University of Maryland
College Park, Maryland
A Model for Determining Group Counseling Outcomes
Susan S. Gerson, William E. Sedlacek, and
Robert W. Freeman
Research Report # 1-91
Computer time for this study was furnished by the Computer
Science Center, University of Maryland at College Park.
College Park, Maryland
A Model for Determining Group Counseling outcomes
Susan S. Gerson, William E. Sedlacek, and Robert W. Freeman
Research Report # 1-91
The goal of this article is to address areas of concern, raised by previous research, in order to provide a model for measuring outcomes of group interventions in school settings.
The researchers utilized objective measures of behavior change, experimental and control groups, and carefully documented procedures so that they might easily be reproduced. This model is intended to encourage counselors to measure group outcomes so that they may have tangible information about what group interventions work most effectively.
Since their inception in the 1940's group therapies have grown in popularity and frequency. The variety of interventions that fit the label "group counseling" as well as the frequency of their use has caused group interventions to be integral to the field of psychological counseling (Yalom, 1985).
Small group guidance and counseling is one of the primary activities of school counselors in elementary, middle/junior, and senior high schools (Bowman, 1987). However, a variety of problems plague counselors leading these groups. Resistance from teachers or parents and lack of confidence in their group skills lead many counselors to be unsure about the future of counseling groups. In addition, counselors often do not determine whether their groups are an effective or just an efficient method of working with students. Systematic methods of measuring outcomes are not typically used by school counselors.
The majority of research on interventions with children has been done in a school setting (Riester & Dies, 1986). The school is the primary environment for school-aged children (other than the home) and is therefore a comfortable and familiar place for the intervention to occur. Additionally, the students often do not need to be transported for intervention, and therefore minimal disruption to their lifestyle will occur.
The most common approaches to measuring student change have included self-report measures, teacher's observations, and to a lesser extent, independent observations (Riester & Dies, 1986). The problems with these approaches are that some important
perspectives are missed (most notably that of the parents). Additionally, the behavior has often been rated with poorly developed measures or by only one source. LaGreca (1983) points out the importance of using several people in the child's environment to do assessments. Barret, Hampe, and Miller (1978) also support this position.
Another area of concern is that much research relies on the counselor's own perception of change because there are no comparison groups present. Freeman, Anderson, Kairey, and Hunt, 1982; DuPlessis and Lochner, 1981; and McKibbin and King, 1983; like numerous others have found behavior change in children but due to the lack of control groups, were not able to know if the change was due to a given treatment.
It is clear that a variety of problems have plagued the study of the effectiveness of group interventions with children (Riester & Dies, 1986). The greatest concerns have been a lack of objective measures being used, failure to employ non-treatment control groups, and a lack of reproducible research projects. The goal of this article is to address each of these areas of concern in order to provide a model for measuring outcomes of group interventions in school settings.
In order to be able to effectively attribute treatment effects to the intervention, this study utilized both an experimental group and a control group. The control group engaged in an alternate activity for the same period of time as the experimental
group in order to control for the Hawthorne effect (the tendency f or any form of attention paid to the students to off ect results) .
The Behavior Rating Profile (BRP) was used because it is a well researched and reliable measure for gaining information about a child's behavior at home, school, and with peers. Test-retest results for emotionally disturbed and learning disabled elementary aged students ranged from .76 - .97, with the teacher's ratings being the most reliable (.95 - .97), (Brown & Hammill, 1983). It is a measure which has met the standards set by Public Law 94-142 for testing special education students. The greatest concerns raised about using behavior checklists are the subjectivity of the rater and the necessity of a basal reading level of the rater (Barkley, 1987). While the BRP's reliability indicates that subjectivity is not a problem, reading level could have been a problem for some of these students or parents.
Three points were specified in this research to enhance reproducibility: a clear definition of the sample, a description of the therapists, and a description of the group activity.
The participants (18 experimental, 18 control) had one or more of the following disabilities: mental retardation, "serious emotional disturbance", or learning disability. All of the students were recommended by their teachers for the program. Students ranged in age from 8-12 years old. Parents' permission was required to participate.
Groups were conducted over 2 years with 8 students in each group during year 1, and and 10 in each group during year 2. A PhD psychologist as well as two graduate students in counseling and psychology worked with the experimental groups on Friday mornings for 2 hours. Concurrently, the control group went bowling with classroom teachers or were involved in a special activity program.
