Examination of the Effects of Video Prompting Across Different Types of Tasks

Abstract

Video prompting has a strong evidence base as an effective strategy to teach students with severe disabilities a variety of skills including daily living skills. This study examined whether differences in daily living tasks (i.e., multistep, multicomponent, and sequential tasks) would impact skill acquisition using video prompting by three participants with severe to profound disabilities. Results indicated that although video prompting was effective broadly, aspects of task type taught may affect the pace of an individual’s acquisition. Implications for future research and practice are discussed.

Keywords

video prompting intellectual disability severe disabilities daily living skills

Adults with severe disabilities tend to have poorer long-term outcomes in employment, community involvement, and independent living than their peers without disabilities (Newman, Wagner, Cameto, Knokey, & Shaver, 2010). However, teaching students with severe disabilities to engage in daily living, vocational, and leisure skills can increase functional independence as well as opportunities for participation in the community (Mechling, 2007). The Council for Exceptional Children (Council for Exceptional Children & CEEDAR Center, 2017) recognized the use of assistive technology, like video prompting, in instruction as one effective strategy for teaching young adults with severe disabilities these skills. This intervention involves the student viewing a short video clip of one step of a task, then having the opportunity to complete that portion of the task before watching the next video clip. Video prompting breaks a large task down into discrete steps for students to follow, rather than a video modeling approach that presents a longer video of the whole task start to finish. Video prompting has been used to teach a variety of skills related to activities of daily living (e.g., Cannella-Malone et al., 2011; Cannella-Malone et al., 2006; Sigafoos et al., 2007; Wu, Cannella-Malone, Wheaton, & Tullis, 2014), vocational skills (e.g., Mechling, Ayres, Purrazzella, & Purrazzella, 2014; Van Laarhoven, Johnson, Van Laarhoven-Meyers, Grider, & Grider, 2009), and leisure skills (Cannella-Malone et al., 2015; Chan, Lambdin, Van Laarhoven, & Johnson, 2013).

In their review of 18 studies evaluating video prompting, Banda, Dogoe, and Matuszny (2011) concluded that video prompting is an effective intervention for improving skill performance of a range of skills and a variety of participants with intellectual and developmental disabilities. In addition, skills learned through video prompting maintained over time and generalized to novel settings or behaviors. The authors suggested several possible reasons for the success of video prompting as an intervention, including consistency in presentation of the prompt; incorporation of other effective strategies including task analysis, prompting, repetition, and feedback; presentation of video clips in multiple modes; and inclusion of repeated practice. Even though video prompting is effective, several aspects of the intervention are still being analyzed to determine the most effective and efficient way to use it (Banda et al., 2011).

In a recent study on the use of continuous video modeling (presenting an audiovisual depiction of a whole task in a continuous loop, whereas video prompting breaks down the depiction into individual prompted steps), Mechling and colleagues (2014) suggested one area for such exploration is the type of task being taught. They taught four 29- to 35-year-old males with Down syndrome and moderate intellectual disabilities to complete three multicomponent tasks using video modeling. The authors noted that the tasks taught (i.e., sorting recycling containers, setting a buffet table, folding towels) were repetitive in nature, which could be partially responsible for the positive results. Acquisition and performance on tasks that involve a chain of behaviors (e.g., changing a tire) may differ from repetitive tasks. Providing repetition and multiple opportunities to perform the steps of a task supports skill acquisition for students with severe to profound disabilities (Browder & Spooner, 2011). Given that, it is possible that students may acquire repetitive skills more quickly than skills involving different steps. To date, however, research investigating the effectiveness of video prompting has not evaluated whether different types of tasks will impact students’ acquisition of those tasks.

Significant differences in students’ acquisition related to the complexity of tasks—specifically repetitive versus nonrepetitive steps—could have implications for research and practice. For example, when first introducing a student to video prompting, teachers may want to focus on tasks that are easier to acquire and save more complex tasks for when the student is more familiar with video prompting. Likewise, research studies using multiple baseline designs may need to consider the complexity of tasks when planning to provide equivalent baselines. For example, a study exploring the effectiveness of video prompting to teach washing a window, washing a table, and tying a tie may not show experimental control if the student quickly acquires the repetitive skills of washing the window and table but does not acquire the nonrepetitive skill of tying a tie. Alternatively, if there is no appreciable difference in performance between different types of tasks, then teachers and researchers may not need to consider this aspect when planning and evaluating interventions. Therefore, the purpose of this study was to determine whether students with severe to profound intellectual disabilities would differentially acquire different types of skills using video prompting with error correction.

Method

Participants

Three students with severe to profound intellectual disability participated in this study. All three participated in the statewide alternate assessment for students with significant cognitive disabilities. We completed the Vineland Adaptive Behavior Scales (Sparrow, Balla, & Cicchetti, 1984) for each participant. Results indicated that all participants had significant deficits in all adaptive living skills domains, which included personal skills (e.g., self-help, hygiene), domestic skills (e.g., cooking, laundry), and community skills (e.g., using money, safety skills). Results also indicated that all participants had significant deficits in all communication domains, which included expressive language, receptive language, and written communication. None of the participants used a reliable or systematic method of communication. All three participants were accustomed to interacting with a digital device (e.g., iPod) and attending to video as leisure activities, but none had engaged with video prompting or modeling before this study.

