Teaching On-Task Rollerblading and Ice-Skating to a Child with Autism

Abstract

The present study used a multi-component intervention package to teach on-task rollerblading and ice-skating to a boy with autism. Intervention consisted of response prompts, stimulus prompts, multiple-exemplar training, and a conditioned reinforcement system. The participant learned to remain on-task while rollerblading in a circular route marked by cones for up to 26 min. Both stimulus and response generalization of skating were demonstrated in a variety of non-training settings, including ice-skating at a rink.

Participation in physical leisure activities may be challenging for individuals diagnosed with autism due to deficits in social skills, motor skills, generalization, and self-monitoring (McDonnell, Johnson, Polychronis, & Risen, 2002; Todd & Reid, 2006). In general, few studies have evaluated procedures for teaching athletic skills to individuals with developmental disabilities, with most of the literature focusing on sedentary activities (e.g., Internet skills) or intermediate community living skills (e.g., dart throwing), rather than sports (Lagomarcino, Reid, & Ivancic, 1984; Schleien, Wehman, & Kiernan, 1981). Sports that have been taught to children with autism include indoor rock climbing, swimming, dancing, gymnastics, soccer, and cooperative games (e.g., Kaplan-Reimer, Sidener, Reeve, & Sidener, 2010; Luyben, Funk, Morgan, Clark, & Delulio, 1986; O’Conner & Cuvo, 1989; Rapp, Vollmer, & Hovanetz, 2005).

Physical activities that do not require collaborative team play may provide an opportunity to practice skills without the challenge of potentially demanding social interactions (e.g., indoor rock climbing, yoga, karate, running, swimming, and rollerblading/skating). Ice-skating has advantages of allowing for parallel engagement with peers and family members of all ages and having relatively uncomplicated rules (e.g., propel forward, follow a designated route). A crowd of people moving in a particular direction serves as a discriminative stimulus for on-task skating. In many locations, roller rinks and ice rinks offer a year round, consistent schedule of public sessions and are popular sites for birthday parties.

To date, no previous studies have evaluated teaching ice-skating to children with autism. The present study used a multi-component intervention package to teach on-task skating to a boy with autism. Initially, the goal was to teach ice-skating; however, it was decided that teaching ice-skating at a rink could be financially prohibitive and potentially stigmatizing. Therefore, rather than teaching on-task ice-skating directly, we taught on-task rollerblading using multiple exemplars of characteristics of an ice arena, including skate route, size, and direction, noise, and the presence of other people. Only generalization probes were conducted in the ice arena. We hypothesized that rollerblading might generalize to ice-skating due to the similarities between the two activities. Although there are some differences (e.g., stopping), these differences are minimally relevant to the skills targeted in this study (http://healthyliving.azcentral.com/rollerblading-vs-ice-skating-3375.html). The intervention consisted of response prompts, stimulus prompts, multiple exemplar training, and a conditioned reinforcement system.

Method

Participant, Settings, and Materials

The participant was Colby, a 6-year and 2-month-old boy diagnosed with autism. Prior to the study, he could propel himself forward independently for 15 m on both rollerblades and ice skates with no more than one instance of falling. Colby did not demonstrate a preference for skating, but his parents reported they wanted him to learn so their family could go ice-skating together. Rollerblading sessions were conducted in an indoor gymnasium at Colby’s school and in several local parks that were otherwise unoccupied during sessions. Ice-skating generalization sessions took place in two indoor rinks during public sessions. Sessions were conducted three to five times a week for approximately 30 to 60 min at different times throughout the day. We selected this duration after observing that people at rinks typically skated this long with periodic breaks.

Large, medium, and small orange cones marked the skating route during training sessions. An iPod® with speaker was used to play music. A stainless steel cup was used to deliver snacks. Rollerblades/ice skates, a safety helmet, and knee/elbow pads were worn by the participant during all skate trials.

Delayed Reinforcement Training

Prior to experimental sessions, a conditioned reinforcement system was taught that would allow the experimenter to reinforce on-task skating without interrupting skating. Contingent upon mastered responses from Colby’s school program, snacks were dropped into a stainless steel cup to establish the distinct sound as a conditioned reinforcer. This was done until Colby waited, without reaching for the cup, for delivery of as many snacks as each phase of ice-skating teaching would require for a duration of up to 3 min (1 to 15 snacks).

