCIT program was implemented at an urban, religiously affiliated day
care center in Chicago following approval by the university’s IRB. The center
was comprised of 90% low-income children, all of whom were charged a
reduced fee. A total of 70% of the low-income children were classified as
such by the state’s Department of Human Services (DHS), based on family
size, income, and parental employment status. An additional 20% did not
meet DHS criteria, but were charged a reduced fee by the center based on
their financial status.
As the teachers in each classroom were the direct recipients of the TCIT
program, ethnicity data for individual child participants were not collected.
However, the ethnicity data for the children in each classroom are likely

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

862 Behavior Modiication 33(6)

representative of the center’s ethnic composition as a whole: 73% African

American, 14% Latino, 9% non-Hispanic White, 2% Asian/Pacific Islander,
and 2% other. Group trainings for participating teachers were conducted in
the center’s conference room, and individualized coaching was conducted in
the preschool classrooms where the teachers worked.

Participants
Four preschool classrooms were selected to participate in the current study.
Classrooms were selected from the total of six preschool classrooms by the
center’s supervisory staff, with the expectation that the remaining two class-
rooms would receive training at a later date pending resources. In some cases,
selection was based on teacher interest and, in other cases, on supervisors’
perceptions of teacher training needs. Priority was given to classrooms that
were fully staffed (i.e., those with three teachers), as the time required for
participation in the intervention may not have been as feasible for teachers in
understaffed classrooms. Training fully staffed classrooms also reduced the
need to conduct training with individual teachers when new staff were hired.
The demographic characteristics of the teachers are provided in Table 3.
All teachers were female of diverse ethnicities, ages, and levels of experi-
ence. Although not all teachers were certified by the state, the term teacher
was used by the center for all classroom service providers in lieu of a more
general term such as caregiver. At the request of the teachers, continuing
education credits were provided. For some teachers, these credits may have
served as an incentive to participate.
A total of 78 children between the ages of 3 and 5 years old, with a distri-
bution of 19-21 students per classroom, received the intervention. Parents
were informed of the program through a letter explaining the classroom-level
intervention and stating that teachers would complete periodic behavioral
ratings on the children, which would be de-identified subsequently for
research purposes. Parents were informed that they could notify the school to
request that their child be excluded from the behavioral ratings if they chose,
but none did so.

Measures
Teacher TCIT Skill Use. Observations of each classroom were conducted
one to two times per week across a span of 4 months (excluding follow-up).
Classrooms A and B were observed during the first 4 months, and classrooms
C and D were observed during the second 4 months. A team of

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 863

Table 3. Teacher Demographic Data

Sample Characteristic Number

Gender
Female 12
Male 0
Ethnicity
African American 6
Non-Hispanic White 4
Latina 2
Age
Mean 36.7
Range 23-61
Years of Experience
Mean 8.1
Range 0.5-24
Education
Master’s with teaching license 1
Bachelor’s with teaching license 2
Bachelor’s degree 1
Associate’s degree 4
Some college 4

graduate students in psychology and the doctoral level principal investigator

conducted these observations using the Dyadic Parent-Child Interaction
Coding System (DPICS-III; Eyberg, Nelson, Duke, & Boggs, 2005), which
was modified to focus only on a subset of behaviors (see below; Modified
DPICS coding manual is available from the last author upon request). In
addition, the observation format was changed from tallies of observed behav-
iors to a 10-second interval system in which behaviors were coded as present
or absent after observing for the entire interval. Observers were trained in the
DPICS-III through weekly didactic meetings that consisted of reviewing the
DPICS-III manual in detail, practicing coding from mock PCIT session vid-
eotapes, and completing homework assignments and review quizzes from the
DPICS-III manual. As coding progressed over the course of the study, weekly
meetings also provided a forum for continued training and review of obser-
vational coding issues in order to minimize observer drift.
There was a total of 115 observations conducted across the four class-
rooms (three of the classrooms were observed 28 times and one of the
classrooms was observed 31 times). Each classroom’s observation took
approximately 30 minutes and usually began at 9:00 a.m., 11:00 a.m., or
3:00 p.m. depending on the class and observer schedule. In order to capture

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

864 Behavior Modiication 33(6)

within-classroom variation, teachers were typically observed during a variety

of activities, including circle time, lessons, free play, and transition periods,
although classroom schedules varied somewhat across days.
Upon entering the classroom, the observer began coding behaviors for one
teacher (the order was determined in advance to balance order across days).
Individual teachers were observed for 2 minutes each, and the presence or
absence of each category of teacher behavior was coded at the end of each
10-second interval. This process was repeated by rotating observations from
teacher to teacher. If three teachers were present, each teacher was coded for
8 minutes; if fewer teachers were present, each teacher was coded for
10 minutes. Given that interval time sampling is a common procedure in
behavioral observations, 10-second intervals were used as a feasible approxi-
mation to real-time events. The rotation across teachers was used to ensure
proper sampling of different individuals in the group setting. Data across all
observed teachers were combined to create a mean percent of skill use over
total intervals observed in the classroom for each observation day.
Categories of recorded teacher behavior included Behavioral Descriptions
(BD; e.g., “You’re coloring a picture.”), Reflections (RF; e.g., a child states,
“I’m going to color!” and the teacher responds, “You’re going to color!”),
Labeled Praise (LP; e.g., “Thank you for coloring on the paper!”), Unlabeled
Praise (UP; e.g., “Great!”), and Criticism (e.g., “You’re not very good at
coloring within the lines.”). Observers also recorded instances of Only Talk
(OT) for teacher verbalizations toward children that did not fit in the catego-
ries described above, as well as use of the Sit and Watch technique (described
below). For each instance of Sit and Watch, observers recorded the length,
reason for, and content (i.e., correct or incorrect use) of the technique.
Because Sit and Watch often continued for several 10-second intervals,
observers recorded these details immediately below the intervals on the
observation coding sheet.
Reliability coding between a pair of observers who independently
observed the same teachers occurred on 23 of the 115 total observations
(20%) interspersed across classrooms and experimental conditions. Result-
ing overall kappas (k) and the range between specific pairs of observers (in
parentheses) were as follows: BD = .60 (.48-1.00); RF = .72 (.46-.79); LP =
.78 (.55-1.00); UP = .84 (.77-.87) and Criticism = .74 (.49-1.00). Cohen’s
kappa (1960) statistic is widely used and recommended for assessing interob-
server agreement because it calculates the proportion of agreement over and
above what would be expected by chance (Hux, Sanger, Reid, & Maschka,
1997; Kratochwill & Wetzel, 1977; McDermott, 1988; Watkins & Pacheco,
2000). According to Landis and Koch (1977), kappa values between .41