Andrew was a 20-year-old African American male diagnosed with organic brain syndrome, profound intellectual disability, and congenital atresia, which resulted in a profound hearing impairment. On the Vineland, his age equivalence was 1 year 2 months in the Communication subdomain, 4 years 2 months in the Daily Living Skills subdomain, 1 year 9 months in the Social subdomain, and 2 years 4 months in the Adaptive Behavior Composite. Andrew primarily communicated through a limited sign repertoire (i.e., approximately 15 recognizable signs for basic messages such as “all done” to indicate he was finished with an activity), pointing, and gesturing. When engaged in work tasks, Andrew often attended well and worked independently but occasionally struggled with some challenging behavior (e.g., work refusal, minor physical aggression).

Peter was a 20-year-old White male diagnosed with autism, intellectual disability, and anxiety disorder NOS (not otherwise specified). On the Vineland, his age equivalence was 1 year 2 months in the Communication subdomain, 2 years 5 months in the Daily Living Skills subdomain, 10 months in the Social subdomain, and 1 year 5 months in the Adaptive Behavior Composite. Peter primarily communicated through pointing or gesturing. When engaged in work tasks, Peter frequently struggled with prompt dependence and at times could wait indefinitely for a prompt.

Shelly was an 18-year-old White female diagnosed with autism, severe intellectual disability, schizophrenia, and bipolar disorder. On the Vineland, her age equivalence was 1 year 6 months in the Communication subdomain, 2 years 9 months in the Daily Living Skills subdomain, 1 year 1 month in the Social subdomain, and 1 year 9 months in the Adaptive Behavior Composite. Shelly primarily communicated through gestures and unintelligible vocalizations that signaled her affect. When engaged with a work tasks, Shelly often struggled with challenging behavior such as work refusal, planting on the ground and refusing to move, and minor verbal aggression.

Setting

This study was conducted in an urban self-contained school that served students with moderate to profound intellectual, developmental, and physical disabilities between the ages of 5 and 22 years. The typical student–teacher ratio in the school was approximately 8:3. Sessions for all three tasks were conducted in the school’s home economics room or designated research classroom. The home economics room was broken into four areas: living room, dining room, kitchen, and general space for a walkway and teacher planning. In the home economics room, loading the dishwasher and making lemonade were completed in the kitchen area, and folding shirts was completed on the dining room table. The research classroom was a classroom used by graduate students who conducted research at the school, and it contained three large tables, chairs, and a sink. Making lemonade sessions occurred at the sink, and folding shirts occurred at one of the tables. Loading the dishwasher sessions only occurred in the home economics classroom. Other than the interventionist and data collector, no other individuals were present for sessions during any phase of the study.

Tasks and Materials

The three types of tasks examined in this study were multistep, multicomponent, and sequential. A multistep task was defined as completing a single task made up of many, discrete steps. For example, changing a tire and tying a tie are examples of multistep tasks. The multistep task selected for this study was making lemonade, which required 15 unique, nonrepeated steps. A multicomponent task was defined as a single, multistep task completed repetitively. For example, stacking up chairs from around a table is a multicomponent task. The multicomponent task selected for this study was folding long-sleeve T-shirts, which required repeating the smaller (i.e., eight-step) discrete task three times. Given that this task repeated the same component, we chose to include folding three long-sleeve T-shirts so that the steps in the task analysis and the video length were similar to the other tasks. Finally, a sequential task was defined as completing a group of tasks that make up a larger task. For example, putting away the groceries or sorting clean silverware into a drawer is a sequential task. The sequential task selected for this study was loading dishes into a dishwasher, which requires three distinct subtasks (i.e., five steps of loading the top rack with one cup and one spoon, five steps of loading the bottom rack with one plate and one bowl, and 10 steps of loading the soap and starting the cycle). Assessment results indicated that each of the students struggled with domestic skills like those included in this study. In collaboration with the teachers, we chose skills that students had not received instruction at the time the study began.

Materials for making lemonade included a cup, a 30 oz container of powdered lemonade with a screw top lid, and a spoon. Materials for folding shirts included three long-sleeve T-shirts and a basket. Materials for loading the dishwasher included dishes (i.e., spoon, cup, bowl), liquid dishwasher soap, and two different dishwashers—a new dishwasher was installed in the home economics room over the summer break.

