750 words
Overview
By successfully completing this assignment, you demonstrate your proficiency in the following competency and specialized behaviors:
- Competency 4: Engage in practice-informed research and research-informed practice.
C4.SP.A: Apply leadership skills, decision making, and the use of technology to inform evidence-based research practice to develop, implement, evaluate, and communicate interventions across the specialization of advanced generalist practice settings.
Related Assignment Criterion:
1: Critique single-subject designs.
C4.SP.B: Apply leadership skills, decision making, and the use of technology to inform program evaluation to develop, implement, evaluate, and communicate interventions across the specialization of advanced generalist practice settings.
Related Assignment Criterion:
2: Apply critical thinking in verbal and written communication through the use of leadership and technology.
Assignment Description
After reading Whitfield’s 1999 article, “Validating School Social Work,” you will answer a series of questions aimed at critiquing single subject designs.
Assignment Instructions
Read Whitfield’s 1999 article, “Validating School Social Work: An Evaluation of a Cognitive-Behavioral Approach to Reduce School Violence,” and answer the following questions in an APA-formatted document:
- What was the purpose of this study?
- What was the sample?
- What dependent variables were studied? How were the dependent variables operationalized?
- Why was a single-subject design used? What are the strengths of a single-subject design?
- What were the key findings (the results of the data analysis)?
- How would you critique the findings? What are limitations of a single-subject design?
RESEARCH ON SOCIAL WORK PRACTICE Whitfield / VALIDATING SCHOOL SOCIAL WORK
Validating School Social Work:
An Evaluation of a Cognitive-Behavioral
Approach to Reduce School Violence
Gary W. Whitfield
Hopkins County Adolescent Day Treatment Program
Objective: This study evaluated the effectiveness of anger control training with conduct-
disordered male adolescents at a day treatment program. Method: A multiple baseline single-
subject design across subjects was used to assess matched pairs of students (8 experimental and
8 control students). Combined visual analyses and group comparison methods were used in the
assessment. Results: The experimental students significantly improved in their weekly self-
reports of using better anger control and experiencing more positive management and expres-
sion of anger. The experimental students also significantly improved in their use of self-control
as shown by a pretest through 6-month follow-up assessment. Conclusion: This research effort
provides some evidence that a cognitive-behavioral approach is one method of effectively reduc-
ing school violence when compared with a nonspecific counseling approach alone. Recommen-
dations for future research efforts are offered given some of the limitations of the study.
Recently, the school social work literature has been concerned with docu-
menting the effectiveness of social work services in schools. In this age of
accountability, educational reform, and budgetary constraints, there is a
growing concern that school social work may be left behind or even forgot-
ten. Thus, school social work’s relevance as a profession and its influence on
educational goals (e.g., grades, attendance, and classroom behavior) are cru-
cial concerns (Allen-Meares, Washington, & Welsh, 1996; Bailey-Dempsey,
1997; Michals, Cournoyer, & Pinner, 1979). Empirical validation in the
school social work literature regarding theoretically based interventions
linked to educational goals and interests is limited. This is especially true
concerning outcome studies for older adolescents (Bailey-Dempsey, 1997;
LeCroy & Flynn, 1982; Sabatino, Timberlake, & Hooper, 1991). Yet, school
social workers are often in an ideal situation to conduct theoretically based
experimental research. School social work settings allow some flexibility for
399
Author’s Note: Correspondence may be addressed to G. Whitfield, Hopkins County Adolescent
Day Treatment Program, 5770 Anton Road, Madisonville, KY 42431. I gratefully acknowledge
the assistance of William Nugent in completing the statistical analyses for this research.
Research on Social Work Practice, Vol. 9 No. 4, July 1999 399-426
© 1999 Sage Publications, Inc.
random assignment to experimental and control groups and to test different
treatment approaches (Alexander & Curtis, 1995).
A major interest of school social workers, educational staff, and policy
makers is reducing violence in schools (Alexander & Curtis, 1995; Allen-
Meares et al., 1996; Astor, Behre, Fravil, & Wallace, 1997). In their review,
Alexander and Curtis (1995) identified the conduct-disordered or aggressive
behaviors exhibited by students as major underlying components and causes
of school violence. Taken in this context, violent behavior in schools may
involve students evincing the following behaviors: (a) disrespect to teachers,
(b) cursing and threatening comments to staff and students, (c) criminal acts
such as theft and vandalism, and (d) assaultive behavior, either toward staff or
students. The above behaviors often denote a deficiency in self-control and
anger control skills or an unwillingness by such students to use these skills
when able to do so.
This article reports the results of a study evaluating the effectiveness of a
cognitive-behavioral intervention (i.e., anger control training) with explosive
and conduct-disordered male adolescents in reducing school violence. In
conducting this evaluation, a relatively new analytic strategy, hierarchical
linear modeling, was used as a group comparison strategy. First, a brief
review of the problems associated with conduct-disordered behavior in chil-
dren and adolescents is undertaken. A brief description of some intervention
efforts, such as day treatment services, to deal with these problems is also
presented. The theoretical underpinnings in remediating the maladapting
cognitive-behavioral products of aggressive youths are specifically high-
lighted. Finally, important research questions about the study are posited
before outlining the actual research effort and results.
