(Baker and Heifetz, 1976; Rinn, Vernon and. Wise, 1975) and to placebo control (Walter and. Gilmore, 1973). Few studies, however, have directly com-.
J. Behov.Ther.&Exp.Psychiat.Vol.13,No.Z.pp. 113-117. 1982. Printed
OOOS-7916/82/020113-05 Cl 1982 Pmgamon
in Great Britain.
EFFECTIVENESS
OF ALTERNATIVE
PARENT
TRAINING
$03.00/O Press Ltd.
FORMATS
RICHARD P. BRIGHTMAN, BRUCE L. BAKER, DUNCAN B. CLARK and STEPHEN A. AMBROSE University of California. Los Angeles Summary-Alternative formats were compared for training parents of retarded children to teach self-help skills and manage problem behaviors. Sixty-six families with moderately to severely retarded children ages 3-13 were assigned for 3 months to group parent training (n = 37), individual parent training (n = la), or delayed training control (n = 13). Measures administered before and after training evaluated: (1) parent knowledge of behavior modification, (2) a behavior sample of parent teaching, and (3) child self-help skills and behavior problems. Trained families gained significantly more than control families on parent measures but not on the child self-help skill measure. Group and individually trained families demonstrated almost identical gains. At a 6-month follow-up, group and individually trained families continued to show equal performance. Group training requires about half the professional time per family as individual training, and therefore seems a more cost-effective approach.
In the two decades since Williams (1959) taught a mother to ignore contingently the bedtime tantrums of her 5-year-old son, numerous investigators have documented the efficacy of behaviorally based parent training in ameliorating a range of childhood problems (see Baker, 1976; Berkowitz and Graziano, 1972; Cone and Sloop, 1974; Johnson and Katz, 1973 for reviews). Individual consultation with parents dominated early reports and is still the common practice in clinical settings. However, in the light of the chronic shortage of child mental health services, researchers have increasingly sought to enhance the cost-efficiency of behavioral parent training programs by training parents in groups, using a standardized curriculum (Tams and Eyberg, 1976). Group parent training has been found superior to waiting list control (Baker and Heifetz, 1976; Rinn, Vernon and Wise, 1975) and to placebo control (Walter and Gilmore, 1973). Few studies, however, have directly com-
pared individual vs group parent training, with random assignment of families. Two reports of training for parents of behavior problem but intellectually normal children, found group and individual training to be equally effective and more beneficial than either “informal training” (Kovitz, 1976) or minimal contact bibliotherapy (Christensen et al., 1980). The latter study is particularly convincing as in-home data obtained via an unobrusive audio recording device revealed significant and equivalent decreases in child non-compliance for both clinic conditions while no such effect was found for bibliotherapy alone. Although it appears that group parent training is more cost-effective, further study is necessary. These studies trained parents of oppositional children in behavior management methods; it remains to be determined if parents of retarded children, who must learn skill teaching as well as behavior management strategies, are as well served in groups. These
This paper is based upon activities of the UCLA Project for Developmental Disabilities, sponsored by grant 5 ROl HD10962 from the National Institute of Child Health and Human Develooment to Dr. Bruce L. Baker. We would like to gratefully acknowledge the contribution of our colleagues on the project staff. Requests for reprints should be sent to Bruce L. Baker, Department of Psychology, UCLA, Los Angeles, CA 90024, U.S.A. 113
114
RICHARD
P. BRIGHTMAN,
BRUCE
L. BAKER,
studies were also somewhat limited by small sample sizes which might have precluded finding modest format differences; the number of families divided into group and individual formats was n = 14 (Kovitz, 1976) and n = 18 (Christensen et al., 1980). Finally, no follow-up comparisons of group vs individual training have been reported. The present study extended the comparison of group vs individual formats by training 46 families with retarded children and conducting a 6-month follow-up.