A third-generation Apple iPod Touch was used as the prompting device in this study, and the video prompts were presented using the device’s preinstalled video application. The video clips for each task were developed beforehand using functionally similar materials as those used during training. The individual steps were filmed using a digital video camera and edited using iMovie. Video clips were filmed from two perspectives depending on the task. Task steps were filmed from the perspective of a spectator when the step required a gross motor skill (e.g., taking a shirt out of the basket, opening the dishwasher), and the student saw another person completing the step of the task. Task steps were filmed from the perspective of the student when a when the step required a fine motor movement (e.g., pushing a specific button on the dishwasher), and the student saw the performer’s hand completing the step. No text appeared on the screen with the videos; however, at the beginning of each clip, a female voice provided an auditory prompt stating what the student was to do in that step. For example, for the first step of loading the dishwasher, the student heard, “First, open the dishwasher,” and saw the performer in the video open the dishwasher. For making lemonade, the video clips were an average of 5 s (range: 2–11 s). For folding shirts, the video clips were an average of 5 s (range: 3–9 s). For the loading the dishwasher task, video clips were an average of 6 s (range: 3–11 s). Table 1 provides the task analyses for each task, the duration of each video clip, and the total duration of the video.

Table 1.

Task Analyses Including Duration of Each Video Clip and Total Duration.

Making lemonade	Time (s)	Folding shirts	Time (s)	Loading dishwasher	Time (s)
Unscrew lid and place on table	6	Take first shirt out of basket	4	Open dishwasher	8
Pick up spoon	2	Place shirt face down on table	5	Pull out bottom rack	5
Scoop one spoonful of powder	4	Flatten out shirt	7	Load plate into rack	7
Pour powder in cup	4	Fold left sleeve in	4	Load utensil into holder	6
Scoop another spoonful of powder	4	Fold right sleeve in	3	Push in bottom rack	5
Pour powder into cup	3	Fold bottom of shirt up to collar	4	Pull out top rack	5
Put spoon on table	3	Fold in half vertically	4	Load cup into rack	6
Turn on cold water	8	Place shirt to side	3	Load bowl into rack	6
Fill glass with water	11	Repeat for remaining two shirts		Push in top rack	5
Place cup on table	3			Close dishwasher	4
Turn off water	5			Get soap out of cabinet	10
Pick up spoon and put in cup	4			Open soap	4
Stir lemonade with spoon	6			Open dishwasher	5
Remove spoon and place in sink	6			Fill soap tray	11
Screw lid onto container	6			Close soap tray	4
				Close dishwasher	5
				Close soap	3
				Put soap in cabinet	9
				Start dishwasher	8
Total clip duration	129		102		116

Dependent Measures and Data Collection

Correct performance of each of the target tasks was the primary dependent measure. For each task, we collected data on whether each step of the task was completed correctly (with or without video prompting) or incorrectly. A correct response was defined as the student accurately completing a step within 20 s with or without first viewing the video prompt. Incorrect responses were defined as the student either correctly or incorrectly completing a step following a second viewing of the video. The percentage of steps completed correctly was calculated by dividing the number of steps completed correctly by the total number of steps and multiplying by 100. We also collected data on the total duration of each session and the number of sessions needed to reach mastery, defined as at least three consecutive sessions with 100% correct responding with or without the video prompt.

Experimental Design

A multiple probe across participants design (Gast & Ledford, 2014) was used to demonstrate a functional relation between video prompting and change in student behavior for each task. The phases were baseline, intervention, and maintenance. To move from baseline to intervention, we required at least three data points that were stable or had a decreasing trend. To move the next participant from baseline into intervention, the student in intervention had to demonstrate a clearly increasing trend over at least three data points. Students were placed in maintenance once they achieved mastery.

Interobserver Agreement

Interobserver agreement (IOA) data were collected by independent observers during an average of 37% (range = 25%–42%) of sessions across conditions, tasks, and students. The observer was positioned behind and away from where the participant was working, as to not distract the participant. Independent observers were trained by the first author, who explained the task analysis for each task, explained how to complete the data sheets, and provided examples of correct and incorrect responses. IOA was calculated on a session-by-session basis by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. Mean IOA was calculated to be 97% (range = 95%–99%) for folding shirts, 99% (range = 98%–99%) for making lemonade, and 99% (range = 98%–100%) for loading the dishwasher.

Procedures

The procedures for all three tasks were identical and are described in the sections below.

Baseline phase

During baseline sessions, students were individually brought to the setting for the task. Using a multiple opportunity method, we asked the students to complete the task (e.g., “Make a glass of lemonade.”). If they did not start the first step within 5 s or started it incorrectly, we blocked their response and turned them away from the task by physically prompting the students to turn around or asking them to look at something on the opposite side of the room. When the participant was faced away, we completed that step. We then turned them back to the task and said, “Keep going.” In this way, they had an opportunity to attempt each step in the task. Upon completion of the task, they were thanked for participating regardless of performance.