LITERATURE REVIEW AND RESEARCH QUESTIONS
Conduct-disordered or aggressive behavior constitutes from one third to
one half of all child and adolescent mental health referrals. Rates as to a clini-
cal diagnosis of conduct disorder range from 4% to 10% for boys and about
2% for girls. As many as 3.8 million children show evidence for this disorder
(Hinshaw, 1992; Kazdin, 1995).
Child and adolescent antisocial or conduct-disordered behavior is also
highly stable and resistive to treatment (Kazdin 1995; Loeber, 1991; McMa-
hon, 1994). Antisocial behavior in childhood that goes untreated or is highly
resistive to treatment greatly increases the risk for problems extending to
adulthood. These problems include further criminal behavior, a diagnosis of
400 RESEARCH ON SOCIAL WORK PRACTICE
antisocial personality and other psychiatric disorders, alcoholism and drug
use, poor school adjustment and low educational attainment, marital and
family disruptions, poor occupational adjustment, and poor physical health
(Farrington, Loeber, & Van Kammen, 1990; Kazdin, 1995).
The costs to society are difficult to precisely quantify but would seem to be
exorbitant considering the expenditures involved in providing the various
services engendered by antisocial behavior including mental health treat-
ment, social services, juvenile adjudications and incarcerations, and special
education programs (Kazdin, 1995). Vandalism and destruction of school
property are estimated to cost more than $200 million annually, and the
expense of vandalism to private property would seem to be even higher
(Ziglar, Taussig, & Black, 1992). Also difficult to quantify is the personal
misery and suffering of the victims of juvenile crime and the later victims of
the juvenile delinquent as an adult perpetrator.
Efforts to provide more effective school intervention for behavior-
disordered and delinquent youths include day treatment programs. These
programs provide community-based services and are seen as alternatives to
institutionalization. The specific structure of these programs and the types of
services offered vary greatly. Very little specific outcome research has been
reported regarding school-sponsored day treatment programs (Comer, 1985;
Epstein & Maragos, 1983; Franklin, 1992).
More specifically, cognitive-behavioral training or treatment has been
recommended as a microintervention strategy to reduce school violence
and address its underlying causes (Alexander & Curtis, 1995). Cognitive-
behavioral training is an attempt to alter a youth’s maladaptive cogni-
tions and in the process change the affective component and the dysfunc-
tional overt behaviors (Finch, Nelson, & Moss, 1993). Research has
shown that aggressive and antisocial youths display serious cognitive
distortions and deficiencies. Aggressive youths present attributional
biases that distort their perceptions of social situations. They value
aggression more and are less skilled with assertive and prosocial
responses as means to solve conflict than their nonaggressive peers. Self-
management of interfering emotional arousal, appropriate use of self-
control, and initiation of problem-solving skills are particularly trouble-
some for conduct-disordered youths (for reviews, see Crick & Dodge,
1994; Finch et al., 1993; Kendall,1993).
Cognitive-behavioral training is effective in reducing aggressive behavior
in children and adolescents when compared with other interventions,
although the effects have not always been uniform or maintained in follow-up
for all samples, measured outcomes, and settings (Durlak, Fuhrman, &
Whitfield / VALIDATING SCHOOL SOCIAL WORK 401
Lampman, 1991; Kendall, 1993). Feindler and her colleagues (Feindler,
1979, 1991; Feindler & Ecton, 1986; Feindler, Marriott, & Iwata, 1984)
found evidence that specific anger control training as a cognitive-behavioral
approach reduces the disruptive behavior and improves self-control with
aggressive adolescents in various settings including residential centers, hos-
pital programs, and public schools. The treatment efficacy of anger control
training as a cognitive-behavioral intervention for aggressive children and
adolescents has been well documented in other studies as well, ranging from
residential to school settings (Dangel, Deschner, & Rasp, 1989; Glick &
Goldstein, 1987; Hains, 1989).
Thus, many challenges are present with testing a theoretically based inter-
vention and documenting school social work’s relevance in addressing a cru-
cial educational concern such as school violence. With these challenges in
mind, I conducted a study at a day treatment program serving conduct-
disordered students. Three relevant research questions helped to guide this
study: One, what is the effectiveness of a cognitive-behavioral school social
work intervention in reducing school violence (i.e., aggressive behaviors of
students)? Two, can a specific approach be more effective than a general
counseling approach in improving anger control and general self-control
among a targeted group of students? And three, can treatment gains be main-
tained at a 6-month follow-up assessment?
METHOD
Participants
Students attending an adolescent day treatment program in a public school
system were the target population. This was essentially a sample of conven-
ience and therefore the generalizability of the findings is compromised some-
what. The specific participants included in this study were the male students
that I served as the principal school social worker in the facility. I selected
participants for the study who were referred to the program because of dem-
onstrated problems with self-control. Students earn a successful comple-
tion of the program by adhering to the program’s rules, reaching certain
point totals as prescribed by individual treatment plans, and exhibiting a
general reduction of the problematic behaviors that caused the initial place-
ment. Thus, the time of exit from the program is open-ended and individu-
ally determined.