METHOD Overview of procedures Sixty-six families with retarded children were assigned to group parent training (n = 37), individual parent training (n = 16) or delayed training control (n = 13). Assignment was random except for three control families who applied for training shortly after the study began. Parents were seen in the clinic before and after training to assess their knowledge of behavioral principles and their behavior modification teaching proficiency. Measures were taken from the parent who would be primarily involved in teaching; in all but two cases this was the mother. Measures of child self-help skills and behavior problems were also taken before and after training. Six months after training, a project staff member who was not their trainer conducted a structured interview with each family at home to assess the extent and quality of follow-through teaching. Families Families with retarded children were recruited from the Los Angeles area through newspaper and television coverage and letters from regional centers. The demographic characteristics of these families are summarized in Table 1. Target children were moderately to severely retarded. The most frequent diagnosis was Down’s Syndrome (31%), followed by organic retardation of unknown etiology (29%) and autism (9%). Most of the children (94%) were receiving some educational services, primarily in preschool or special classes. One-way Analysis of Variance conducted on family and child variables revealed no significant differences among conditions. Training Trainers. The eight trainers were experienced in developing behavior modification programs for retarded children and had worked previously with families of retarded children; five trainers conducted both group and individual training. Description of Paining formats. The training curriculum was kept as similar as possible across group and individual formats. In both formats, parents attended a large group orientation meeting followed by nine training sessions (2 hr) each in the group training format, 1 hr each in the individual training format). Sessions l-4 focused on techniques for self-help skill training, sessions 5-7 focused
DUNCAN
B. CLARK Table
Age Mother Father Child Education
1. Family
A. AMBROSE
demographic
data
Mean
Range
36 39 6
25-60 22-65 2-15
Mean
Mother Father Marital
and STEPHEN
14 15 status
%
Single Intact
17 83
Employment
Full time %
Part time %
Mother Father
26 90
13 0
Unemployed % 61 10
on stategies for developing and implementing behavior problem management programs, and sessions 8-9 extended these strategies to enhancing speech and play skills. The first six sessions were weekly, the last three bi-weekly. Between sessions, parents read manuals and completed homework assignments relevant to the current topic (e.g. setting target goals, task analysis, monitoring progress, keeping behavior problem data). During the training period, most families worked on several self-help skills and at least one behavior problem. Although the content and sequence of training were comparable for both formats, the process of training differed markedly. Parents in the five groups attended meetings in one of four community-based centers with two trainers and seven or eight families in each group. In these sessions, training was generally more didactic, although co-consulting and small group media, role-playing, problem solving were used extensively. Parents in the individual format brought their children with them to one of two centers, to meet with a trainer who would observe their teaching and provide suggestions, modeling and videotaped feedback. Training materials. Parents in both training conditions received behavior modification training manuals from the Steps to Independence series, covering self-help skills, toilet training, behavior problems, and speech and language (Baker ef al., 1976, 1977, 1978). Brief training videotapes were also utilized in each session (Blatt, Brightman and Baker, Note 1). Measures Behavioral Vignettes Tesf (BVT). Parents’ knowledge of behavior modification principles was assessed by the BVT, a 20-item multiple choice questionnaire (Heifetz, Baker and Pease, Note 2). Brief vignettes describe situations that reflect a range of common problems in the formulation and implementation of skill teaching and behavior management programs; parents select the response they think would be
115
PARENT TRAINING FORMATS most effective. In the present sample, the Spearman-Brown corrected split half reliability of pre-training BVT’s was r = 0.15. Modified training proficiency scale. Parents’ proficiency in teaching their children to use behavior modification principles was assessed by a modification of Gardner, Brust and Watson’s (1970) Training Proficiency Scale. Measures were obtained for teaching self-help and play skills. Each primary parent was aided by a trainer in selecting three level-appropriate self-help skills for which the child had skill mastery of lo-15%. For play skills, parents were given a level-appropriate puzzle and also instructed to select an additional one of several available toys and games to teach. Parents were then videotaped as they tried to teach their child the three self-help and two play skills. Trained coders independently viewed the 15-20 min session and scored 29 elements of teaching, each on a five-point scale. These item scores were then combined to form a self-help skill and play skill teaching score for each parent. The interrater reliability based upon 100 sessions was r = 0.69 for selfhelp and r = 0.77 for play. Performance inventory: self-help. Child performance levels on 38 specific skills in the areas of dressing, eating, grooming, toileting and housekeeping were assessed by the PI Self-Help Scale, completed by parents prior to the first training meeting. The PI arranges component parts of each skill in a hierarchical scale (ranging from “Child cannot do any part of this skill” to “Child can perform this skill completely independently”), and instructs parents to check the step that most accurately reflects their child’s performance. All items are transformed into per cent scores and summed to form a Self-Help Skill score for each child. To test PI reliability, observers who were blind to the parent ratings visited randomly selected sub-samples of 18 families prior to training and observed 11 families after training. They observed the child perform eight skills for which parents had reported a level of mastery of between 30 and 70% (on the assumption that reports would be less accurate in the middle ranges than at the extremes). A Pearson r was computed between observer and parent scores for each family. The average Spearman-Brown corrected correlation between trainer and parent ratings was r = 0.94 pre-training and r = 0.97 post-training, similar to previous findings with this measure (Baker and Heifetz, 1976). Behavior problems. Child behavior problems were assessed by a checklist of 51 commonly reported behavior problems for retarded children; parents indicated the three problems that concerned them most. Follow-up interview. A B-month follow-up interview of about 1 hr duration was conducted in the home. Hierarchical questions about teaching effort were posed by the interviewer; as parents reported more programming, the interviewer sought more detailed and specific information. Interviews were tape recorded and later scored on two major dimensions: extent of continued programming (teaching or behavior problem management) and appropriateness of behavioral techniques employed. Interrater reliabilities for a sub-sample of 14 interviews were r = 0.87 and 0.90 for the two dimensions, comparable to the reliability previously reported for an earlier version of this interview (Baker, Heifetz and Murphy, 1980).