Intervention phase: Video prompting with error correction

In the intervention phase, we used a video prompting with error correction procedure consistent with previous video prompting research (e.g., Cannella-Malone et al., 2015; Sigafoos et al., 2007). Students were brought individually to the appropriate location for the targeted task and provided video prompting with error correction. The student was positioned in front of the materials for the task, and a timer was started to track session duration. The trainer held the iPod Touch in front of the student so that he or she could easily see the screen, said, “Watch this,” and played the video clip of the first step. When the clip ended, the trainer said, “Now you do it.” The student had 20 s to complete the step. If the student did not initiate the step within 5 s or began to complete the step incorrectly, the trainer blocked the response and said, “Not quite right. Here, watch this again.” The trainer then played the video clip a second time. If the student did not respond correctly following the second showing of the video clip, the researcher completed the step with the participant watching. If the second viewing of the video clip followed by a model did not evoke a correct response across three consecutive sessions, three hierarchical most-to-least prompts were provided following the first error rather than reviewing the video. The first prompt was a full physical prompt. Using this prompt, the experimenter repeated the auditory prompt given in the video while physically guiding the student by the hand to complete the step. Following three sessions with full physical prompts, the prompt was reduced to a partial physical prompt at the elbow in which the experimenter repeated the auditory prompt given in the video while physically guiding the student at the elbow to complete the step. After three sessions with the partial physical prompt at the elbow, the prompt level was reduced to a partial physical prompt at the shoulder in which the experimenter repeated the auditory prompt given in the video while providing a light tap at the student’s shoulder. After three sessions with the partial physical prompt at the shoulder, the prompt level was reduced to a gesture prompt in which the experimenter repeated the auditory prompt given in the video while pointing to the start of the step.

Once the full physical prompt was faded, if the student made an error during a session, the prompt was immediately increased back to the full physical prompt during that session. In the next session, the prompt level was increased by one level. In other words, if the student was receiving a prompt at the shoulder, a prompt at the elbow was provided in the next session. All error correction decisions were made on a step-by-step basis. Therefore, if a student was unsuccessful with one step in the entire task, only that step received the tiered error correction. Upon completion of the task, the timer was stopped, and the student was given a preferred item and thanked for participating regardless of performance. For the multistep task, students were allowed to consume the cup of lemonade they made during the session. For the other tasks, preferred items included verbal praise, high fives, or watching a short portion of a preferred movie. Preferred items were specific to each individual participant.

Return to baseline with loading the dishwasher

During the summer break, a new dishwasher was installed in the home economics classroom. Although the steps in the task analyses were identical for both dishwashers, there were qualitative differences between the two machines (e.g., the first dishwasher had square protruding buttons, where the second dishwasher had round buttons that were flush with the surface). As such, we updated the videos to show the new dishwasher and collected new baseline data using the procedures described above.

In vivo instruction

For several tasks, Shelly and Peter did not master the skill with video prompting with error correction. The decision to move to in vivo instruction was made when they stopped making progress toward mastery and attempts to fade the error correction procedures were unsuccessful. As such, a least-to-most instructional procedure was implemented in an attempt to help them master the skills. Using these procedures, they were provided the same vocal prompt as in the video prompts, and if they did not respond within 5 s or began to respond incorrectly, the instructor blocked the incorrect response, repeated the verbal prompt and paired it with a gesture. This process was repeated while increasing the prompt level (i.e., pairing a verbal prompt with a model then a full physical prompt) until the student responded correctly. When the student completed the task, the timer was stopped, and the student received the same preferred item as in the video prompting with error correction phase regardless of performance.

Additional procedures

Two additional procedures were used. First, Peter struggled to mix the lemonade with the spoon (i.e., he would hold the spoon above the glass, barely touching the lemonade, and swing the spoon like a pendulum). Beginning in Session 79, Peter began massed practice trials stirring the lemonade with a spoon prior to the experimental session. Graduated guidance was used to support his practice of this one step. The instructor paired verbal prompts with a model of what he was to do and provided physical guidance only as needed.

In the second year of this study, a new dishwasher was installed at the school. This dishwasher required the students to press a specific set of buttons to start it. When the data indicated that they were struggling with the step of pushing the start button on the machine, we added a stimulus prompt to the dishwasher (i.e., a blue circular sticker), which remained in place through the end of the study.

Maintenance phase

After students reached mastery, a return to baseline was implemented. If a student’s performance dropped below 80% correct, video prompting was reintroduced until the student again demonstrated mastery. Maintenance data were collected during each session until the student demonstrated 100% performance, then they were collected once per week for 6 weeks, then once every 2 weeks for the duration of the study.

Procedural integrity

Independent observers collected data on the interventionist’s procedural integrity during an average of 33% (range = 21%–43%) of the sessions for all students across all phases and tasks. The procedures were listed sequentially on a checklist and a secondary observer marked off which steps were completed correctly or incorrectly. Over the course of the study, the interventions’ implementation of wait times before error correction drifted slightly resulting in one session with a procedural integrity score of 79%. When procedural integrity dropped below 80%, the researcher implementing the intervention was retrained on the procedures and subsequent procedural integrity remained high. Procedural integrity was calculated by dividing the number of steps completed correctly by the total number of steps and multiplying by 100 and was calculated to be 99% (range = 79%–100%).

Results

Three multiple baseline experiments were conducted across three different types of tasks with three participants, allowing for nine opportunities to demonstrate effects. Through visual analysis of the data (see Figures 1 –3), we identified eight effects for video prompting with error correction out of nine opportunities. To further compare the impact of task difficulty on the participants’ acquisition, in Table 2, we provide means and ranges for session number and duration while in the video prompting intervention phase. Below, we describe visual analysis of all opportunities by focusing on the trend, level, and variability of the data.