402 RESEARCH ON SOCIAL WORK PRACTICE
Research Design
As the school year progressed, 8 pairs of male participants who were part
of my caseload and had consented to participate were matched as closely as
possible on the following variables: (a) age, (b) self-control abilities as indi-
cated by total problem scores on the Self-Control Rating Scale (Kendall &
Wilcox, 1979), (c) parent- and teacher-rated aggressive and delinquent char-
acteristics as signified by externalizing problem scores on the Child Behavior
Checklist (Achenbach, 1991a) and the Teacher’s Report Form (Achenbach,
1991b), and (d) general cognitive abilities or deviation IQ as measured by the
Henmon-Nelson Tests of Mental Ability (revised) (French, Lamke, & Nel-
son, 1973). The above standardized measures are well-validated instruments
with high reliabilities. For instance, the widely used Achenbach scales have
test-retest reliabilities ranging from .89 to .92 and the correlations with other
behavior problem rating scales are significant, generally in the .80s (Achen-
bach, 1991a, 1991b; Pellegrini, Galinski, Hart, & Kendall, 1993). Additional
information on the Self-Control Rating Scale and data generated on these
measures are reported later.
I used a single-subject, multiple baseline design across subjects as the
main research design. An additional evaluation included an extended assess-
ment using the Self-Control Rating Scale measure. The research plan called
for yoked and matched pairs of appropriate students (i.e., permission
obtained and placement in the program due to problems with anger control
and aggression) to be selected for assessment and intervention. Following
baseline assessment, one of the students in each pair was randomly chosen to
begin treatment sessions. These 8 male clients comprised the main study. The
other 8 students not chosen for participation in the experimental or interven-
tion program were used as control clients for the group comparison analyses.
These comparison students continued to be assessed in the same manner as
the treatment clients. The complete selection and assessment process is
detailed in the original research (Whitfield, 1996).
Intervention Program
A 12-session Anger Control Training program as outlined by Feindler and
her colleagues was implemented as the independent variable in this study.
They have developed a cognitive-behavioral training program with five basic
components that include self-instruction, self-assessment, self-evaluation,
arousal management, and adaptive skills development.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 403
Some of the specific skills covered in the training include brief relaxation
training, deep breathing exercises, active and passive progressive muscle relaxa-
tion training, and autogenic relaxation training. Anger reducers such as back-
ward counting, pleasant imagery exercises, and self-talk cue statements are also
reviewed. Assertion techniques and role-playing are heavily emphasized as well
as the modeling of these skills and subsequent behavioral rehearsals. Additional
self-instructional training includes thought-stopping techniques and use of self-
guiding speech. Instruction in problem-solving skills and extensive discussions
of hypothetical and actual anger-provoking situations are also covered in the
training. A more detailed description of the procedures and techniques is out-
lined in a treatment manual (Feindler &Ecton, 1986).
This intervention involved 1-hour individual training sessions with each
client that occurred one to two times per week. All of the students in the study
continued to receive day treatment services as usual. This included crisis
intervention, individual counseling, group counseling, family counseling,
and other interventions as needed. The day treatment staff, including teach-
ers, were not aware of which students actually received this more focused
intervention. Thus, the participants’ progress and release from the program
followed individual treatment plans and program guidelines irrespective of
their participation and status in this study.
Outcome Measures
Two self-report measures and a behavioral count or measure of acting-out
behaviors were the dependent variables assessed weekly. Parent observations
were originally included in the study but proved to be unreliable measures
(Whitfield, 1996). The State-Trait Anger Expression Inventory (Spielberger,
1991) was completed each week by the students. This inventory consists of
44 items and several scales, but only two scales—Anger Control and Anger
Expression—were used as self-report measurements. Higher scores on the
Anger Control scale indicate more investment by the person in monitoring
and preventing the experience and expression of anger, generally something
lacking with these students. Conversely, higher scores on the Anger Expres-
sion scale show intense angry feelings that may be suppressed or expressed in
explosive behavior. Thus, lower scores on this scale show better internal
management of anger and less suppression or the acting out of anger. Anger
Control scale scores range from 8 to 32 and Anger Expression scale scores
range from 0 to 72. Median reliability (alpha coefficients) for all of the
State-Trait Anger Expression Inventory scales is about .82 with a reliability
of .84 for the Anger Control scale and .78 for the Anger Expression scale. Face
validity is reported to be good (Biskin, 1992). These two scales were used to
404 RESEARCH ON SOCIAL WORK PRACTICE
obtain a client perspective of progress and to assess whether the specific
cognitive and affective instructions of the Anger Control Training (or lack
of) would be reflected in the self-reported cognitive responses of the
participants.
The Staff Daily Report (Whitfield, 1996) was also another weekly
repeated measurement. This measurement was patterned after the Parent
Daily Report (Chamberlain & Reid, 1987), a well-established clinical tool
emanating from the Oregon Youth Study research project. Essentially, the
Staff Daily Report was a behavioral count of the aggressive episodes and the
specific instances of rules violations. In calculating these scores, the severity
of the offense was not being rated but whether a specific problem occurred or
not. The daily scores were aggregated each week to make the total problem
scores. Lower scores would, of course, indicate less behavioral and aggres-
sive problems by the students.