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 865

and .60 are considered moderate, between .61 and .80 are substantial, and
above .81 are almost perfect.
TCIT Training Participation. Indices of training participation included
teacher attendance at group training sessions, number of CDI and TDI coach-
ing sessions completed per teacher, and amount of daily homework practice
completed per teacher. Total attendance at training sessions included both the
group didactics and individual make-up sessions with a TCIT trainer.
Teacher Satisfaction. Teacher satisfaction with TCIT training was assessed
via ratings collected through anonymous surveys at the end of the CDI and
TDI phases. Teachers provided ratings along each of six dimensions on a
5-point (0-4) scale, ranging from strongly disagree (0) to strongly agree (4).
Dimensions measured included the extent to which (a) the CDI and TDI
skills were useful, (b) training helped the teachers feel more effective in their
jobs, (c) the activities utilized in CDI and TDI were helpful to solidify the
material, (d) the presenters were knowledgeable, (e) training was organized
and clear, and (f) the overall training was useful.

Procedure
Training Sequence. The training sequence was implemented twice, each
time with teachers from two different classrooms. All the teachers from the
two selected classrooms participated in small-group training sessions, for a
total of six teachers per sequence. TCIT training sessions were held once a
week for a total of nine sessions. This included four CDI sessions, four TDI
sessions, and one “graduation” session. Training sessions lasted 1.5 hours
each and were led by two or three trainers (two doctoral graduate students
and, in one sequence, a clinical psychologist). Meetings were held with
absent teachers (typically as individual sessions for 30-40 minutes) in order
to cover the necessary session content before the next training session
occurred. In addition, teachers received individualized coaching on their
skills, during which they were coached between one and three times per
week for 20 minutes. CDI coaching began after the third CDI group session
and continued for 3-4 weeks. Likewise, TDI coaching began after the third
TDI group session and continued for 3-4 weeks. Teachers’ availability for
coaching varied due to their rotating schedules, differing classroom duties
(primary or aide), as well as occasional sick days or staffing shortages during
which they were temporarily assigned to another classroom. The total train-
ing sequence lasted between 11 and 13 weeks (due to scheduling availability
and vacation days), not counting major holiday breaks. Detailed, step-by-step

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

866 Behavior Modiication 33(6)

outlines of the TCIT protocol and training sessions described below can be
obtained by contacting the last author.
TCIT Protocol. The TCIT protocol was adapted from the orginal PCIT
manual in order to be time-limited, emphasize a different set of skills each
session, provide support to teachers for their hard work in the classroom, and
create an atmosphere in which teachers could support one another. Table 2
lists the skills taught in TCIT and the rationale for modifcations from PCIT.
All sessions began with an “ice-breaker” warm-up activity, in which teachers
and trainers shared answers to a question relevant to the training topics (for
example, “As a child, who was your favorite teacher and why?”). Sessions
(except the final graduation session) concluded by assigning a structured
“practice activity” for homework, which generally involved noting class-
room behavior problems and self-monitoring skill use for a set time (such as
5 minutes per classday), and inviting teacher feedback. Practice activity
forms were collected at the beginning of the following session.
In-class coaching began following the third CDI session and third TDI
session, and it continued for 1 to 2 weeks after the completion of the phase.
Each coaching session involved a period of observation followed by live
feedback and concluded with oral and/or written feedback to teachers. The
written feedback sheet was suggested by the first group of teachers and was
incorporated into the TDI phase for both groups. The first 5 minutes of coach-
ing sessions were spent observing teacher behaviors and assessing TCIT skill
competence, the next 10 minutes were spent providing live coaching, and the
final few minutes were spent in feedback. Coaching occurred during increas-
ingly complex situations, beginning with a small number of children and
more controlled activities (e.g., drawing at a table) and eventually including
more children and a wider variety of activities (e.g., free play, circle time,
clean-up). During most sessions, teachers were encouraged to select target
behaviors (e.g., Behavioral Descriptions) to be the focus of coaching, and
trainers provided extra prompts and encouragement surrounding the identi-
fied skill(s). In addition to these target behaviors, trainers also provided
feedback on teachers’ use of all TCIT skills. Unlike coaching in PCIT, in
which feedback is provided following nearly every parent verbalization,
TCIT coaching sessions involved greater variability in the amount of real-
time feedback provided. Depending on the nature of the activity, continuous
verbal feedback was sometimes contraindicated in the classroom setting. For
example, when teachers were engaged with the entire class (e.g., during free
play or transition), they frequently moved around the classroom, dividing
their attention across a number of children and activities. In other instances
(e.g., circle or story time), coaching sometimes interfered with the teacher’s

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 867

primary task (e.g., reading to students). In those situations, coaches provided

less immediate feedback and relied more heavily on summary feedback fol-
lowing the coaching period.
Child-Directed Interaction (CDI). The first session of the CDI phase began
with an introduction to TCIT and the distribution of TCIT binders and train-
ing handouts for review during the sessions. (Copies of the training materials
are available from the last author.) Teachers were asked to discuss the behav-
ioral challenges they faced in their classrooms. Trainers responded with
active listening and supportive statements to teachers surrounding their exist-
ing use of positive skills. Following an initial rationale for CDI (i.e.,
strengthening relationships and providing a foundation for effective disci-
pline), the specific CDI skills (i.e., PRIDE skills and verbalizations to reduce
or avoid) were introduced and modeled by trainers. Teachers were then asked
to identify classroom activities that were suitable for PRIDE skill use and, as
a practice activity, record opportunities for skill use each day on a structured
homework sheet.
The second, third, and fourth CDI training sessions each focused on a
different set of CDI skills (one PRIDE skill and one additional skill):
(a) Behavioral descriptions and ignoring unwanted behavior, (b) Reflec-
tions and reducing unnecessary or rapid-fire questions, (c) Labeled praise
and reducing criticism. In all sessions, the initial warm-up activity was
followed with a discussion of the prior week’s practice activity and a review
of the specific CDI skills to focus on that week. This review was accompa-
nied by handouts describing the specific PRIDE skill, providing a rationale,
and giving specific examples. Some handouts were adapted from the PCIT
training manual, and others were based on materials from previous pro-
grams derived from PCIT principles for use in preschool classrooms
(McNeil, 1999). Each handout also included a short written activity in which
teachers identified correct examples of each skill, and answers were shared
as a group. All CDI sessions also involved practice periods in which teach-
ers role-played the use of the CDI skills in small groups. Typically, two
teachers role-played, trainers coached, and the remaining teacher observed
and recorded CDI skill use. In-class coaching at prearranged times began
following the third CDI training session.
Teacher-Directed Interaction (TDI). The sequence of TDI training proceeded
similarly to CDI. The first TDI session began with a brief review of the
CDI skills and teachers’ experiences using them. This was done in order to
reinforce their application in the classroom and facilitate the transition into
TDI skills, the use of which requires a firm foundation in CDI. The specific
rationale, goals, and skills in TDI were introduced in the first TDI session.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