RESULTS Completion
of training
The rate of program completion was 87% and essentially equivalent for the two training formats; 14 of 16 individual training families and 32 and 37 group training families completed the program and post-measures. Parent measures A 2 x 3 ANOVA on pre- and post-training
mothers’ BVT scores showed significant Condition (F(2.55) = 4.00, P = 0.02), Time (F(1.55) = 46.96, P < 0.001) and Condition x Time effects (F(2.55) = 4.08, P = 0.02). Trained mothers showed a significant BVT gain (t(45) = 8.62, P < 0.001). The gain for trained mothers was significantly greater than the gain for control mothers (t(57) = 2.91, P = 0.003). BVT gain scores for group versus individual formats did not differ (t(44) = 0.07, ns). A 2 x 3 ANOVA on pre- and post-training mothers’ TPT scores for self-help teaching showed significant Condition (F(2.51) = 5.78, P < O.Ol), Time (F(1.51) = 29.46, P < 0.001) and Condition x Time effects (F(2.51) = 6.23, P < 0.01). Trained mothers showed a significant TPT: self-help gain (t(43) = 7.87, P < 0.001). The gain for trained mothers was greater than the gain for control mothers (t(52) = 3.52, P < 0.001). TPT: self-help gain scores for group versus individual formats did not differ (t(42) = 0.02, ns). Improvement in TPT: plays skills teaching somewhat reflected generalization of principles to a skill area not stressed in training. A 2 x 3 ANOVA on pre- and post-training mothers’ TPT scores for play skills teaching showed significant Condition (F(2.50) = 10.75, P < O.OOl), Time (F(1.50) = 17.89, P < 0.001) and Condition x Time effects (F(2.50) = 3.53, P = 0.04). Trained mothers showed a significant TPT: play gain (t(43) = 4.86, P < 0.001). The gain for trained mothers was significantly greater than the gain for control mothers (t(52) = 2.91, P < 0.01). TPT: play gain scores for group versus individual formats did not differ (t(42) = 0.98, ns).
