Exceptional Children, Vol 59, No. I, pp. 77-86. © 1992 The Council for Exceptional Children
Teaching Children with Autism Through Task Variation in Physical Education ROBERT C. WEBER Texas Tech University JOANNE THORPE Southern Illinois University
ABSTRACT: The purpose of the study was to determine whether the technique of task variation (with; maintenance tasks interspersed) (TV) is more effective in the acquisition ofgross motor skills for students with autism than a constant task (CT) condition in a physical education setting. Subjects were 12 male students with autism, ages 11 to 15 years. The study included pretest-posttest administration of the I CAN assessment of Gross Motor Skills to assess skills such as overhand throw, kick, and vertical jump. After a 6-week treatment period. the TV condition was significantly more effective than the CT condition, at the .05 level.
D For more than 40 years, researchers and teachers in physical education have been concerned with the way certain variables affect the learning experiences of students (e.g. Ammons, 1947; Caplan, 1969; Digman, 1959). Many researchers have compared the influence of constant practice and distributed practice methods on students' acquisition of motor skills. In constant practice, only one experimental task or activity is presented per session. According to Drowatzy (1970), this is commonly referred to as "massed practice," "fixed-format," or the "serial" approach. Distributed practice typically includes rest pauses or alternate skills learning between practice trials. Physical educators and coaches have successfully used both forms of learning with students without disabilities. According to Ammons (1950), Caplan (1969), Digman (1959), and Koonce, Chambliss, and Irion (1964), distributed practice conditions are superior for people without disabilities. However, Ammons (1947), Harmon and Oxendine (1961) and Young (1954) have indicated that
massed practice is also effective. Other studies showed no significant differences in learning for people without disabilities, for massed and distributed practice (Carron, 1969; Graw, 1968; Stelmach, 1969; Whitley, 1970). In most cases, the conditions of these studies were not comparable. Moreover, these studies did not address the needs of people with disabilities-nor, more specifically, those with autism. In the past 10 years, however, research in special education concerning the academic classroom performance of children with autism has become more prevalent (Dunlap, 1984; Dunlap & Koegel, 1980a; Winterling, Dunlap, & O'Neill, 1987). Many of these studies involved various forms of distributed practice. A classic case was Dunlap's (1984) investigation. In Dunlap (1984), a specific type of distributed learning referred to as "task variation with maintenance tasks interspersed" was compared with a constant task condition in a special education classroom for students with autism. Dunlap's results indicated that the task variation condition 77
Exceptional Children
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
was significantly more effective than the constant task condition for learning cognitive skills in the special education classroom. However, Dunlap's findings are not necessarily transferable to a physical education setting. Singer (1975) and Gagne (1973) suggested that there are five domains oflearning-motor skills, verbal skills, intellectual skills, cognitive strategies, and attitudes-and that a particular learning condition or technique can be successful for one of the five domains of learning and not necessarily for another. In fact, Singer and Gagne both indicated that the conditions or techniques that are the most powerful or successful are usually unique to a specific domain. The question addressed here is: Is the distributed-learning technique used by Dunlap (1984) more effective than a constant task condition in teaching gross motor skills to students with autism? We have been concerned that procedures used in physical education classes for students with autism have not maximized instructional time. Nor have the procedures accounted for individual learning characteristics such as slow acquisition, short attention span, and a reduced ability to generalize (French & Jansma, 1982; Sherrill, 1986), which differentiate people with autism from other learners (Mulligan, Guess, Holvoet, & Brown, 1980). In the area of motor learning, most of the research has been directed toward the acquisition of fine motor skills in novel motor tasks. Little information is available concerning the acquisition of gross motor skills, and the literature appears to be void in studies involving practice conditions for gross motor skills by people with autism.