868 Behavior Modiication 33(6)

These skills included using differential social attention to strategically “praise

the opposite” of challenging behaviors, giving effective (i.e., direct, posi-
tively stated, specific) commands, following through after giving commands,
and use of the Sit and Watch classroom discipline procedure (described
below). The practice activity for this session involved teacher monitoring and
recording of serious, challenging behavior to serve as the basis for later Sit
and Watch planning.
The second TDI session built on the positive attention skills of CDI by
focusing on giving effective (i.e., direct, positively stated, specific, etc.)
commands and initial planning of Sit and Watch. The third TDI session
focused on following through on commands, the completion of Sit and
Watch planning for each classroom, and role-playing the introduction of
Sit and Watch procedures to the children. All between-classroom variations
in TCIT procedures were established during the Sit and Watch planning
process in response to teacher preferences (see below). There were no
within-classroom differences in procedures. The fourth TDI session involved
a review of Sit and Watch implementation, problem-solving difficult situa-
tions, and role-play practice of all CDI and TDI skills. Finally, a concluding
graduation session was held in which teachers’ impressions of the training
program and its effectiveness were reviewed, teachers were congratulated
for individual strengths and successes, and continuing education completion
certifications were awarded.
Sit and Watch. Our Sit and Watch framework was adapted from the
contingent observation techniques originally described by Porterfield,
Herbert-Jackson, and Risley (1976). Porterfield and colleagues found that
briefly placing disruptive children (aged 1-3 years) on the periphery of a
group activity following misbehavior was superior to corrective instruction
and redirection. Formal Sit and Watch planning began in the second TDI ses-
sion and involved the use of a structured planning sheet. The sheet allowed
the teachers of each individual classroom to reach consensus surrounding
implementation of Sit and Watch in their room (see Table 4 for an outline of
the basic procedure components and examples of classroom variations).
TCIT trainers provided support to make sure that the procedure was consis-
tent with behavioral principles. Engaging teachers as collaborative partners
in determining the specifics of Sit and Watch was believed to be an essential
element in creating an acceptable and useful intervention, as well as facilitat-
ing teacher buy-in. As a result, the Sit and Watch procedure looked slightly
different in each participating classroom. For example, some groups of teach-
ers decided to provide a Sit and Watch warning for certain behaviors prior to
implementing the procedure, and the location or length of Sit and Watch

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 869

Table 4. Sit and Watch Planning in TCIT

Sit and Watch Element Examples
Behaviors for which Sit and Fighting—hitting, kicking, biting, pushing, or
Watch will be used. pinching another child or the teacher
Disrupting group time by yellinga
Rude talka
Rude behavior
Throwing objectsa
Defiancea
Running away into bathroom/hall
Brief statement at beginning of “Because you threw a block, you have to sit and
Sit and Watch of what the watch how the other children play.”
child did. “Because you hit your classmate, you have to sit
in the Sit and Watch chair.”
Location for child to be seated Approximately 5 feet outside the activity area,
for Sit and Watch. facing the activity
Near the middle of the room (at the division of
the children’s desks and the group/circle area),
facing the activity
Time length for Sit and Watch One or two minutes in chair, with five seconds of
and requirement to end. quiet at end
Procedure for handling if a. Return child to the chair (“Stay here until I tell
child gets out of chair or you Sit and Watch is over”) and restart time.
misbehaves during Sit and b. If child gets up more than two times move
Watch. chair to a quiet corner of room.
c. Extend time by one to two minutes if needed.
d. Child loses sticker from existing classroom
rewards chart.
e. Remove child from room. Later, have child
return to Sit and Watch chair and sit for one
minute.
Brief statement at end of Sit “You can join the group now”
and Watch. “You may come back to the activity now”
Teacher attention when child Labeled praise of child’s appropriate behavior
returns to activity and begins
to behave appropriately.
a. Behaviors for which a warning was provided prior to Sit and Watch implementation.

varied across classrooms. During our planning process, teachers identified

specific behaviors for which Sit and Watch would be used (e.g., hitting
others) and also specified a statement to be used at the beginning of the pro-
cedure that would label the rule-breaking behavior. Decisions about identified
behaviors were made by teacher consensus and sometimes represented