A more global and longer-term assessment included the Self-Control Rat-
ing Scale measure as a dependent variable. Besides being used for matching
purposes, it provided a measure in which to evaluate self-control evinced by
the students before intervention (pretest), at the end of Anger Control Train-
ing (posttest), and at 6-month follow-up assessment. Day treatment program
teachers, blind to the treatment status of the students, completed this rating
scale for the pretest and posttest assessments. For the students that returned to
a regular school placement, the regular school teachers completed the
6-month follow-up assessments. The efficacy of using different raters at
follow-up from the initial raters for standardized measures was recently dem-
onstrated in a major National Institute of Mental Health research project
(Glisson, 1994). Self-control ability and impulsivity as measured by this
scale pertain to how the student regulates his or her behavior to attain certain
goals. This regulation of behavior involves being able to generate and evalu-
ate behavioral alternatives and refraining from acting on discarded options
while engaging the selected alternative (Pellegrini et al., 1993).
The Self-Control Rating Scale is a well-validated 33-item scale with high
internal consistency (alpha = .98) and test-retest reliability estimated at .84.
Interrater reliability is reported at .66, and significant correlations with
observed frequency counts and other similar psychological indices are pres-
ent as well. Total problem scores range from 33 to 231. The higher the total
problem score, the greater the lack of self-control (Kendall & Wilcox, 1979;
Pellegrini et al., 1993).
Both the Staff Daily Report measure and the Self-Control Rating Scale
were used to assess the effects of the intervention in the natural school envi-
ronments of the participants. Further evidence of Anger Control Training’s
effects beyond the self-reported domain is needed. Generalization of the
Whitfield / VALIDATING SCHOOL SOCIAL WORK 405
intervention’s benefits to extratherapeutic contexts such as school and main-
tenance of those effects have not been readily apparent in the research litera-
ture (Feindler, 1991).
Data Analysis
Visual analysis of the weekly time-series data was conducted for the 8
Anger Control Training clients. This was aided by graphing separate mean
lines and regression (slope) lines for baseline and intervention phases for
each weekly repeated measure for each student to discern any discontinuities
in the within-subject data and to determine the magnitude and directionality
of each data set (see Bloom, Fischer, & Orme, 1995, for a description of vis-
ual analysis procedures). For the purposes of this article, only the raw obser-
vations are provided. The complete and extensive graphs with the plotted
mean and regression lines for each client are outlined in the original research
(Whitfield, 1996).
Hierarchical linear modeling (Bryk & Raudenbush, 1987, 1992) was used as
the group comparison method for this study to discern between-subject differ-
ences. Thus, data from both the 8 treatment clients and the 8 comparison clients
were used for the analyses. Hierarchical linear modeling was used to assess any
differences in theoverall intervention trendsofeachweeklymeasurement for the
two groups. A similar hierarchical linear model was used to compare the overall
trend of the Self-Control Rating Scale data forboth groups.
Based on the previously outlined research questions, four statistical
hypotheses were tested at the .05 alpha levels (two-tailed tests). Regarding
the State-Trait Anger Expression Inventory scores, it was posited that the stu-
dents receiving Anger Control Training would have higher Anger Control
scale scores and lower Anger Expression scale scores relative to the control
students’ self-reports for these two measures. It was also posited that the
Anger Control Training students would have lower Staff Daily Report total
problem scores relative to the scores of the comparison students. In addition,
Anger Control Training students were expected to show greater self-control
(lower Self-Control Rating Scale total problem scores) following Anger
Control Training relative to the comparison students’ ratings. The complete
hierarchical linear modeling procedures and statistical analyses are found in
the original research (Whitfield, 1996).
Some sample attrition for the Self-Control Rating Scale 6-month
follow-up hierarchical linear modeling analysis occurred. All of the partici-
pants had data for the posttest (end of Anger Control Training) evaluation.
However, only 9 students were in a school program (either day treatment or
406 RESEARCH ON SOCIAL WORK PRACTICE
regular school) at the 6-month follow-up assessment. These included 5
Anger Control Training students and 4 comparison students.
RESULTS
Anger Control Results
Figures 1 through 4 provide the raw weekly data for the 8 intervention cli-
ents for the Anger Control scale scores (see the original research for other
descriptive and statistical analyses for each client data set for each weekly
measure). Data for Client 1 (Figure 1) indicated a pattern of improvement
(i.e., increasingly higher Anger Control scale scores during intervention).
Data for Client 2 (Figure 1) indicated a pattern of very little change from
baseline to intervention phase.
The data for Client 3 (Figure 2) were highly variable for both the baseline
and intervention phases. A pattern of deterioration was evident with this par-
ticipant. Client 4’s data (Figure 2) indicated slight improvement from base-
line to intervention phase. Observations for Client 5 (Figure 3) indicated a
general pattern of improvement. Client 6’s data (Figure 3) reflected basically
a pattern of little change from baseline to intervention phase. A pattern of
improvement from baseline to intervention phase was present for Client 7’s
data (Figure 4). The highly variable data for Client 8 (Figure 4) contributed to
an ambiguous pattern generally.