Figure 1.

Percentage correct of multistep task (making lemonade) completion for Shelly (top tier), Peter (middle tier), and Andrew (bottom tier).

Figure 2.

Percentage correct of multicomponent task (folding shirts) completion for Shelly (top tier), Peter (middle tier), and Andrew (bottom tier).

Figure 3.

Percentage correct of sequential task (loading the dishwasher) completion for Peter (top tier), Shelly (middle tier), and Andrew (bottom tier).

Table 2.

Number of Sessions and Session Duration (Mean and Range) to Mastery, Move to in Vivo, or Study Termination Across Tasks.

Participant	Tasks
	Multistep (making lemonade)		Multicomponent (folding shirt)		Sequential (loading dishwasher)
	Sessions	Minute per session (range)	Sessions	Minute per session (range)	Sessions	Minute per session (range)
Andrew	26 sessions to mastery	1.6 (0.7–4.9)	20 sessions to mastery	4.2 (2.7–7.3)	35 sessions to mastery	3.6 (1.9–5.8)
Shelly	64 sessions to mastery	3.2 (2–7.9)	35 sessions to in vivo	8.6 (3.4–18.7)	60 sessions to end of study	6.1 (3.1–16.4)
Peter	101 sessions to end of study	4.2 (1.9–8.8)	27 sessions to in vivo	14.8 (10.4–17.4)	71 sessions to in vivo	9.5 (6–16.1)

Multistep Task (Making Lemonade)

Shelly

Shelly’s data for making lemonade are presented in the top tier of Figure 1. In baseline for making lemonade, Shelly completed an average of 26% (range = 20%–27%) of the steps correctly. When video prompting with error correction was implemented, she completed an average of 85% (range = 33%–100%) of the steps correctly and mastered the skill in 65 sessions. She maintained the skill for six sessions (range = 87%–93%). In the seventh session, her performance dropped to 73%. When video prompting was reinstated, her performance immediately improved, and baseline conditions were implemented again after four sessions. Shelly maintained her performance above 80% for the remainder of the study (M = 89%; range = 80%–100%). Video prompting sessions were an average of 3.2 min per session (range = 2–7.9 min).

Peter

Peter’s data for making lemonade are presented in the second tier of Figure 1. In baseline for making lemonade, Peter completed an average of 20% (range = 13%–27%) of the steps correctly. When video prompting with error correction was implemented, Peter demonstrated progressive improvement (M = 76%; range = 20%–100%), though he never mastered the skill. Peter completed video prompting sessions in an average of 4.2 min (range = 1.9–8.8 min).

Andrew

Andrew’s data for making lemonade are presented in the bottom tier of Figure 1. Andrew’s performance in baseline was highly variable (M = 47%; range = 20%–73%). When video prompting with error correction was implemented, he mastered making lemonade in 31 sessions (M = 91%; range = 67%–100%) and maintained this performance (M = 99%; range = 93%–100%) for 18 weeks. Video prompting sessions were an average of 1.6 min (range = 0.7–4.9 min).

Multicomponent Task (Folding Shirts)

Shelly

Shelly’s data for folding shirts are presented in the top tier of Figure 2. In baseline, Shelly performed an average of 23% (range = 13%–29%) of the shirt folding steps correctly with stable responding. When video prompting with error correction was introduced, her correct responding increased (M = 75%; range = 54%–92%), but she never performed more than 92% of the steps correctly. Video prompting sessions lasted an average of 8.6 min (range = 3.4–18.7 min). In an attempt to increase her performance to 100%, we implemented in vivo instruction. During these sessions, she did complete 100% of the steps correctly twice, but she was not consistently performing the task with 100% accuracy (M = 92%; range = 80%–100%). In vivo sessions lasted an average of 3.2 min per session (range = 2.5–3.9 min).

Peter

Peter’s data for folding shirts are presented in the middle tier of Figure 2. In baseline, Peter completed an average of 1.7% (range = 0%–4%) of the shirt folding steps correctly, and his responding was stable. When video prompting with error correction was implemented, he demonstrated a slight improvement (M = 32%; range = 15%–50%), but he never performed more than 50% of the steps correctly. Video prompting sessions were an average of 14.8 min (range = 10.4–17.4 min). After 27 sessions with video prompting with error correction, we introduced in vivo instruction, after which he demonstrated a stable increase in responding (M = 88%; range = 54%–100%), ending the study with two sessions with 100% correct responding. In vivo sessions lasted an average of 4.5 min (range = 2.7–7.3 min).

Andrew

Andrew’s data for folding shirts are presented in the bottom tier of Figure 2. In baseline, Andrew performed an average of 58% (range = 50%–71%) of the shirt folding steps correctly. When video prompting with error correction was introduced, his performance quickly improved (M = 93%; range = 79%–100%), and he mastered the skill in 20 sessions. Andrew maintained his performance for 11 weeks above 80% (M = 96%; range = 88%–100%). Video prompting sessions were an average of 4.2 min (range = 2.7–7.3 min).