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

870 Behavior Modiication 33(6)

common misbehaviors in a particular classroom (e.g., children saying “no” in

response to a command) or behaviors that certain teachers found especially
stressful, potentially increasing their likelihood of burnout. Teachers selected
a location for the Sit and Watch chair and the length of time a child needed to
remain in Sit and Watch (1 minute was suggested, based on Porterfield et al.,
but individual classroom teams made the choice). Procedures for dealing
with children who misbehaved during the procedure were also determined
and often included repeated returns to the chair, starting over, extended time,
or a combination. When necessary, additional scripts were developed in order
to clearly inform children of their response to misbehavior (e.g., “Because
you got up before Sit and Watch was done, you have to start over.”). Finally,
each classroom decided on a brief statement to be made at the end of Sit and
Watch and a type of positive attention for appropriate behavior following the
child’s return to the activity.
Following the third TDI session, Sit and Watch was introduced to class-
rooms at a time the teachers identified as convenient. Each classroom selected
one teacher to introduce the protocol to classroom children. Some teachers
chose to use a large stuffed animal (“Mr. Bear”) or puppets to model the use
of Sit and Watch for child misbehavior, while others used other teachers to
role-play child misbehavior. TCIT trainers were present during the classroom
introduction to provide support as needed.
Assessment Sequence. There were four experimental phases: baseline,
CDI, TDI, and follow-up (for two classrooms). Baseline assessment occurred
for 2-4 weeks, during which the teachers were observed in their classrooms
using their typical classroom management procedures. Observational assess-
ments continued throughout the CDI phase. At the conclusion of CDI,
teachers provided satisfaction ratings of the trainers and training program.
Assessment during and following the TDI phase was similar, with ongoing
behavioral observations throughout and completion of satisfaction ratings at
the end of TDI. Four months after the completion of TDI, follow-up observa-
tions were conducted on 4 of the 6 teachers who participated in the first TCIT
sequence. Two teachers were unavailable at follow-up because one had left
the center and another had been transferred to a different classroom.
Experimental Design. We utilized a multiple baseline design to examine
effects of intervention across the four participating classrooms (Kazdin,
1982; Kratochwill, 1992). For logistical reasons, it was not possible to
observe all classrooms concurrently, so we used a two-stage multiple base-
line in which two classrooms were observed for 4 months (before and during
TCIT training) and the other two classrooms were observed for the next
4 months. This design combines elements of concurrent and nonconcurrent

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 871

multiple baseline designs (Christ, 2007; Harvey, May, & Kennedy, 2004;
Watson & Workman, 1981; Winn, Skinner, Allin, & Hawkins, 2004), in that,
although a different number of baseline data points was obtained for each
baseline classroom, simultaneous data collection occurred for only two base-
lines at a time. Nonconcurrent multiple baselines do not assess for or control
history effects, and thus they are considered less rigorous than concurrent
multiple baseline designs. However, as Harvey and colleagues (2004) noted,
“nonconcurrent designs have a degree of flexibility that may allow for their
use in studying complex social contexts, such as educational settings, that
might otherwise go unanalyzed” (p. 267). Furthermore, the “principle of suc-
cessive coincidences,” which states it is highly unlikely that extraneous
variables repeatedly coincided with and caused intervention effects across
staggered intervention applications, helps to rule out history effects via
repeated demonstration of effects across baselines (Winn et al., 2004). Some
authors (Christ, 2007; Watson & Workman, 1981) have recommended
random, preplanned assignment of baseline data points to specific baselines
and/or graphing using a continuous x-axis when nonconcurrent multiple
baseline designs are used. Others (Harvey et al., 2004; Winn et al., 2004)
have disagreed or advocated greater flexibility in application of nonconcur-
rent designs in educational research. The authors believe the design
modifications employed in this study were merited in order to provide a pre-
liminary test of an innovative, adapted intervention in a community context.
Primary assessments of the effects of the intervention on teacher behavior
were conducted at the classroom level. This was done for a variety of rea-
sons. First, the Sit and Watch procedures were tailored by individual
classroom teams, and implementation could involve more than one teacher,
making classroom-level analyses most meaningful. Second, teachers within
a single classroom had different professional roles (e.g., lead teacher, aide),
and the amount and type of attention provided to students varied as a result.
Third, variability in teachers’ work schedules and presence during observer
visits led to the greater stability of classroom-level data. Because we con-
ducted observations across multiple activities (e.g., circle time, free play,
transition) and rotated across teachers each observation, the samples were
necessarily open to fluctuation.

Results
Teacher behavior change over time was assessed by examining mean, trend,
level, and latency (i.e., points prior to shift in trend or level) across phases for
each classroom (Kazdin, 1982). As measured by in-class observations,

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

872 Behavior Modiication 33(6)

Classroom A
50% Baseline CDI TDI Follow-
40% Up
30%
20%
10%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Days
Classroom B
50% Baseline CDI TDI Follow-
40% Up
30%
20%
10%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Days
Classroom C
50% Baseline CDI TDI
40%
30%
20%
10%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Days

Classroom D
50% TDI
Baseline CDI
40%
30%
20%
10%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Days

Figure 1. Session-By-Session Levels of Positive Teacher Behavior (i.e., Praise,

Reflection, Behavioral Description) Across Classroom Teachers, Graphed as the
Mean Percentage of Intervals Over Total Intervals Observed
Note: Classrooms A and B were observed concurrently, and Classrooms C and D were
observed concurrently after intervention was completed with Classrooms A and B. BL =
Baseline; CDI = Child Directed Intervention; TDI = Teacher Directed Intervention.

teachers demonstrated low mean levels of positive attention (i.e., praise,

reflections, behavioral descriptions) in three of the four participating class-
rooms prior to intervention. Trends during this baseline phase were relatively
horizontal and stable in each participating classroom (see Figure 1). One
classroom (Classroom C) demonstrated a somewhat higher level of positive

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 873

attention initially followed by a downward trend over the course of the mul-
tiple observations, and reaching the level of the other three classrooms before
the introduction of CDI. Another classroom (Classroom D) had a stable or
slight increasing trend during baseline.
Between baseline and CDI, mean positive behaviors increased across each
classroom, with an overall mean change from 9% to 19%. Changes in trend,
in which more horizontal slopes at baseline were replaced by accelerating
trends, were observed for all participating classrooms. Changes in level from
baseline to CDI varied across classrooms, with two classrooms (B and D)
exhibiting little or no immediate discontinuity between phases and the other
two classrooms exhibiting a more marked shift. For the classrooms that did
not show an immediate level shift, a latency period occurred prior to the
observed behavior change.
Teacher positive behavior changes from CDI to TDI were less pronounced,
with overall mean levels of positive behavior remaining essentially unchanged
(increase from 19% to 20%). Trend shifts between CDI and TDI phases were
apparent, and the majority of classrooms demonstrated deceleration of the
positive CDI trend during TDI. A shift in level was only apparent between
CDI and TDI phases for one classroom (Classroom C), but sesssion-to-
session variability clouds interpretation of the effect. Latency effects were
not generally observed in teacher positive behavior during TDI.
For the four teachers (two per classroom) for whom follow-up data were
available, some of the positive behavioral changes had attenuated 4 months
following the completion of TCIT. A trend shift was apparent and revealed a
negative slope during follow-up, indicating decreasing use of positive
behaviors. Changes in level and latency between TDI and follow-up were not
assessed because of the 4-month gap between phases. Among individual
follow-up teachers, those who demonstrated positive change relative to base-
line during TCIT (3 of 4 teachers) also maintained those gains at follow-up.
On average, the 4 follow-up teachers demonstrated positive attention during
12% of their verbalizations during TDI observations and during 13% of
verbalizations at follow-up. Both of the teachers who were unavailable for
follow-up had also demonstrated improvement over the course of TCIT, and
one showed some of the greatest positive changes from baseline to CDI. The
loss of these two teachers to follow-up likely affected the measured behavior
change at that time point.
As for specific TCIT skills, the largest mean behavioral gains across
classrooms were observed in the use of unlabeled praise, which increased
from an overall mean of 5% at baseline to 9% at CDI and TDI. Across
phases, teachers increased their use of behavioral descriptions from a