The experimental group time included an indoor art project, an outdoor group game and a snack time. These active periods were the times used for meaningful group interaction. Due to the high energy levels of most group emembers, just sitting and talking was not a strategy that would be conducive to the group process. During the project time group members were encouraged to work diligently on their projects and to keep on task. Positive verbal reinforcemnt was utilized to encourage these behaviors and negative behaviors were ignored. During the outside games, organized group sports ensued. Group members were encouraged to take turns, cooperate, and to be "good sports." Snack time was utilized as a "wind down" period before their departure back to the classroom and a time of discussion of the day's activities.
Groups met once a week for ten weeks. Experimental groups met during the fall term only. Their behavior was measured in September (pre), January (post) and June (post). Control groups engaged in the non-treatment activity in the fall term and then engaged in the treatment during the spring semester. Their behavior was measured in September (pre), January (non-treatment
post) and June (treatment post). A comparison was made between experimental and control group means over time.
Year 1: Group means were examined on 5 different areas: parent rating of home behavior, teacher rating of school behavior, child's rating of home behavior, child's rating of school behavior, and child's rating of peer relations. Both control and experimental groups showed slight group changes over time from pre to post tests. While the experimental group had slightly higher means than the control group, both groups' scores increased slightly from pre to post test on all dimensions. The changes from pre to post test for experimental and control groups were not statistically significant (.05 level).
Year 2: Group means were examined on teacher rating of school behavior, child's rating of home behavior, child's rating of school behavior, and child's rating of peer relations. This change was made because many parents experienced difficulty in understanding the questionnaire items during Year 1. Both control and experimental groups showed slight group changes over time from pre to post tests, but the differences were not statistically significant (.05 level).
Therefore, it is possible that the treatment did not result in any change in the experimental group members' behavior. However, there are a variety of concerns in the data collection that indicate other issues that may have affected the outcome of the study.
Three primary concerns may be raised about the results found, indicating needed changes for future research. The first concern is in the use of the BRP. Several parents as well as students were unable to understand the language used in the questions. The authors of the BRP indicate that it is an appropriate measure for this population, however, our experience indicates that this measure was beyond the comprehension of several students and their parents. Caution should be used in future work with this measure for special populations.
Second, it is possible that the goal of this research was too comprehensive. To observe overall behavior improvements in children whose primary difficulty is with their behavior is a large goal. Many of these children spend years in special placements in order to improve their behavior. Perhaps, a more attainable goal for ten weeks work could have been chosen. Future researchers might measure willingness to share, ability to wait one's turn or a variety of other discrete behaviors to determine the success of group therapy with behaviorally disordered children.
This research did successfully meet many of the previous concerns of researchers of groups treatments for children. Control comparison groups, multiple behavior ratings, clearly defined intervention, and credentialed group leaders were achieved.
However, some new concerns were raised: size of sample, using measures appropriate for the sample, and devising goals appropriate to the abilities of the students. Future research in this area is
still promising given that these areas of difficulty are improved.
The counseling needs of children are growing every day. School counselors will be using groups with greater frequency in order to meet the increasing needs of school children. It is therefore necessary to continue to research improved methods of measuring outcomes of groups, in order to determine under what conditions children most benefit from group interventions.
Barkley, R. A. (1987). The assessment of attention deficit hyperactivity disorder. Behavioral Assessment, 9,
Barret, C. L., Hampe, I. E., & Miller, L. C. (1978). Research on child psychotherapy. In: Handbook of
psychotherapy and behavior change: An empirical analysis, ed. S.L. Garfield & A.E. Bergin. New York:
Brown, L. L., & Hammill, D. D. (1983). Behavior rating profile-. Austin, TX: Pro-ed.
DuPlessis, J. M., and Lochner, L. M. (1981). The effects of group psychotherapy on the adjustment of four
12-year-old boys with learning and behavior problems. Journal of Learning Disabilities, 14, 209-212.
Freeman, R. W., Anderson, C., Kairey, I., & Hunt, P. F. (1982). Evaluation of Camp Tortuga, a two week
children's therapeutic day camp via goal attainment scaling and locus of control. Children's Youth Service Review, 4, 375-388.
LaGreca, A. M. (1983). Interviewing and behavioral observations. In: Handbook of clinical child psychology,
eds. C. E. Walker & M. C. Roberts. New York: Wiley, pp.109-131.
McKibbin, E., and King, J. (1983). Activity group counseling for learning disabled children with behavior
problems. American Journal of Occupational Therapy, 37, 617-623.
Riester, A. E., & Dies, R. R. (1986). Research on child group
therapy. In: Child group psychotherapy: Future tense, eds.
A. E. Riester & I. A. Kraft. Madison, CT: International
Universities Press, pp. 173-220
Yalom, I. D. (1985). The theory and practice of group
psychotherapy. New York: Basic Books.