METHOD
Selection of Subjects The subjects for this study were 12 autistic males, ages 11-15, representing the total population from one rural school. All were previously diagnosed as autistic. For the purposes of inclusion in this study, the subjects had to satisfy the criteria suggested by Coleman (1976). All students displayed rocking, hand flapping, and mouthing and spinning of objects. According to Rimland (1964) and Ritvo and Freeman (1978), these characteristics represent classical self-stimulatory behavior for autistic individuals. Each student was estimated to have a social quotient between 20 and 35 on the Vineland Social Maturity Scale when tested by the school psychologist. The subjects were randomly assigned to the constant (massed) task treatment group (EI) or the distributed-learning technique of task variation (with interspersed maintenance tasks) treatment group (E2). The six students in E1 had a mean social quotient on the Vineland Social Maturity Scale of 28.5 (range 24-35, SD 3.62) and a mean age of 13 years, with a range of 11-15 years. The six students in E2 had a mean social quotient on the Vineland Social Maturity Scale of 28.0 (range 24-35, SD 3.52) and a mean age of 13 years, with a range of 11-15 years. All children were nonverbal and were considered to be homogeneous in their psychological behavior, which included behavior disorders.
EXPERIMENTAL PROCEDURES The experimental procedures used throughout this study are similar to those suggested by Dunlap (1984), except that the skills involved are gross motor skills rather than cognitive. According to Singer (1975), a great disparity may exist because students react differently to various teaching methods when learning motor skills.
Experimental Conditions
Design The design for this study was a pretest-posttest configuration; two experimental task sequences were analyzed for skill acquisition. No control group was used. This design is useful for observing the comparative effect of two treatments (Thorpe, 1986). To control for order effects, the group taught first each day was alternated between the two conditions (Hersen & Barlow, 1976; Kazdin & Hartmann, 1978).
Physical Environment. All teaching and evaluation sessions were conducted in a 30 ft by 20 ft room with alOft (3m) ceiling. Access to the room was through one set of double doors that had a small window, through which sessions could be observed from an adjoining room. The room was equipped with three bicycle ergometers, one minitrampoline, and a storage box filled with balls of various sizes.
78
September 1992
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
Instructional Conditions. A four-teachers-tosix-students format was maintained for all sessions. The teachers worked with four students on a one-to-one basis while the other two students were riding the bicycle ergometers. During the course of each class session, each teacher normally worked at least once with each child so that the child did not become dependent on one specific teacher. The instructors for this study were two females and two males ranging in age from 23 to 42. All four were adapted physical education aides with a minimum of 8 months of teaching experience in adapted physical education. Each instructor had completed extensive inservice training in adapted physical education under the supervision of an adapted physical education specialist. In addition, as suggested by Dunlap (1984) and Koegel, Russo, and Rincover (1977), each had completed inservice training courses in operant conditioning and had been trained to an 80% criterion level in the use of behavior modification for.children with autism by the Director of Special Education. All four worked with both treatment groups. Regular assessments of on-task behavior for all the instructors were conducted, as recommended by Dunlap (1984) and Dunlap and Koegel (1980a). These procedures had been adopted as a part of the standard operating code at this school; however, for the present study, these assessments were conducted twice per week rather than once per month. Similar to Dunlap (1984), the teachers were expected to maintain at least an 80% performance level of ontask activity throughout all their teaching for all tasks under both conditions. Because all four teachers conducted instructional periods with both treatment groups, the potential for teacher bias was a concern. To control for this potential effect, as recommended by Drew (1980), the teachers were not aware of the experimental hypothesis. The teachers were told that the students had been grouped and assigned to their respected treatment classes to address their individual needs for varying instructional methods as suggested by their test scores. Instructional Procedures. The instructional procedures used in this study were similar to those used by Dunlap (1984) and Dunlap and Koegel (1980b). During the study, the standard teaching techniques of physical shaping, physical assistance, physical prompting, and social reinforce-
