The purpose of this study was to determine if parents of students could improve the motor performance of their moderately and severely mentally retarded children through the implementation of the Data-Based Gymnasium (DBG) Instructional Model as a supplement to the school program. The DBG Instructional Model is an individualized system designed to provide successful physical education experiences for moderately and severely handicapped students (Dunn, 1980). A total of 45 children, ages 3-9, with moderate and severe mental retardation from the states of Oregon and Washington took part in this study. Each child was randomly selected to participate in one of three groups based upon baseline data gathered for the three tasks: underhand roll, catch, overhand throw. During the eight-week treatment period, subjects in experimental group one received DBG instruction in school and at home, subjects in experimental group two received DBG instruction in school only, and subjects in the control group participated in the physical education program offered in their respective special education settings. The analysis of covariance was used to test the null hypothesis. The .05 level of significance was selected for use in this study. The null hypothesis that there was no significant difference between children who received DBG instruction in the home and school (E1) and those who received school DBG instruction only (E2) or no DBG instruction (C) was rejected on the basis of an obtained F value of 33.847. Application of the Student-Newman-Keul's Multiple Comparison's procedure across groups (E1, E2, and C) revealed that E1 and E2 were not statistically different (E1 = E2) and that E1 and E2 were statistically different from the control group (E1 # C, E2 # C). An examination of the raw and adjusted mean gain data for E1, E2, and C groups revealed a trend favoring the utilization of parents with the implementation of home instruction as a supplement to the school program.