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

874 Behavior Modiication 33(6)

mean of 0.5% (baseline) to 2% (CDI) and 3% (TDI), increased reflections

from 3% to 6% and 7%, and increased labeled praise from 1% to 5%
and 6%, respectively. Inspection of individual teachers’ data suggested that
10 teachers demonstrated increased use of at least one positive behavior
over the course of training, and two did not demonstrate any meaningful
changes. In addition, five teachers increased their use of multiple skills. In
general, teachers showed the most positive change in their use of reflec-
tions and unlabeled praise.
Across all teachers and classrooms, critical statements were generally
observed infrequently and were therefore not investigated using single-
subject visual analysis techniques. Critical statements were only recorded for
a mean of 7% of intervals during baseline, 4% during CDI, and 5% during
TDI. Within each classroom, the largest changes in criticisms typically
occurred between baseline and CDI.
Sit and Watch was used a total of 9 times by 7 of the 12 teachers over 58
TDI observations (15.5% of observations). In addition, three instances of
procedures similar to Sit and Watch were observed during CDI in classrooms
from the second sequence prior to the introduction of the Sit and Watch pro-
cedure with those teachers.
Participation in TCIT was evaluated in order to provide an integrity check
on the training. In all, teachers received between 4 and 9 total coaching ses-
sions across CDI and TDI phases. On average, teachers received 4.1 coaching
sessions during CDI (SD = 1.4, range = 2-7) and 2.4 coaching sessions during
TDI (SD = 1.1, range = 0-4). Homework completion varied from 0% to 72%
with an average of 37.5% (SD = 28.9). Including make-up, all teachers
attended 100% of CDI training sessions. Attendance during TDI was slightly
lower, averaging 95.8% (SD = 9.7, range = 75%-100%). The two teachers
who did not demonstrate any notable behavioral changes over the course of
the TCIT intervention also had lower levels of training participation. These
two teachers had among the lowest homework completion percentages, 0%
and 3%. In addition, they were the only participants to have incomplete atten-
dance at TCIT training sessions when make-up sessions were included in our
counts. Anecdotally, these two teachers also experienced significant family
health problems during TCIT’s implementation and were among the older-
aged teacher participants.
Results from teacher satisfaction ratings were extremely positive,
with average ratings ranging from 3.2 to 4 (on a 4-point scale) on all items
in CDI and PDI. Due to the small sample size, no statistical tests were
conducted using satisfaction ratings. The largest difference between CDI

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 875

and TDI satisfaction ratings was observed on the item of skill utility, with
mean teacher ratings of 3.25 for CDI and 3.92 for TDI.

Discussion
Preschool teachers can play an important role in preventing behavior prob-
lems and promoting psychological health in the classroom. Among
school-based prevention methods that have been advanced, adaptations
of PCIT for the classroom setting represent a promising approach (Filcheck
et al., 2004; McIntosh et al., 2000; Tiano & McNeil, 2006). The current
research sought to extend previous studies by implementing an entire-
classroom TCIT program in an urban, ethnic minority sample using a
multiple-baseline research design. The findings provided support for our
hypothesis that teachers’ positive attention skills would increase systemati-
cally following training in CDI, although the observed changes were
moderate. Due to low baseline levels of criticism, we were not able to test
whether TCIT would reduce teachers’ use of criticism following CDI. The
results provided partial support for our hypothesis that teachers would
implement a structured discipline procedure following training in TDI;
further, they addressed our secondary research question by indicating
that teachers found the TCIT intervention acceptable. Overall, our TCIT
program yielded a number of positive outcomes.
Findings indicated that TCIT was effective in bringing about meaningful
teacher behavior change across classrooms. Nevertheless, these changes only
represented small to moderate effects. Specifically, teachers’ use of positive
behaviors increased systematically, although modestly, from baseline to the
completion of CDI and then remained consistently higher than baseline
through TDI. The immediate level change observed between baseline and
CDI for classrooms A and C came after the introduction of all of the PRIDE
skills during the first CDI session and the simultaneous initiation of self-
monitoring as a homework practice activity. In addition, for three classrooms,
positive behavior seems to have peaked or nearly peaked close to the end of
the CDI phase, suggesting that both the group didactic training and the
individualized, in-class coaching that followed it were instrumental in
facilitating uptake. Variability among classrooms was observed in the degree
to which behavior gains during CDI appeared to diminish (classroom A),
remain stable (classroom B), or improve further (classrooms C and D) during
TDI. Some of this variability might have resulted from TDI’s shifting empha-
sis away from promoting positive teacher behavior toward the management
of inappropriate child behavior (via Sit and Watch, etc). Although some