ment were applied (Fait, 1960; French & Jansma 1982; Sherrill, 1986).These techniques were provided and then were slowly eliminated until the students responded successfully on their own. If a student failed to respond successfully after three consecutive trials, the teacher went back to the next lower level in the teaching techniques used. All correct responses, whether with physical shaping, physical assistance, prompting, or independent action, were reinforced with social reinforcement and a "pat on the back." Similar to Dunlap (1984), incorrect responses resulted in the teacher's saying "No _ " (using the subject's name) and withdrawing attention. An off-task or disruptive behavior was treated according to the procedural guidelines suggested by Arnheim, Auxter, and Crowe (1969). These procedures had been approved by the school's interdisciplinary team. The procedures routinely consisted of planned ignoring, signal interference, proximity control, or brief physical restraint (Arnheim et aI., 1969). Instructional sessions were held for a 45-miIi duration for both conditions. The performance criterion of 9 correct independent responses out of 10 consecutive trials of the task, which Dunlap (1984) used, was selected as the successful performance criterion for ending training. This criterion applied to all experimental tasks in both conditions. However, throughout the training period and during instruction, these criteria were never met. Classes for all students were conducted 5 days per week. A minimum of 15 min was provided between the two classes so that teachers could arrange equipment, use the bathroom, organize data sheets, and allow students to move to and from classes. Independent Variables The question addressed in this experiment is whether the distributed practice technique of task variation (with maintenance tasks interspersed) is more effective than a constant task condition (commonly referred to as massed practice) when teaching gross motor skills to people with autism. Six experimental acquisition tasks were randomly presented under both conditions. In addition, the task-variation group practiced the extra-experimental tasks of run, catch, and hop, which they had previously learned but were maintaining. A complete description of each of the experimental and extra-experimental tasks
Exceptional Children
79
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
FIGUREl Class Performance Score Sheet for the Six Basic Gross Motor Skills
Continuous Bounce FOCAlPOINTS
SCORING
STD
Au_mont:
~:~::r-
R_...ment: II • Achieved f). Not
"""i_
Underhand Roll FOCAL POINTS
SCORING Aa.... ment: X.Ac/li_ O.NoIachievod
"-
b R_.oIB..
R.......ment: II.Adliovod f). NoIochievod
-. -,
STD
• Pondu,", Arm Swing
• El'OO on Torgot
10ft.
"-,
b Pondu,", Arm Swinw
• Sh""ld_ "Squarod"
""-
2Oft.laIgol II 15 ft..213
"-
b PondulorArmMotion
"
c Palm F.ang Diroction d _
Teaching Children with Autism Through Task Variation in Physical Education ROBERT C. WEBER Texas Tech University JOANNE THORPE Southern Illinois University
ABSTRACT: The purpose of the study was to determine whether the technique of task variation (with; maintenance tasks interspersed) (TV) is more effective in the acquisition ofgross motor skills for students with autism than a constant task (CT) condition in a physical education setting. Subjects were 12 male students with autism, ages 11 to 15 years. The study included pretest-posttest administration of the I CAN assessment of Gross Motor Skills to assess skills such as overhand throw, kick, and vertical jump. After a 6-week treatment period. the TV condition was significantly more effective than the CT condition, at the .05 level.
D For more than 40 years, researchers and teachers in physical education have been concerned with the way certain variables affect the learning experiences of students (e.g. Ammons, 1947; Caplan, 1969; Digman, 1959). Many researchers have compared the influence of constant practice and distributed practice methods on students' acquisition of motor skills. In constant practice, only one experimental task or activity is presented per session. According to Drowatzy (1970), this is commonly referred to as "massed practice," "fixed-format," or the "serial" approach. Distributed practice typically includes rest pauses or alternate skills learning between practice trials. Physical educators and coaches have successfully used both forms of learning with students without disabilities. According to Ammons (1950), Caplan (1969), Digman (1959), and Koonce, Chambliss, and Irion (1964), distributed practice conditions are superior for people without disabilities. However, Ammons (1947), Harmon and Oxendine (1961) and Young (1954) have indicated that