116
RICHARD P. BRIGHTMAN,
BRUCE L. BAKER, DUNCAN B. CLARK and STEPHEN
Child measures
A 2 x 3 ANOVA on PI: self-help skills showed a significant main effect for Time (F(1.56) = 90.43, P < 0.001). However, there was no significant Condition effect or Condition x Time interaction. Children in all three conditions showed significant skill gains. Self-help gain scores for group vs individual formats did not differ (t(44) = 1.09, ns). A 2 x 3 ANOVA on Behavior Problem scores showed a significant main effect for Time (F(1.42) = 18.93, P < 0.001). There was no conditions effect and the Conditions x Time interaction did not reach significance (F(2.42) = 2.41, P = 0.10). Since the interaction approached significance, t-tests were conducted. Children in trained families showed a highly significant decrease in behavior problems (t(37) = 6.32, P < 0.001) and decreased significantly more than controls (t(43) = 2.12, P = 0.04). Behavior problem improvement for group vs individual formats did not differ (t(36) = 0.59, ns). Folio w-up
Follow-up interviews were conducted with 41 of the 46 families who completed training. Parents trained in the group versus individual formats did not differ on either the extent of continued programming (t(39) = 0.71, ns) or on the quality of behavioral techniques employed (G > I, t(39) = 1.57, ns). Families were categorized as high, medium or low follow-through based upon a combination of the programming and technique dimensions. Families above the mean for the sample on both dimensions constituted the high group; generally these 15 families had productively continued the programs they began during training and initiated some new teaching and/or behavior problem management following training. Families at least one standard deviation below the mean on either dimension constituted the low group; these 10 families reported’ little or no continued teaching or demonstrated inadequate behavioral technique. The remaining 16 families constituted the
A. AMBROSE
medium group; these had continued some degree of useful teaching. Group and individually trained families did not differ by follow-through category (Chi Square (2) = 1.21, ns). DISCUSSION Multimethod assessment procedures were used to evaluate the utility of a media-based behavioral training curriculum administered either in groups or individually to parents of retarded children. Two findings emerged consistently. First, trained parents gained significantly more than controls. They acquired more knowledge about behavior modification and increased considerably more in their teaching proficiency during observed interactions with their child. Children in trained families tended to improve more than controls on the behavior problems of most concern to their parents. Child self-help skills, however, showed a comparable gain in trained and control families. Almost all children were in school during the training period and several control children showed large self-help gains which precluded demonstrating a Condition x Time interaction with these small samples. Secondly, the group and individual training formats were equally effective. Moreover, training formats did not produce differential effects at a B-month follow-up. Since group training required half the number of professional contact hours per family as individual it was the more cost-efficient training, approach. These findings are consistent with earlier cited results reported by researchers working with oppositional non-retarded children. Each of the training formats studied may have unique advantages. Individual training affords the opportunity for greater flexibility. In attempting to keep curriculum content consistent across formats in the present study, however, trainers tended not to address clinical issues that went beyond program content and thereby did not fully exploit the potential flexi-
PARENT TRAINING FORMATS
bility of the individual format. Group training affords more support for trainers and parents alike. Staff preferred group training since the co-trainer model includes a colleague to share responsibilities and aid in problem solving. For parents, group training provides peer support and the opportunity to exchange information about experiences and services; however, our measures did not assess potential benefits from these characteristics. Overall in the present study, the strong reliance upon written and visual media in both training formats and the use of a stardard training curriculum served to reduce the variability between formats. In conclusion, although there may be some families who would derive more benefit from individual training, limited resources suggest a broader adoption by agencies of a group training format. REFERENCE
NOTES
1. Blatt J., Brightman A. J. and Baker B. L. (1975) Parents as Teachers Video Series: (1) Self-help Skills for Children with Special Needs, (2) Play Skills for Children with Special Needs. 2. Heifetz L., Baker B. L. and Pease L. A. (1981) TheBehoviorul Vignettes Test. Unpublished manuscript, Syracuse University.
REFERENCES
Baker B. L. (1976) Parent involvement in programming for for the developmentally disabled child. In Communicution Assessment and Intervention (Ed. by Lloyd L. L.), pp. 691-733. University Park Press, Baltimore, MD. Baker B. L., Brightman A. J., Carroll N. B., Heifetz B. B. and Hinshaw S. P. (1978) Steps to Independence Series, Speech and Language: Level I, Speech and Language: Level 2. Research Press, Champaign, Ill.
117
Baker B. L., Brightman A. J., Heifetz L. J. and Murphy C. (1976) Steps to Independence Series, Bahavior Problems, Early Self-Helu Skills, Intermediate Self-He10 Skills. Ad&we; Sel;f-Help Skills (1977), Toilet ?roiningl Research Press, Champaign, Ill. Baker B. L. and Heifetz L. J. (1976) The READ Project: Teaching manuals for parents of retarded children. In Intervention Strategies for High Risk Infants and Young Children (Ed. by Tjossem T.D.). University Park Press, Baltimore, MD. Baker B. L., Heifetz L. J. and Murphy D. (1980) Behavioral training for parents of retarded children: Oneyear follow-up, Am. J. Ment. Deft. 85 (1). 31-38. Berkowitz B. P. and Graziano A. M. (1972) Training parents as behavior therapists: A review, Behov. Rex Ther. 10,297-317. Christensen A., Johnson S. M., Philips S. and Glasgow R. E. (1980) Cost effectiveness in behavioral familv therapy, Behuv. Ther. 11,20-226. Cone J. D. and Slooo E. W. (1974) Parents as aaents of change. In The Gioup us kgeni of Change