In summary, patterns of improvement (i.e., higher Anger Control scale
scores) with this self-report data were evident with 4 Anger Control Training
clients. One experimental student displayed a negative pattern of change
from baseline to intervention phase and 3 students’ data patterns were
ambiguous or unchanged.
The hierarchical linear modeling analysis revealed that the difference
between treatment status was a significant predictor for the trend for the
Anger Control measure. With the introduction of Anger Control Training, the
intervention trend was 0.57 units per week larger than the trend associated
with control group status. This was a significant difference witht(15) = 2.44,
p = .03. Thus, the Anger Control Training students significantly improved in
their self-reports as to their use of anger control in a positive direction com-
pared with the control students. There was, however, significant parameter
variance associated with the intervention trend with the residual parameter
variance estimated at 0.15 andχ2(14,N= 16) = 55.43,p= .00. This suggested
that other variables besides Anger Control Training may have been
Whitfield / VALIDATING SCHOOL SOCIAL WORK 407
(text continues on p. 412)
408 RESEARCH ON SOCIAL WORK PRACTICE
Figure 1: Weekly Anger Control scale scores for Clients 1 and 2.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 409
Figure 2: Weekly Anger Control scale scores for Clients 3 and 4.
410 RESEARCH ON SOCIAL WORK PRACTICE
Figure 3: Weekly Anger Control scale scores for Clients 5 and 6.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 411
Figure 4: Weekly Anger Control scale scores for Clients 7 and 8.
functioning to produce different client responses to treatment. The estimate
of the proportion of explained variance was 28%.
Anger Expression Results
Figures 5 through 8 present the raw repeated measures data for the 8 par-
ticipants for the Anger Expression scale scores. Overall, a pattern of
improvement from baseline to intervention phase (i.e., decreasing Anger
Expression scale scores) was evident for Client 1 (Figure 5). Due to the
change in the levels of data, a steady pattern of improvement was also present
for Client 2 (Figure 5).
Client 3’s overall pattern (Figure 6) indicated a slight deterioration when
assessing the impact of intervention. Client 4’s data (Figure 6) were much
more variable in the intervention phase than in the baseline phase. Some
extreme scores were evident. Some improvement was present with the
change in the levels of data. The data were highly variable for both phases for
Client 5 (Figure 7). However, a slight pattern of improvement was discerned
in the overall pattern for this client. Overall, a pattern of deterioration was
present for Client 6 (Figure 7). Client 7’s general pattern of data (Figure 8)
was unchanged, although some instability in the data was observed in the
baseline phase. The data pattern for Client 8 (Figure 8) was also relatively
stable across both phases. In summary, 4 Anger Control Training students
presented favorable response patterns (i.e., decreasing Anger Expression
scale scores). Two clients displayed general patterns of deterioration. The
other 2 participants’ data reflected unchanged patterns of data with these
self-reports.
Similar to the hierarchical linear modeling analysis for the Anger Control
data, the difference between treatment status was a significant predictor for
the intervention trend for the Anger Expression weekly time-series data.
With the introduction of Anger Control Training, the intervention trend was
1.07 units per week less than the trend associated with the control group
status. This was a significant difference witht(15) = –2.19,p = .04. Thus,
Anger Control Training participants significantly improved (decreasing
direction) their Anger Expression scale scores compared with the control stu-
dents. As with the Anger Control scale scores, however, there was significant
parameter variance associated with the intervention trend with the estimated
residual parameter variance estimated at 0.71 andχ2(14,N= 16) = 70.81,p=
.00. This suggested again that other variables may have contributed to the dif-
ferent client responses in addition to group status. The estimate of the propor-
tion of explained variance was 26%.
412 RESEARCH ON SOCIAL WORK PRACTICE
(text continues on p. 417)
Whitfield / VALIDATING SCHOOL SOCIAL WORK 413
Figure 5: Weekly Anger Expression scale scores for Clients 1 and 2.
414 RESEARCH ON SOCIAL WORK PRACTICE
Figure 6: Weekly Anger Expression scale scores for Clients 3 and 4.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 415
Figure 7: Weekly Anger Expression scale scores for Clients 5 and 6.
416 RESEARCH ON SOCIAL WORK PRACTICE
Figure 8: Weekly Anger Expression scale scores for Clients 7 and 8.
Staff Daily Report Results
Figures 9 through 12 provide the raw weekly data for the 8 participants for
the Staff Daily Report total problem scores. These scores represent the
behavioral incident scores at the day treatment program. Observations for
Client 1 (Figure 9) displayed an ambiguous pattern of change. Client 2’s data
(Figure 9) indicated a highly variable pattern for the intervention phase as
there were several extreme scores. The overriding impression was a lack of
improvement for this client.
Observations for Client 3 (Figure 10) indicated very little change overall
from baseline to intervention phase. Client 4’s data (Figure 10) presented
some discontinuity between phases. Some outliers affected the general pat-
tern with some deterioration present. A stronger general pattern of deteriora-
tion was present with Client 5 (Figure 11). A general pattern of improvement
was seen with the data for Client 6 (Figure 11).