massed practice is also effective. Other studies showed no significant differences in learning for people without disabilities, for massed and distributed practice (Carron, 1969; Graw, 1968; Stelmach, 1969; Whitley, 1970). In most cases, the conditions of these studies were not comparable. Moreover, these studies did not address the needs of people with disabilities-nor, more specifically, those with autism. In the past 10 years, however, research in special education concerning the academic classroom performance of children with autism has become more prevalent (Dunlap, 1984; Dunlap & Koegel, 1980a; Winterling, Dunlap, & O'Neill, 1987). Many of these studies involved various forms of distributed practice. A classic case was Dunlap's (1984) investigation. In Dunlap (1984), a specific type of distributed learning referred to as "task variation with maintenance tasks interspersed" was compared with a constant task condition in a special education classroom for students with autism. Dunlap's results indicated that the task variation condition 77
Exceptional Children
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
was significantly more effective than the constant task condition for learning cognitive skills in the special education classroom. However, Dunlap's findings are not necessarily transferable to a physical education setting. Singer (1975) and Gagne (1973) suggested that there are five domains oflearning-motor skills, verbal skills, intellectual skills, cognitive strategies, and attitudes-and that a particular learning condition or technique can be successful for one of the five domains of learning and not necessarily for another. In fact, Singer and Gagne both indicated that the conditions or techniques that are the most powerful or successful are usually unique to a specific domain. The question addressed here is: Is the distributed-learning technique used by Dunlap (1984) more effective than a constant task condition in teaching gross motor skills to students with autism? We have been concerned that procedures used in physical education classes for students with autism have not maximized instructional time. Nor have the procedures accounted for individual learning characteristics such as slow acquisition, short attention span, and a reduced ability to generalize (French & Jansma, 1982; Sherrill, 1986), which differentiate people with autism from other learners (Mulligan, Guess, Holvoet, & Brown, 1980). In the area of motor learning, most of the research has been directed toward the acquisition of fine motor skills in novel motor tasks. Little information is available concerning the acquisition of gross motor skills, and the literature appears to be void in studies involving practice conditions for gross motor skills by people with autism.
METHOD
Selection of Subjects The subjects for this study were 12 autistic males, ages 11-15, representing the total population from one rural school. All were previously diagnosed as autistic. For the purposes of inclusion in this study, the subjects had to satisfy the criteria suggested by Coleman (1976). All students displayed rocking, hand flapping, and mouthing and spinning of objects. According to Rimland (1964) and Ritvo and Freeman (1978), these characteristics represent classical self-stimulatory behavior for autistic individuals. Each student was estimated to have a social quotient between 20 and 35 on the Vineland Social Maturity Scale when tested by the school psychologist. The subjects were randomly assigned to the constant (massed) task treatment group (EI) or the distributed-learning technique of task variation (with interspersed maintenance tasks) treatment group (E2). The six students in E1 had a mean social quotient on the Vineland Social Maturity Scale of 28.5 (range 24-35, SD 3.62) and a mean age of 13 years, with a range of 11-15 years. The six students in E2 had a mean social quotient on the Vineland Social Maturity Scale of 28.0 (range 24-35, SD 3.52) and a mean age of 13 years, with a range of 11-15 years. All children were nonverbal and were considered to be homogeneous in their psychological behavior, which included behavior disorders.
EXPERIMENTAL PROCEDURES The experimental procedures used throughout this study are similar to those suggested by Dunlap (1984), except that the skills involved are gross motor skills rather than cognitive. According to Singer (1975), a great disparity may exist because students react differently to various teaching methods when learning motor skills.
Experimental Conditions
Design The design for this study was a pretest-posttest configuration; two experimental task sequences were analyzed for skill acquisition. No control group was used. This design is useful for observing the comparative effect of two treatments (Thorpe, 1986). To control for order effects, the group taught first each day was alternated between the two conditions (Hersen & Barlow, 1976; Kazdin & Hartmann, 1978).
Physical Environment. All teaching and evaluation sessions were conducted in a 30 ft by 20 ft room with alOft (3m) ceiling. Access to the room was through one set of double doors that had a small window, through which sessions could be observed from an adjoining room. The room was equipped with three bicycle ergometers, one minitrampoline, and a storage box filled with balls of various sizes.