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

876 Behavior Modiication 33(6)

teachers might have used TDI as a continued opportunity to hone their CDI
skills in addition to implementing structured discipline, others could have
focused their attention more exclusively on discipline. Furthermore, the neg-
ative trend observed during follow up observation, although partially
attributable to the loss of two teachers at follow-up who had shown positive
behavior changes during intervention, suggests that additional support (e.g.,
booster sessions) may be necessary in order to sustain long-term behavioral
change for a subset of participating teachers.Although substantial variation
was discovered in the degree of change and the specific skills each teacher
improved, praise was increased most consistently across teachers. This was
likely due to the fact that, unlike the other TCIT skills (e.g., behavioral
descriptions), teachers are typically familiar with the concept of praise. As a
result, the participating teachers might have been more comfortable (and,
consequently, efficient) increasing their use of praise statements. Neverthe-
less, our results differ from McIntosh et al.’s (2000) single-teacher application
in which behavioral descriptions showed the most change. When compared
to behavioral descriptions, praises can be more easily delivered to larger
groups of children. This difference might account for the discrepancy
between our findings and those of McIntosh and colleagues.
With regard to praise and criticism, changes in the percentage of intervals
observed were lower than previous research utilizing similar methodology.
Tiano and McNeil (2006) demonstrated increases in labeled praise from
1.3% to 19.5% following their intervention, a considerably larger change
than in our study (1% to 6%). Our findings were more consistent with these
authors regarding changes in critical statements. Tiano and McNeil described
a change from 5.5% to 1.1% while our teachers decreased their criticism
from 7% to 5%. Changes in unlabeled praise were most equivalent between
studies (7.1% to 11.2% for Tiano and McNeil; 5% to 9% in our study). It is
noteworthy that the observation period utilized by Tiano and McNeil was
significantly longer than each of our observations (40 minutes vs. 8 minutes
per teacher), only occurred during structured activities, included only the
primary teachers, and represented a single time point immediately following
the completion of training. In contrast, our data represent an average of all
observations within a training phase and likely captured greater natural vari-
ability across time and activities, as well as the intermediate behavior changes
that occurred during each phase. It is also possible that the teachers in Tiano
and McNeil’s study maximized their skill use during the single post-treatment
observation period, when they were conscious of the observers. Due to the
frequency with which our observations occurred, performance demands were
likely reduced, resulting in more generalizable data.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 877

Our observed decrease in teachers’ use of critical statements was small,

and findings regarding critical statements should be interpreted cautiously
due to the low base rates of the behaviors. The greatest change occurred from
baseline to CDI, when teachers were specifically focused on providing posi-
tive attention, but still yielded a decrease of 3% in mean intervals of critical
statements. Not surprisingly, teachers’ use of critical statements increased
slightly during TDI. This might have been due to TDI’s specific emphasis on
discipline and responses to misbehavior.
Interestingly, some teachers from both of the classrooms participating in
the second TCIT sequence began using Sit and Watch before it was officially
introduced. This presumably occurred after teachers observed the first cohort
of teachers successfully using the technique in neighboring classrooms.
Although somewhat threatening to the internal validity of the study, this
phenomenon underscored the teachers’ desire to find useful strategies for
behavioral management, which is a reality when working with teachers in
school settings. It also suggests that simultaneous application of TCIT in a
greater number of classrooms or implementing TCIT school-wide might be
warranted.
Wide variability in TCIT participation across teachers indicated differing
levels of teacher engagement. Indeed, anecdotal reports suggest that, for
some teachers, external life circumstances (e.g., family health problems)
interfered with engagement in didactics and coaching. We have encountered
comparable barriers in the provision of PCIT services to a similar population
of children (i.e., low-SES, urban, ethnic minority; Lyon & Budd, 2009). Not
surprisingly, teachers who received lower levels of participation also demon-
strated lower levels of behavioral change, a finding consistent with data
indicating less effectiveness for school-based interventions with incomplete
implementation (Wilson, Lipsey, & Derzon, 2003). This finding simultane-
ously provides support for the effectiveness of TCIT in bringing about the
desired behavior changes (when participation is high) and underscores the
limitations of the model in impacting a diverse range of teachers.
Despite wide variety in teacher participation, overall satisfaction ratings
indicated uniformly high levels of teacher acceptance of the TCIT program.
Satisfaction was high in both CDI and TDI phases, a finding consistent with
Filcheck et al. (2004). Nevertheless, teachers’ higher satisfaction with TDI
skill utility suggests that teachers found discipline procedures such as Sit and
Watch more useful than positive attention alone. TDI’s focus on directly
addressing more serious misbehavior might have been particularly appealing
to teachers, especially those with limited classroom experience. The discrep-
ancy between teacher satisfaction and participation is also notable and

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

878 Behavior Modiication 33(6)

suggests that satisfaction might be a necessary, but not sufficient, component

of intervention engagement.

Limitations
The location of all classrooms in the same urban day care center limits the
degree to which the findings are generalizable to other schools. Resources
required for implementing TCIT is another potential limitation. TCIT is rela-
tively time intensive for both the teachers and the trainers, particularly
because it entails individualized coaching outside the group sessions and
requires trainers with expertise in the core skills. Some of the costs of train-
ing can be lessened by manualizing the training materials and preparing
on-site trainers or early childhood consultants to deliver the TCIT program in
community schools. Once trained, teachers potentially can serve as mentors
for other teaching staff as personnel changes occur over time. Such an
approach could have the benefit of enhancing the long-term institutional sus-
tainability of the intervention.
The multiple baseline design across classrooms with repeated observa-
tions provided experimental control in the current study, yet it does not
entirely rule out the possibility that the documented changes were affected by
factors external to TCIT. In particular, observation in the current study and
TCIT intervention focused on two of the four classrooms at a time, whereas,
ideally, all baselines (in this case, classrooms) in a multiple baseline
design begin data collection at the same point in time to control for history
effects. Although concurrent baselines provide the strongest experimental
control, nonconcurrent baselines are common in applied research, and it is
highly unlikely that important external events would repeatedly coincide
with condition changes (Harvey et al., 2004; Winn et al., 2004). A further
limitation is that this design does not separate out the effects of TCIT from
the larger consultative collaboration component. In addition, although we
included multiple measures of teacher participation, no fidelity data were
collected to monitor trainers’ adherence to the TCIT protocol. Future research
could include audio recordings of TCIT sessions in order to establish training
fidelity and its association with outcomes. Finally, our classroom selection
process, in which fully staffed rooms were identified by supervisors based on
teacher interest and training needs, might have resulted in a nonrepresentative
sample of teachers. Nevertheless, this decision was made in order to meet the
stated needs of the implementation setting and, as such, more closely reflects
the realities of school-based intervention.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 879