Client 7’s levels of data (Figure 12) were basically the same for both the
baseline and intervention phases. Generally, however, a pattern of moderate
improvement was present when assessing the trends of the data. Overall, a
stronger pattern of improvement was present with the data for Client 8 (Fig-
ure 12). In summary, the visual analysis of the Staff Daily Report data indi-
cated clear improvement (i.e., decreasing Staff Daily Report total problem
scores or behavioral incidents) for 3 experimental clients. General patterns of
deterioration in behavior were present for 3 Anger Control Training students.
The other data results were negligible or ambiguous for the 2 other clients in
the study.
As was the case for the two State-Trait Anger Expression Inventory
(Anger Control and Anger Expression) measures, the data from the hierar-
chical linear modeling analysis confirmed much of the within-subject visual
analysis for this dependent variable as well (see the original research for
details, Whitfield, 1996). The hierarchical linear modeling analysis revealed
that with the introduction of Anger Control Training, the intervention trend
was 0.56 units per week less than the trend associated with the control group
status for the Staff Daily Report total problem scores. This approached a sig-
nificant difference witht(15) = –2.05,p = .055. Thus, the Anger Control
Training students reduced their acting-out or aggressive episodes at the day
treatment program at a nearly significant level compared with the control par-
ticipants. There was, however, significant parameter variance associated
with the intervention trend with the estimated residual parameter variance at
0.12 andχ2(14,N= 16) = 25.72,p= .03. Thus, other variables may have been
Whitfield / VALIDATING SCHOOL SOCIAL WORK 417
(text continues on p. 422)
418 RESEARCH ON SOCIAL WORK PRACTICE
Figure 9: Weekly Staff Daily Report total problem scores for Clients 1 and 2.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 419
Figure 10: Weekly Staff Daily Report total problem scores for Clients 3 and 4.
420 RESEARCH ON SOCIAL WORK PRACTICE
Figure 11: Weekly Staff Daily Report total problem scores for Clients 5 and 6.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 421
Figure 12: Weekly Staff Daily Report total problem scores for Clients 7 and 8.
influential in differentiating the two groups besides treatment status. The
estimate of the proportion of explained variance was 40%.
Self-Control Rating Scale Results
Table 1 gives the pretest, posttest, and 6-month follow-up Self-Control
Rating Scale total problem scores data for the Anger Control Training clients
and the comparison clients. Pretest scores for both groups were essentially
the same, however, there was a 7-point difference at the posttest measurement
and a 57-point difference at the 6-month follow-up measurement with the
Anger Control Training students having the lower group averages.
Consult the original research (Whitfield, 1996) for detailed results of the
hierarchical linear modeling analysis of the trend from pretest through
6-month follow-up for the Self-Control Rating Scale data. With the introduc-
tion of Anger Control Training, the trend was 1.49 units less per assessment
interval than the trend associated with control group status. This was a sig-
nificant difference witht(14) = –2.51,p = .03. Thus, the difference between
treatment status influenced the self-control behaviors of the students with the
Anger Control Training students showing greater improvement. Also, there
was not significant variation in the trend, with the estimated residual parame-
ter variance at 0.06 andχ2(13,N= 15) = 9.21,p= .50. This indicated a good fit
of the model to this data with the estimate of the proportion of explained vari-
ance at 54%.
DISCUSSION AND APPLICATIONS
FOR SOCIAL WORK PRACTICE
The combined visual and hierarchical linear modeling analyses suggest
that a majority of the Anger Control Training students improved in their
weekly self-reports of using better anger control and experiencing more posi-
tive management and expression of anger. Also, these analyses reveal that,
generally, the Anger Control Training students presented fewer behavioral
problems on a weekly basis at the day treatment program when compared
with the students not receiving Anger Control Training. Barring any serious
problems with the statistical model, the hierarchical linear modeling analysis
also suggests that the Anger Control Training students fared much better with
their school behavioral adjustment (improved teacher-rated self-control) fol-
lowing the Anger Control Training than did the comparison students. Treat-
ment status in this case accounted for a fairly high proportion of explained
variance at 54%. Thus, evidence is present that some treatment gains for
422 RESEARCH ON SOCIAL WORK PRACTICE
some Anger Control Training students were maintained over several months
and generalized to regular school.
Other client developmental and environmental variables influence treat-
ment outcomes with this type of problem and clientele, however. Treatment
status explained relatively modest proportions of variances for the two
State-Trait Anger Expression Inventory scales (28% and 26%) and a some-
what higher proportion at 40% for the Staff Daily Report measure. The analy-
ses for the weekly time-series data suggest that other variables may have
affected the variation with the clients’ responses. With this type of limited
sample, many moderating and mediating variables (e.g., amenability to treat-
ment, drug and alcohol use, and family characteristics) could not be con-
trolled for, either through the design or statistically. A more discriminating
and fine-grained hierarchical linear modeling analysis would be possible
with larger samples and with additional hierarchical ordering of levels of
analysis (e.g., assessing several school programs across two or more school
systems).