Sequential Task (Loading Dishes Into a Dishwasher)

Peter

Peter’s data for loading the dishwasher are presented in the top tier of Figure 3. In baseline, Peter did not complete any steps for loading the dishwasher correctly. When video prompting with error correction was implemented, his performance did not initially improve, but in the eighth intervention session, his performance began to steadily improve (M = 45%; range = 0%–84%). When baseline was reinstated with the new dishwasher, Peter completed an average of 10% (range = 5%–15%) of the steps correctly. When video prompting with error correction was reinstated, he demonstrated a slightly variable increasing trend (M = 67%; range = 25%–90%). After Session 70, his performance indicated a decreasing trend. In Session 78, in vivo instruction was introduced, and his proficiency increased, with his last intervention session at 100% (M = 86%; range = 70%–100%). Video prompting sessions were an average of 9.5 min (range = 6.0–16.1 min). In vivo instructional sessions were an average of 1.6 min (range = 0.7–4.9 min).

Shelly

Shelly’s data for loading the dishwasher are presented in the second tier of Figure 3. In baseline, Shelly completed an average of 8% (range = 0%–16%) of the steps correctly. When video prompting with error correction was initiated, her performance increased (M = 75%, range = 58%–84%). When baseline was reimplemented, she completed an average of 20% (range = 10%–30%) of the steps correctly. When video prompting with error correction was reintroduced, she completed an average of 85% (range = 60%–95%) of the steps correctly. Video prompting sessions were an average of 6.1 min (range = 3.1–16.4 min).

Andrew

Andrew’s data for loading the dishwasher are presented in the bottom tier of Figure 3. In baseline, Andrew performed an average of 40% (range = 11%–68%) of the steps correctly. When video prompting with error correction was initiated, his performance increased (M = 80%, range = 63%–89%). When baseline was reintroduced, he completed an average of 64% (range = 55%–70%) of the steps correctly. When video prompting with error correction was initiated again, he mastered the skill (M = 93%; range = 80%–100%). He maintained his performance for 3 months (M = 95%; range = 85%–100%). Video prompting with error correction sessions was an average of 3.6 min (range = 1.9–5.8 min).

Comparison Across Tasks

Table 2 illustrates how skill acquisition, mastery, and session duration across tasks varied across participants. For the multicomponent task, making lemonade, two of the three participants mastered the skill using video prompting. Andrew mastered the task in 26 sessions, and Shelly did so in 64 sessions. The average session length aligned with the speed of acquisition (i.e., as the number of sessions to mastery increased, the average session time increased as well). Participants demonstrated increased performance with the multicomponent task, but only one participant, Andrew, mastered the skill. Similar to shirt folding, Andrew was the only participant to master the sequential task of loading the dishwasher. The session duration seemed to align with the number of steps required to complete the task. Sessions for the 15-step process required to make lemonade were the shortest, and sessions for the 24-step process required to fold three shirts were the longest.

Discussion

The purpose of this study was to determine whether students with severe to profound intellectual disabilities would differentially acquire different types of skills using video prompting with error correction. Although not all skills were mastered by all participants, they all made gains across skills using video prompting with error correction with all nonmastered skills. At the end of the study, all participants performed the skills with greater than 80% accuracy and several performed with greater than 90% accuracy. One participant mastered all three targeted skills, one participant mastered one of the three skills, and the third did not master any of the skills. Across the skills, a pattern emerged suggesting that the skills were not equivalent. Making lemonade (i.e., multistep) was the only skill that two of the three participants mastered, and we did not move this entire skill into in vivo instruction for any of the three. The multicomponent skill (i.e., folding shirts) was moved into in vivo instruction for two of the three participants when they stopped making progress within a relatively short period of time (27–35 sessions), though one participant mastered the skill in 20 sessions. Finally, loading the dishwasher (i.e., sequential) had the most varied performance, with one student mastering the skill, one making clear progress toward mastery, and one moving to in vivo instruction after 71 sessions. Although preliminary, these findings suggest that video prompting with error correction has an initial positive effect on skill acquisition but overall may be influenced by the type of task. Specifically, Andrew mastered each skill at different rates. The multicomponent task had the fewest sessions (20), followed by 26 sessions to master the multistep task, and 35 sessions to master the sequential task. For Andrew, at least, it seems clear that some aspect of the tasks impacted the efficacy of video prompting. That is, although the multicomponent task required 24 total steps, the repetitive nature of the task seemed to influence speed and rate of acquisition.

There are some possible reasons for these findings. For the multicomponent skill (i.e., folding shirts), the benefits of massed practice are well researched and suggest that students with severe disabilities can benefit from explicit and repeated practice of skills in acquisition (Browder & Spooner, 2011). The massed practice built into folding multiple shirts seemed to benefit Andrew and Shelly, but not Peter. Peter consistently missed the same steps for each component—namely folding in the sleeves—translating two missed steps into six missed steps across the three repeated components. These findings suggest that for students who respond quickly to video prompts, the repetitive nature of the skill may provide sufficient within-session practice that the whole skill can be acquired quickly, yet students who do not may end up practicing errors repeatedly, leading to less efficient acquisition patterns.