78
September 1992
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
Instructional Conditions. A four-teachers-tosix-students format was maintained for all sessions. The teachers worked with four students on a one-to-one basis while the other two students were riding the bicycle ergometers. During the course of each class session, each teacher normally worked at least once with each child so that the child did not become dependent on one specific teacher. The instructors for this study were two females and two males ranging in age from 23 to 42. All four were adapted physical education aides with a minimum of 8 months of teaching experience in adapted physical education. Each instructor had completed extensive inservice training in adapted physical education under the supervision of an adapted physical education specialist. In addition, as suggested by Dunlap (1984) and Koegel, Russo, and Rincover (1977), each had completed inservice training courses in operant conditioning and had been trained to an 80% criterion level in the use of behavior modification for.children with autism by the Director of Special Education. All four worked with both treatment groups. Regular assessments of on-task behavior for all the instructors were conducted, as recommended by Dunlap (1984) and Dunlap and Koegel (1980a). These procedures had been adopted as a part of the standard operating code at this school; however, for the present study, these assessments were conducted twice per week rather than once per month. Similar to Dunlap (1984), the teachers were expected to maintain at least an 80% performance level of ontask activity throughout all their teaching for all tasks under both conditions. Because all four teachers conducted instructional periods with both treatment groups, the potential for teacher bias was a concern. To control for this potential effect, as recommended by Drew (1980), the teachers were not aware of the experimental hypothesis. The teachers were told that the students had been grouped and assigned to their respected treatment classes to address their individual needs for varying instructional methods as suggested by their test scores. Instructional Procedures. The instructional procedures used in this study were similar to those used by Dunlap (1984) and Dunlap and Koegel (1980b). During the study, the standard teaching techniques of physical shaping, physical assistance, physical prompting, and social reinforce-
ment were applied (Fait, 1960; French & Jansma 1982; Sherrill, 1986).These techniques were provided and then were slowly eliminated until the students responded successfully on their own. If a student failed to respond successfully after three consecutive trials, the teacher went back to the next lower level in the teaching techniques used. All correct responses, whether with physical shaping, physical assistance, prompting, or independent action, were reinforced with social reinforcement and a "pat on the back." Similar to Dunlap (1984), incorrect responses resulted in the teacher's saying "No _ " (using the subject's name) and withdrawing attention. An off-task or disruptive behavior was treated according to the procedural guidelines suggested by Arnheim, Auxter, and Crowe (1969). These procedures had been approved by the school's interdisciplinary team. The procedures routinely consisted of planned ignoring, signal interference, proximity control, or brief physical restraint (Arnheim et aI., 1969). Instructional sessions were held for a 45-miIi duration for both conditions. The performance criterion of 9 correct independent responses out of 10 consecutive trials of the task, which Dunlap (1984) used, was selected as the successful performance criterion for ending training. This criterion applied to all experimental tasks in both conditions. However, throughout the training period and during instruction, these criteria were never met. Classes for all students were conducted 5 days per week. A minimum of 15 min was provided between the two classes so that teachers could arrange equipment, use the bathroom, organize data sheets, and allow students to move to and from classes. Independent Variables The question addressed in this experiment is whether the distributed practice technique of task variation (with maintenance tasks interspersed) is more effective than a constant task condition (commonly referred to as massed practice) when teaching gross motor skills to people with autism. Six experimental acquisition tasks were randomly presented under both conditions. In addition, the task-variation group practiced the extra-experimental tasks of run, catch, and hop, which they had previously learned but were maintaining. A complete description of each of the experimental and extra-experimental tasks
Exceptional Children
79
Downloaded from ecx.sagepub.com at TUFTS UNIV on December 1, 2015
FIGUREl Class Performance Score Sheet for the Six Basic Gross Motor Skills
Continuous Bounce FOCAlPOINTS
SCORING
STD
Au_mont:
~:~::r-
R_...ment: II • Achieved f). Not
"""i_
Underhand Roll FOCAL POINTS
SCORING Aa.... ment: X.Ac/li_ O.NoIachievod
"-
b R_.oIB..
R.......ment: II.Adliovod f). NoIochievod
-. -,
STD
• Pondu,", Arm Swing
• El'OO on Torgot
10ft.
"-,
b Pondu,", Arm Swinw
• Sh""ld_ "Squarod"
""-
2Oft.laIgol II 15 ft..213
"-
b PondulorArmMotion
"
c Palm F.ang Diroction d _