Future Directions
Despite above limitations, our findings suggest that TCIT can be effectively
implemented in urban preschool environments that serve low-SES, ethnic
minority children. This result extends previous school-based applications of
PCIT and also has implications for the identified gaps in mental health ser-
vices for children (Stephenson, 2000). TCIT combines group training of
teachers with individualized coaching of their use of skills in the classroom
by adapting PCIT’s core components to a preschool context. TCIT offers an
innovative, potentially less expensive alternative to other behavioral, school-
based prevention programs (Raver et al., 2008; Webster-Stratton, Reid, &
Hammond, 2001; Williford & Shelton, 2008) by focusing only on teachers
rather than combining teacher- and parent-training components. Future stud-
ies should replicate TCIT with additional groups of teachers and trainers in
additional preschools. Adequately powered studies could examine the extent
to which teacher behavioral change varies depending child-level variables
(such as age, SES, and level of behavior problems); teacher-level variables
(such as educational and teaching experience, the completion of homework,
or other indices of TCIT participation); classroom-level variables (such as
the activity, number of children involved, and time of day observations
occur); and school-level variables (such as collaborative consultation and
administrative support for the intervention). Research could also address
whether the collaborative approach to establishing each classroom’s Sit and
Watch procedures results in greater teacher perceptions of intervention own-
ership than a standardized approach. Finally, although parent training was
not offered as part of this preventive program, it would be valuable to inves-
tigate the benefits of combining TCIT for teachers with PCIT for parents of
children with clinical levels of behavior problems.
Another direction for future research is the effects of TCIT on student
behavior. Although not a primary hypothesis in this study, an underlying
assumption of TCIT is that teacher training would ultimately help to prevent
or ameliorate child behavior problems. Due to limited resources, minimal
information was available on child behavior. Furthermore, because the chil-
dren were not referred for clinical problems, the generally low level of
behavior problems prior to intervention may have precluded detection of dif-
ferences (Lyon, Budd, & Gershenson, 2009). Evaluating the effects of TCIT
on child behavior remains an important topic for future study.
Our preventive adaptation of PCIT retains the core principles and training
elements of PCIT but modifies some details for delivery with teachers
as a preventive approach for the entire classroom. Efforts to make TCIT

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

880 Behavior Modiication 33(6)

homework more relevant and engaging are appropriate, given the low com-
pletion rates by some participants. As further evidence is collected on TCIT’s
effectiveness, it will also be useful to compare its effectiveness and efficiency
to other evidence-based behavioral interventions. Finally, if future evidence
continues to support the effectiveness of the TCIT model, it will be essential
to study the feasibility of TCIT for dissemination to and sustainability in a
range of school settings, maximizing its potential benefits.

Declaration of Conflicting Interests

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

Funding
This project was supported in part by a grant from the Kraft Employee Fund of
Chicago, awarded to the last author. Sincere thanks are extended to the teachers for
participating and to the childcare administrators for endorsing and facilitating this
project.

References
Atkins, M. S., Frazier, S. L., Birman, D., Adil, J. A., Jackson, M., & Graczyk, P. A.,
et al. (2006). School-based mental health services for children living in high pov-
erty urban communities. Administration and Policy in Mental Health and Mental
Health Services Research, 33, 146-159.
Birch, S. H., & Ladd, G. W. (1998). Children’s interpersonal behaviors and the
teacher-child relationship. Developmental Psychology, 43, 934-946.
Breitenstein, S. M., Gross, D., Ordaz, I., Julion, W., Garvey, C., & Ridge, A. (2007).
Promoting mental health in early childhood programs serving families from low-
income neighborhoods. Journal of the American Psychiatric Nurses Association,
13, 313-320.
Brinkmeyer, M. Y., & Eyberg, S. M. (2003). Parent-Child Interaction Therapy for
oppositional children. In A. E. Kazdin & J. R. Weisz (Eds.), Evidence-based psy-
chotherapies for children and adolescents. New York: Guilford.
Christ, T. J. (2007). Experimental control and threats to internal validity of concur-
rent and nonconcurrent multiple baseline designs. Psychology in the Schools, 44,
451-459.
Eyberg, S. M., Nelson, M. M., & Boggs, S. R. (2008). Evidence-based psychosocial
treatments for children and adolescents with disruptive behavior. Journal of Clini-
cal Child and Adolescent Psychology, 37, 215-237.
Eyberg, S. M., Nelson, M. M., Duke, M., & Boggs, S. R. (2005). Manual for the
dyadic parent-child interaction coding system (3rd ed.). Retrieved July 1, 2006,
from http://pcit.phhp.ufl.edu/DPICSfiles/DPICS%20Draft%203.03.pdf

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 881

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.
Gershenson, R. A., Lyon, A. R., & Budd, K. S. (2009). Promoting positive interactions
in the classroom: Adapting Parent-Child Interaction Therapy as a universal
prevention program. Manuscript under review.
Harvey, M. T., May, M. E., & Kennedy, C. H. (2004). Nonconcurrent multiple base-
line designs and the evaluation of educational systems. Journal of Behavioral
Education, 13, 267-276.
Hastings, R. P., & Bham, M. S. (2003). The relationship between student behaviour
patterns and teacher burnout. School Psychology International, 24, 115-127.
Hux, K., Sanger, D., Reid, R., & Maschka, A. (1997). Discourse analysis procedures:
Reliability issues. Journal of Communication Disorders, 30, 133-150.
Kazdin, A. E. (1982). Single-case research designs: Methods for clinical and applied
settings. New York: Oxford University Press.
Kratochwill, T. R. (1992). Singe-case research design and analysis: An overview. In
T. R. Kratochwill & J. R. Levin (Eds.), Single-case research design and analysis:
New directions for psychology and education. Hillsdale, NJ: Lawrence Erlbaum.
Kratochwill, T. R., & Wetzel, R. J. (1977). Observer agreement, credibility, and judg-
ment: Some considerations in presenting observer agreement data. Journal of
Applied Behavior Analysis, 10, 133-139.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for
categorical data. Biometrics, 33, 159-174.
Lyon, A. R., & Budd, K. S. (2009). A community mental health application of Parent-
Child Interaction Therapy (PCIT). Manuscript under review.
Lyon, A. R., Budd, K. S., & Gershenson, R. A. (2009). Teacher-child interaction train-
ing: Overview, outcome, and sustainability. Emotional and Behavioral Disorders
in Youth, 9, 27-30.
McDermott, P. A. (1988). Agreement among diagnosticians or observers: Its
importance and determination. Professional School Psychology, 3, 225-240.
McIntosh, D. E., Rizza, M. G., & Bliss, L. (2000). Implementing empirically
supported interventions: Teacher-child interaction therapy. Psychology in the
Schools, 37, 453-462.
McMahon, R. J., & Forehand, R. (2003). Helping the noncompliant child: Family-
based treatment for oppositional behavior (2nd ed.). New York: Guilford.
McNeil, C. B. (1999). Understanding, preventing, and managing misbehavior in
preschool classrooms. West Virginia University.
McNeil, C. B., Herschell, A. D., Gurwitch, R. H., & Clemens-Mowrer, L. (2005).
Training foster parents in Parent-Child Interaction Therapy. Education and
Treatment of Children, 28, 182-196.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