For the multistep task (i.e., making lemonade), there may have been a strong motivation factor in place. Each participant quickly consumed the lemonade once it was made, suggesting that the natural consequence of consuming the lemonade motivated all three participants to make it. Both Andrew and Shelly mastered and maintained the skill. Although Peter did not master this skill, he made significant progress, performing the last three sessions with at least 87% accuracy. Another trend we saw with these data was that the average duration of the sessions was the shortest of all tasks (range = 1.6–4.2 min), despite this task having the longest total duration for video prompts (129 s). It may be that faster access to a preferred drink led to improved performance—though the participants were allowed to drink the lemonade regardless of performance.

The sequential task (i.e., loading the dishwasher) took Andrew the longest to master and neither Shelly nor Peter mastered this skill—though both were performing significantly above baseline with video prompting with error correction, ending the study performing at least 80% of the steps correctly. It is possible that the lack of within-session repeated practice and seemingly less preferred nature of this skill may have led to these results.

Limitations and Future Research

Although video prompting with error correction led to significant improvement in performance, there are several limitations to consider. First, the casual design that we chose could not isolate what exactly about these tasks lead to differences in acquisition. The three task types as defined here may have differed in many relevant ways and may have been affected differently by participants’ unique histories and preferences.

Another limitation of this study is that we found it necessary to add intervention components to enhance skill acquisition. For loading the dishwasher, this included adding a stimulus prompt to the dishwasher to highlight which button to push. For Peter making lemonade, we introduced massed practice trials for putting the spoon down into the cup and stirring. Before this, he would get barely half of the spoon into the water. Finally, we changed from video prompting to in vivo instruction with Peter for loading the dishwasher and for Peter and Shelly for folding shirts. Although both made progress with video prompting with error correction, there were particular steps that they were not learning using video prompting even with error correction. Given these additional components, the addition did not affect students’ performance in ways that take away from the effectiveness of video prompting with error correction. In addition, we cannot say with certainty which aspect of the intervention led to skill acquisition.

A related limitation is that we did not have a preset criterion to move to in vivo instruction. Although visual analysis of the data supported this transition, having a preset criterion may have resulted in students being moved more quickly to in vivo instruction. In vivo procedures were not added in a way that allowed us to draw conclusions about the comparative efficacy of the two interventions. Given that, future research might compare the efficacy of each intervention (i.e., in vivo and video prompting) across different tasks to determine whether the type of task affects participants’ acquisition regardless of teaching method.

Future researchers might take steps to avoid these limitations. Particularly for multicomponent tasks that are repetitive in nature, future researchers might initially teach only one repetition of the task. Then, once the skill is mastered, add in additional repetitions. For example, rather than teaching someone initially to fold a basket of shirts, it might be more efficient to teach them to fold one shirt (i.e., multistep task), then increase the number of shirts to be folded (i.e., multicomponent) only after folding one shirt is mastered. This might avoid repeating the same errors, as we saw with Shelly and Peter. Future researchers may also want to consider using a multiple probe across task types, rather than across participants, as another method of examining the impact of task type. Finally, including a preset criterion for changing the intervention if progress is not being made could enhance the efficiency of instruction.

Another limitation was a lengthy break in sessions (i.e., summer break). None of the participants were available during the summer term, so they all experienced a 2½ month break from making lemonade and loading a dishwasher. All three participants had a drop in performance from the end of one school year to the start of the next. However, their performance with making lemonade returned to prebreak levels within a few sessions. This break also resulted in the introduction of a new dishwasher, which was installed over the summer.

A final limitation is that we changed the data collection procedures. Initially, steps were scored as correct only if they were completed accurately and in the order presented in the task analysis. We noticed that as the students mastered steps, they would initiate a step without first watching the video and would occasionally do the steps in a different order. For example, with making lemonade, the task analysis was written so that water was added to the lemonade powder. Shelly and Peter both started filling the glass with water before adding the lemonade powder. This behavior was functionally equivalent—in that the end product was a glass of lemonade. As such, we made the decision to count steps that were functionally correct as such, even if they occurred out of the order of the task analysis. In reviewing the data, this change in data collection had no impact on the interpretation of each participant’s behaviors. Although a limitation, we did not want to introduce error correction procedures for steps that were functionally and topographically correct. Instead, we wanted to focus our attention on those steps that were both functionally and topographically incorrect. Future researchers and practitioners may want to test their task analyses with other students with disabilities to determine whether there are different ways to correctly complete the task, then follow the student’s lead with respect to order of step completion.