This study is one answer to the calls in the literature for more outcome
research on specific, theoretically based interventions linked to educational
goals and interests. These results suggest that a focused, cognitive-
behavioral approach coupled with a general, relationship-oriented counsel-
ing is more effective in improving anger control and general self-control than
a nonspecific, relationship-oriented counseling approach alone. Thus, evi-
dence is present that a cognitive-behavioral approach is one method of effec-
tively addressing and reducing the problems associated with school violence.
This is a qualified appraisal, however, as the regular counseling component
was not operationalized for this study.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 423
TABLE 1: Means and Standard Deviations for the Self-Control Rating Scale
Total Problem Scores at Pretest, Posttest, and 6-month Follow-Up
Assessments for Each Group
Anger Control Training Control
M SD M SD
Pretest 163.63 22.48 163.88 45.56
Posttest 145.00 19.84 152.00 24.84
Six months 111.00 34.07 168.00 17.68
NOTE: n = 8 for each group for the pretest and posttest assessments, n = 5 for the
Anger Control Training group, and n = 4 for the control group for the 6-month follow-up
assessment.
Relatively rudimentary research designs (e.g., survey and descriptive
designs) and limited statistical analyses have dominated the professional lit-
erature of social work (Glisson, 1995) and school social work in particular
(Bailey-Dempsey, 1997; Sabatino et al., 1991). This study attempts to move
beyond these limited research efforts. Several researchers (e.g.,
Allen-Meares, 1988; Bloom et al., 1995; Nugent, 1987) have advocated
time-series, multiassessment, and multimethod research designs for social
work investigations, and the use of hierarchical linear modeling is an effec-
tive analytic strategy in exploiting the strengths and compensating for some
of the problems associated with integrated evaluation approaches.
The external validity of this study, however, is limited due to a relatively
circumscribed sample. Research as to treatment outcomes involving other
samples of youths in similar school social work settings is greatly needed.
Efforts need to be directed toward finding out what works best with these
conduct-disordered youths in redirecting their negative patterns of behavior
in school and in the community. The psychological, social, and physical costs
to these individuals, the general misery engendered by antisocial behavior,
and the monetary costs to society are too great to ignore research in this area.
REFERENCES
Achenbach, T. M. (1991a).Manual for the child behavior checklist/4-18 and 1991 profile. Bur-
lington: University of Vermont Department of Psychiatry.
Achenbach, T. M. (1991b).Manual for the teacher’s report form and 1991 profile. Burlington:
University of Vermont Department of Psychiatry.
Alexander, R., Jr., & Curtis, C. M. (1995). A critical review of strategies to reduce school vio-
lence.Social Work in Education, 17, 73-82.
Allen-Meares, P. (1988). Interrupted time series design and the evaluation of school practice. In
J. G. McCullagh & P. Allen-Meares (Eds.),Conducting research: A handbook for school so-
cial workers(pp. 103-115). Des Moines: Iowa State Department of Education.
Allen-Meares, P., Washington, R. O., & Welsh, B. L. (1996).Social work services in schools
(2nd ed.). Boston: Allyn & Bacon.
Astor, R. V., Behre, W. J., Fravil, K. A., & Wallace, J. M. (1997). Perceptions of school violence
as a problem and reports of violent events: A national survey of school social workers.Social
Work, 42, 55-68.
Bailey-Dempsey, C. (Ed.). (1997, April).NASW school social work section practice effective-
ness series(Practice Effectiveness Paper No. 1). Washington, DC: NASW School Social
Work Section.
Biskin, B. H. (1992). Review of the State-Trait Anger Expression Inventory, research edition. In
J. J. Kramer & J. C. Conoley (Eds.),The eleventh mental measurements yearbook(pp. 868-
869). Lincoln, NE: Buros Institute of Mental Measurements.
Bloom, M., Fischer, J., & Orme, J. G. (1995).Evaluating practice: Guidelines for the account-
able professional(2nd ed.). Boston: Allyn & Bacon.
424 RESEARCH ON SOCIAL WORK PRACTICE
Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to assessing
change.Psychological Bulletin, 101, 147-158.
Bryk, A. S., & Raudenbush, S. W. (1992).Hierarchical linear models: Applications and data
analysis methods. Newbury Park, CA: Sage.
Chamberlain, P., & Reid, J. (1987). Parent observation and report of child symptoms.Behavioral
Assessment, 9, 97-109.
Comer, R. (1985). Day treatment of adolescents: An alternative to institutionalization.Journal
of Counseling and Development, 64, 74-76.
Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social information process-
ing mechanisms in children’s social adjustment.Psychological Bulletin, 115, 74-101.
Dangel, R. F., Deschner, J. P., & Rasp, R. R. (1989). Anger control training for adolescents in
residential treatment.Behavior Modification, 13, 447-458.
Durlak, J. A., Fuhrman, T., & Lampman, C. (1991). Effectiveness of cognitive-behavior therapy
for maladapting children: A meta-analysis.Psychological Bulletin, 110, 204-214.
Epstein, N., & Maragos, N. (1983). Treating delinquent-prone adolescents and pre-adolescents.
Social Work, 28, 66-68.