882 Behavior Modiication 33(6)

Niec, L. N., Hemme, J., Yopp, J., & Brestan, E. (2005). Parent-child interaction therapy:
The rewards and challenges of a group format. Cognitive and Behavioral
Practice, 12, 113-125.
Porterfield, J. K., Herbert-Jackson, E., & Risley, T. R. (1976). Contingent observation:
An effective and acceptable procedure for reducing disruptive behavior of young
children in a group setting. Journal of Applied Behavior Analysis, 9, 55-64.
Rabinowitz, S., Kushnir, T., & Ribak, J. (1996). Preventing burnout: Increas-
ing professional self-efficacy in primary care nurses in a Balint group. American
Association of Occupational Health Nurses, 44, 28-32.
Raver, C. C., Jones, S. M., Li-Grining, C. P., Metzger, M., Champion, K. M., &
Sardin, L. (2008). Improving preschool classroom processes: Preliminary find-
ings from a randomized trial implemented in Head Start settings. Early Childhood
Research Quarterly, 23, 10-26.
Rawal, P., Romansky, R., Jenuwine, M., & Lyons, J. S. (2004). Racial differences
in the mental health needs and service utilization of youth in the juvenile justice
system. Journal of Behavioral Health Services & Research, 31, 242-255.
Rimm-Kaufman, S. E., & Sawyer, B. E. (2004). Primary teachers’ self-efficacy
beliefs, attitudes toward teaching, and discipline and teaching practice priorities
in relation to the “Responsive Classroom” approach. Elementary School Journal,
104, 321-341.
Shaw, D. S., Lacourse, E., & Nagin, D. S. (2005). Developmental trajectories of con-
duct problems and hyperactivity from ages 2 to 10. Journal of Child Psychology
and Psychiatry, 46, 931-942.
Snowden, L. R. (1999). African American service use for mental health problems.
Journal of Community Psychology, 27, 303-314.
Stephenson, J. (2000). Children with mental problems not getting the care they need.
Journal of the American Medical Association, 284, 2043-2044.
Thomas, R., & Zimmer-Gembeck, M. J. (2007). Behavioral outcomes of Parent-Child
Interaction Therapy and Triple P - Positive Parenting Program: A review and meta-
analysis. Journal of Abnormal Child Psychology, 35, 475-495.
Tiano, J. D., & McNeil, C. B. (2006). Training Head Start teachers in behavior man-
agement using Parent-Child Interaction Therapy: A preliminary investigation.
Journal of Early and Intensive Behavior Intervention, 3, 220-233.
U.S. Department of Health and Human Services. (2001). Mental health: Culture,
race, and ethnicity: A supplement to mental health: A report of the Surgeon
General. Rockville, MD: Author.
Vitelli, R. (1997). Prevalence of childhood conduct and attention-deficit hyperac-
tivity disorders in adult maximum-security inmates. International Journal of
Offender Therapy & Comparative Criminology, 40, 263-271.
Washington, W. (2004). Collaborative/participatory research. Journal of Health Care
for the Poor and Underserved, 15, 18-29.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

Lyon et al. 883

Watkins, M. W., & Pacheco, M. (2000). Interobserver agreement in behavioral

research: Importance and calculation. Journal of Behavioral Education, 10,
205-212.
Watson, P. J., & Workman, E. A. (1981). The non-concurrent multiple baseline across-
individuals design: An extension of the traditional multiple baseline design. Jour-
nal of Behavior Therapy and Experimental Psychiatry, 12, 257-259.
Webster-Stratton, C., & Reid, M. J. (2003). The incredible years parents, teachers,
and children training series: A multifaceted treatment approach for young children
with conduct problems. In A. E. Kazdin & J. R. Weisz (Eds.), Evidence-based
psychotherapies for children and adolescents. New York: Guilford.
Webster-Stratton, C., Reid, M., & Hammond, M. (2001). Preventing conduct prob-
lems, promoting social competence: A parent and teacher training partnership in
Head Start. Journal of Clinical Child Psychology, 30, 283-302.
Weisz, J. R., Sandler, I. N., Durlak, J. A., & Anton, B. S. (2005). Promoting and
protecting youth mental health through evidence-based prevention and treatment.
American Psychologist, 60, 628-648.
Williford, A. P., & Shelton, T. L. (2008). Using mental health consultation to decrease
disruptive behaviors in preschoolers: Adapting an empirically-supported interven-
tion. Journal of Child Psychology and Psychiatry, 49, 191-200.
Wilson, D. B., Gottfredson, D. C., & Najaka, S. S. (2001). School-based prevention
of problem behaviors: A meta-analysis. Journal of Quantitative Criminology, 17,
247-272.
Wilson, S. J., Lipsey, M. W., & Derzon, J. H. (2003). The effects of school-based
intervention programs on aggressive and disruptive behavior: A meta-analysis.
Journal of Consulting and Clinical Psychology, 71, 136-149.
Winn, B. D., Skinner, C. H., Allin, J. D., & Hawkins, J. A. (2004). Practicing school
consultants can empirically validate interventions: A description and demonstra-
tion of the non-concurrent multiple baseline design. Journal of Applied School
Psychology, 20, 109-128.

Bios
Aaron R. Lyon, PhD, is a postdoctoral fellow in the University of Washington
Department of Psychiatry and Behavioral Sciences. He is a graduate of the Clinical
Child Psychology program at DePaul University.

Rachel A. Gershenson, MA, is a PhD candidate in Clinical Child Psychology at

DePaul University.

Farahnaz K. Farahmand, MA, is a PhD candidate in Clinical Child Psychology at

DePaul University.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009

884 Behavior Modiication 33(6)

Peter J. Thaxter, MA, is a PhD candidate in Clinical Child Psychology at DePaul

University.

Steven Behling, MA, is a PhD candidate in Clinical Child Psychology at DePaul

University.

Karen S. Budd, PhD, is a professor of Clinical Child Psychology and the director of
Clinical Training in DePaul University’s Psychology Department. She directs the
Parent-Child Interaction Therapy program at DePaul’s Family and Community
Services.

Downloaded from http://bmo.sagepub.com at UNIV WASHINGTON LIBRARIES on December 7, 2009