Experimental Procedures

Baseline Probes

Pre- and post-training baseline probes were conducted by arranging a rollerblading environment to simulate an ice-skating rink. Eight large cones were arranged in an oval the same size as at the ice-skating rink. Colby was assisted with putting on and safely securing the rollerblades or ice skates, helmet, knee pads, and elbow pads. The experimenter then led him to the skating surface, said, “skate with me,” provided a model to begin skating in a particular direction for 3–5 s, and then stood just outside the skating area; there were no prompts or programmed consequences for skating. A baseline probe was also conducted before beginning each new teaching phase to determine whether additional teaching was necessary and to assess potential stimulus control from the experimenter skating behind Colby during teaching. At the end of each session, the experimenter examined Colby’s feet to determine there were no sores (there never were).

Teaching Phases

Following baseline, Colby was exposed to a sequence of phases in which response prompts and stimulus prompts were gradually faded, and the schedule of reinforcement was gradually thinned. Prior to each skate trial, a stimulus preference assessment was conducted during which six caregiver-reported preferred snacks were presented to the participant. During the skate trial, the selected snack was delivered into the cup as per the criterion for that phase. Praise and access to the cup were provided following every skate trial.

At the beginning of skate trials during teaching phases, after the experimenter skated to model the direction, she skated behind and slightly to the left or right of Colby. The distance and side changed as the response prompts were faded and the directional flow changed. Response prompts were faded using a most-to-least hierarchy across eight levels, beginning with experimenter’s hands on Colby’s shoulders (Level 1), touching Colby’s shoulder with index finger every 10 s (Level 2), touching Colby’s shoulder with index finger every 20 s (Level 3), and skating 1 ft behind Colby (Level 4). For the remaining levels, the experimenter’s distance was faded to 2 ft (Level 5), 4 ft (Level 6), 6 ft (Level 7), and 8 ft (Level 8). The requirement for moving to the next level was skating a single trial to criterion at the current level. If the criterion for a level was not met during a trial, the level was repeated. If that trial was not skated to criterion, Colby moved back to a lower level.

Table 1 shows the stimulus prompts and schedule of reinforcement used during each phase. Phases were presented in a “probe ahead” fashion (LeBlanc, Hagopian, Maglieri, & Poling, 2002): upon reaching criterion, a probe two phases higher was conducted. When the final phase was mastered, baseline probes were conducted in which Colby could skate without being stopped at 3 min; these sessions were terminated if an error was made. Eight additional baseline probes were conducted to demonstrate maintenance.

Table 1.

Stimulus fading and reinforcer thinning across phases

Phase	Number of medium cones	Number of small cones	Schedule of reinforcement (s)
1	24	28	10
2	22	24	20
3	20	20	30
4	16	14	60
5	12	8	90
6	8	2	120
7	4	0	150
8	0	0	180

Generalization

Because ice-skating could not be taught at the ice arena, generalization was programmed for by identifying features of the ice arena that varied across days according to pre-experimental observations and conversations with staff. These included musical genre (e.g., rock, classical, no music), music volume (e.g., 80, 60, 40 db), the presence of people (e.g., center, sidelines, no people), and direction of skating (i.e., clockwise, counterclockwise) across sessions. Generalization across settings was also programmed for by conducting teaching sessions in several environments, including a school gym, a tennis court, and a basketball court at a local park. Generalization probes were conducted under baseline conditions at two ice rinks and a novel park.

Social Validity

To assess the social validity of the intervention, three different surveys using Likert-type scales (1 = lowest; 5 = highest) were administered to Colby’s parents, 12 parents of children with autism who did not ice skate, and seven ice arena staff. Colby’s parents answered six questions about the goals, procedures, and outcomes. Other parents answered six questions about their interest in skating, activities in the community, activities with family and peers, and participation in a similar program. Skate instructors and guards at the ice arena watched four videos of four boys (Colby and three typically developing boys) ice-skating and answered three questions (see Table 2). The typically developing boys were selected from a skate class offered at the arena. Raters were blind to the purpose of the study.

Table 2.

Video comparison questionnaire: ice arena staff

Question	Bruce (typical)		Alan (typical)		Josh (typical)		Colby
	M	R	M	R	M	R	M	R
This child is skating on-task as per the definition.	4.9	4–5	4.7	4–5	4.5	3–5	4.8	4–5
This child’s behavior is appropriate for the situation.	4.7	4–5	4.8	4–5	4.7	4–5	4.3	4–5
This child is engaging in an activity that’s appropriate for his age and gender.	4.8	4–5	4.8	4–5	4.7	4–5	4.8	4–5

M mean, R range

Measurement

Data were collected on duration (minutes) of on-task skating, defined as independently skating forward in the direction modeled (either clockwise or counterclockwise) by maintaining a circular, unidirectional pattern of movement while maneuvering around cones, and maintaining a distance of 1–19 m from the cones without making errors after the experimenter said, “skate with me” and modeled the direction to skate. Possible topographies of error included attempts to leave the skating area, stopping for 5 s or longer, skating into the circle of cones, changing skating direction, or bumping cones.