Farrington, D. P., Loeber, R., & Van Kammen, W. B. (1990). Long term criminal outcomes of
hyperactivity-impulsivity-attention deficit and conduct problems in childhood. In L. N. Rob-
ins & M. Rutter (Eds.),Straight and devious pathways from childhood to adulthood(pp. 62-
81). Cambridge, UK: Cambridge University Press.
Feindler, E. L. (1979).Cognitive and behavioral approaches to anger control training in explo-
sive adolescents. Unpublished doctoral dissertation, West Virginia University, Morgantown.
Feindler, E. L. (1991). Cognitive strategies in anger control interventions for children and ado-
lescents. In P. C. Kendall (Ed.),Child and adolescent therapy: Cognitive-behavioral proce-
dures(pp. 66-97). New York: Guilford.
Feindler, E. L., & Ecton, R. B. (1986).Adolescent anger control: Cognitive-behavioral tech-
niques. Elmsford, NY: Pergamon.
Feindler, E. L., Marriott, S. A., & Iwata, M. (1984). Group anger control training for junior high
school delinquents.Cognitive Therapy and Research, 8, 299-311.
Finch, A. J., Jr., Nelson, W. M., III, & Moss, J. H. (1993). Childhood aggression: Cognitive-
behavioral therapy, strategies, and interventions. In A. J. Finch, Jr., W. M. Nelson III, & E. S.
Ott (Eds.),Cognitive-behavioral procedures with children and adolescents: A practical
guide(pp. 148-205). Boston: Allyn & Bacon.
Franklin, C. (1992). Alternative school programs for at-risk youths.Social Work in Education,
14, 239-251.
French, J. L., Lamke, T. A., & Nelson, M. J. (1973).Examiner’s manual for the Henmon-Nelson
Tests of Mental Ability. Chicago: Riverside Publishing.
Glick, B., & Goldstein, A. P. (1987). Aggression replacement training.Journal of Counseling
and Development, 65, 356-362.
Glisson, C. (1994). The effect of services coordination teams on outcomes for children in state
custody.Administration in Social Work, 18(4), 1-23.
Glisson, C. (1995). The state of the art of social work research: Implications for mental health.
Research on Social Work Practice, 5, 205-222.
Hains, A. A. (1989). An anger control intervention with aggressive delinquent youths.Behav-
ioral Residential Treatment, 4, 213-230.
Hinshaw, S. P. (1992). Academic underachievement, attention deficits, and aggression: Comor-
bidity and implications for intervention.Journal of Consulting and Clinical Psychology, 60,
893-903.
Whitfield / VALIDATING SCHOOL SOCIAL WORK 425
Kazdin, A. E. (1995).Conduct disorders in childhood and adolescence(2nd ed.). Thousand
Oaks, CA: Sage.
Kendall, P. C. (1993). Cognitive-behavioral therapies with youth: Guiding theory, current status,
and emerging developments.Journal of Consulting and Clinical Psychology,61, 235-247.
Kendall, P. C., & Wilcox, L. E. (1979). Self-control in children: Development of a rating scale.
Journal of Consulting and Clinical Psychology, 47, 1020-1029.
LeCroy, C. W., & Flynn, J. P. (1982). Journal reports of school social work practice and research.
In R. T. Constable & J. P. Flynn (Eds.),School social work: Practice and research perspec-
tives(pp. 29-37). Homewood, IL: Dorsey Press.
Loeber, R. (1991). Antisocial behavior: More enduring than changeable?Journal of the Ameri-
can Academy of Child and Adolescent Psychiatry, 30, 393-397.
McMahon, R. J. (1994). Diagnosis, assessment, and treatment of externalizing problems in chil-
dren: The role of longitudinal data.Journal of Consulting and Clinical Psychology, 62, 901-
917.
Michals, A. P., Cournoyer, D. E., & Pinner, E. L. (1979). School social work and educational
goals.Social Work, 24, 138-141.
Nugent, W. R. (1987). Information gain through integrated research approaches.Social Service
Review, 61, 337-364.
Pellegrini, D. S., Galinski, C. L., Hart, K. J., & Kendall, P. C. (1993). Cognitive-behavioral as-
sessment of children: A review of measures and methods. In A. J. Finch, Jr., W. M. Nelson
III, & E. S. Ott (Eds.),Cognitive-behavioral procedures with children and adolescents: A
practical guide(pp. 90-147). Boston: Allyn & Bacon.
Sabatino, C. A., Timberlake, E. M., & Hooper, S. N. (1991). Research in school social work:
Charting the future. In R. Constable, J. P. Flynn, & S. McDonald (Eds.),School social work:
Practice and research perspectives(2nd ed., pp. 403-418). Chicago: Lyceum Books.
Spielberger, C. D. (1991).State-Trait Anger Expression Inventory professional manual. Odessa,
FL: Psychological Assessment Resources.
Whitfield, G. W. (1996).An evaluation of anger control training with male adolescents in a day
treatment program. Unpublished doctoral dissertation, University of Tennessee, Knoxville.
Ziglar, E., Taussig, C., & Black, K. (1992). Early childhood intervention: A promising preventa-
tive for juvenile delinquency.American Psychologist, 47, 997-1006.
426 RESEARCH ON SOCIAL WORK PRACTICE
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