Sessions consisted of multiple increments referred to as skate trials. The criterion for an on-task skate trial was three consecutive minutes on-task skating per trial. When an error occurred, the skate trial ended, the timer was stopped, and the total duration of on-task skating for that skate trial was recorded. The time was rounded down to the nearest 30 s to allow for a conservative measure of duration. For example, if a participant had been skating for 2 min and 40 s and then made an error, then the total duration of on-task skating was scored as 2 min and 30 s.

Secondary data on minutes of on-task skating per skate trial were collected by independent observers. Interobserver agreement (IOA) was calculated by dividing the lower score by the higher score and multiplying by 100 %. IOA was assessed during 82 % of randomly selected skate trials and was 100 %. Procedural integrity was measured during 55 % of randomly selected skate trials on correct presentation of stimuli, use of prompts, and delivery of reinforcement and summarized as percentage of correctly completed components per skate trial (M = 98 %; range = 82–100 %). IOA data on procedural integrity were collected during 45 % of procedural integrity skate trials (M = 99 %; range = 92–100 %).

Results and Discussion

The present study evaluated a multi-component intervention package to teach on-task skating to a boy diagnosed with autism. Colby learned to rollerblade for 3 min across a variety of environmental stimuli, including different settings, music genres, music volume, and skating direction, and ice skate at an arena in a relatively short amount of time (i.e., 14 days; 57 skate trials; Fig. 1).

Fig. 1.

Duration of on-task skating across phases for Colby. Response prompt level is depicted with a dash and is measured on the right y axis

The secondary y axis and horizontal bars depict the response prompts that were systematically decreased throughout Phase 1. Colby skated without response prompts in every skate trial after Phase 1. As baseline probes were conducted after each phase, the duration of on-task skating gradually increased, but did not reach the criterion until the final stimulus fading phase. Colby met the criterion when Phases 4, 6, and 8 were probed ahead. Although Colby met the criterion with all prompts faded during Phase 8, during the next reversal to baseline, he skated only 2 min. It is possible that his performance was different during this session because a choice of snacks was offered prior to teaching trials but not baseline probes. However, during the next two baseline probes, he skated at the mastery criterion.

Following mastery, when given an opportunity to skate without interruption, Colby continued to skate for 26 and 14 min. Without being taught to skate in an ice rink, his rollerblading generalized to ice-skating in three out of four skate trials. During generalization probes at a novel park, Colby met the criterion during both skate trials. During maintenance probes, Colby met the criterion for during all eight skate trials.

Although the design employed does not control for all threats to internal validity, functional control of the intervention is supported by Colby not meeting the criterion during probes that acted as reversals to baseline (i.e., eight large cones, no prompts or programmed reinforcement). Reversals to baseline after Phases 2, 4, and 6 demonstrated that, although skill acquisition was occurring, further teaching was necessary. Future studies might report data on types of errors that occur (e.g., leaving the area, stopping, skating into cones, changing skating direction).

The questionnaires indicated a high degree of social validity for this intervention and the resulting skill level. Colby’s parents rated the goals, procedures, and outcomes of the study with an overall mean of 4.2 (range, 3–5). Anecdotally, they reported that Colby had skated several times since the study ended, had been skating around the cones with consistency, and appeared to enjoy skating. Other parents of children with autism rated their interest in the goals and procedures of this study with an overall mean of 4.8 (range, 3–5). One respondent rated community participation with a 3 (neutral) and commented that the reason for this was “fear.” On the survey comparing video of the participant and typical peers, results for Colby were similar to that of the typical peers, with Colby sometimes scoring higher than his peers (see Table 2). It is possible that this is because two of the peers skated more slowly than Colby. Future studies might ask respondents to rate both pre-intervention and post-intervention videos.

This study adds to the literature on teaching physical leisure skills, specifically skating, to children with autism, which had not been previously investigated. Skating may be an ideal pastime because it provides exercise, an opportunity to engage in activity parallel with others in a community setting, and uncomplicated social rules. Additionally, these outcomes add to the limited research on stimulus fading for individuals with autism (e.g., Schreibman, 1975) and demonstrate a method for using and fading stimulus and response prompts together systematically, which may be useful in teaching other types of skills (e.g., self-help), as well.

Compliance with Ethical Standards

This research received no direct grants or funding.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.

Conflict of Interest

The authors declare that they have no conflict of interest.