In addition to the areas of future research that can be pursued to overcome some limitations in this study, there are some broader directions of future research that might be undertaken. First, it will be essential to replicate these findings with additional participants. More importantly, conducting this study with a new group of participants who all meet mastery will be important to make broader statements about the differential effects across types of tasks. Second, if these findings replicate, and similar challenges arise, it would be worthwhile to explore other means of introducing video prompts for skills so that participants can be as successful as possible within the shortest amount of time. Finally, as noted with the making lemonade task, it would be worthwhile to explore the added effects of using contingent versus noncontingent reinforcement on performance outcomes. In this study, all reinforcement was provided noncontingently (i.e., regardless of accuracy of performance). Making reinforcement contingent could improve the efficiency of the intervention.

Implications for Practice

In conducting this study, we found that the participants struggled with different components of the skills. In particular, Peter struggled with putting the spoon into the glass to stir the lemonade. He also struggled with straightening the shirt so that it lay flat when he folded it. If a teacher was teaching these skills to a student, we would encourage them to immediately focus instructional attention on those steps, possibly introducing contingent reinforcement for specific steps needing more focused attention. We also found that some participants completed steps in an order different from what we had developed. For example, they would fill the glass with water before putting in the lemonade powder. Although this was out of order from the task analysis, it was still functionally correct (i.e., water added to powder OR powder added to water both result in lemonade). Using task analyses in instruction is an effective tool, but instructors must remain flexible with the scoring of accuracy. If a step cannot occur before another (e.g., stirring the lemonade powder before adding water), these errors must be corrected, whereas functionally correct steps should be considered correct.

In terms of the validity of the use of video prompting procedures in classroom settings, teachers may find the benefits of video prompting assist in skill acquisition for students with intellectual and developmental disabilities. Video prompting procedures can be played on a wide range of devices, some of which may already be in the classroom and/or school. Instructional fidelity is also a key component in video prompting as the verbal and visual prompt remain consistent throughout intervention. In other words, teachers would know that students were being given the same prompt across staff members. Furthermore, teachers are able to create videos quickly using the same or similar materials to those used in intervention. If filmed in a controlled setting, teachers would not need to edit the videos but could use them immediately. Last, once created, the videos can be used across the years and students and shared on multiple devices.

Conclusion

Students with severe disabilities can benefit from instruction provided using video prompts, but there is little research available regarding how to best target and introduce skills. This study provides preliminary data that there are at least some differential effects across different types of tasks (e.g., multistep, multicomponent, sequential). Data suggest that students may perform better starting with multistep tasks, before moving on to multicomponent and sequential tasks.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Banda

D. R.

Dogoe

M. S.

Matuszny

R. M.

(2011). Review of video prompting studies with persons with developmental disabilities. Education and Training in Autism and Developmental Disabilities, 46, 514–527.

Browder

D. M.

Spooner

(2011). Teaching students with moderate and severe disabilities. New York, NY: Guilford Press.

Cannella-Malone

H. I.

Fleming

Chung

Y. C.

Wheeler

G. M.

Basbagill

A. R.

Singh

A. H.

(2011). Teaching daily living skills to seven individuals with severe intellectual disabilities: A comparison of video prompting to video modeling. Journal of Positive Behavior Interventions, 13, 144–153. doi:10.1177/1098300710366593

Cannella-Malone

H. I.

Miller

Schaefer

J. M.

Jimenez

E. D.

Page

E. J.

Sabielny

L. M.

(2015). Using video prompting to teach leisure skills to students with significant disabilities. Exceptional Children, 82, 463–478. doi:10.1177/0014402915598778

Cannella-Malone

H. I.

Sigafoos

O’Reilly

de la Cruz

Edrisinha

Lancioni

G. E.

(2006). Comparing video prompting to video modeling for teaching daily living skills to six adults with developmental disabilities. Education and Training in Developmental Disabilities, 41, 344–356.

Chan

J. M.

Lambdin

Van Laarhoven

Johnson

J. W.

(2013). Teaching leisure skills to an adult with developmental disabilities using a video prompting intervention package. Education and Training in Autism and Developmental Disabilities, 48, 412–420.

Council for Exceptional Children, & CEEDAR Center. (2017). High leverage practices for special education teachers. Arlington, VA: Council for Exceptional Children.

Gast

D. L.

Ledford

(2014). Multiple baseline and multiple probe designs. In Gast

D. L.

Ledford

(Eds.), Single case research methodology: Applications in special education and behavioral sciences (pp. 251–296). New York, NY: Routledge.

Mechling

L. C.

(2007). Assistive technology as a self-management tool for prompting students with intellectual disabilities to initiate and complete daily tasks: A literature review. Education and Training in Developmental Disabilities, 42, 252–269.

10.

Mechling

L. C.

Ayres

K. M.

Purrazzella

(2014). Continuous video modeling to prompt completion of multi-component tasks by adults with moderate intellectual disability. Education and Training in Autism and Developmental Disabilities, 49, 3–16.

11.

Newman

Wagner

Cameto

Knokey

A. M.

Shaver

(2010). Comparisons across time of the outcomes of youth with disabilities up to 4 years after high school: A report of findings from the National Longitudinal Transition Study-2 (NLTS2). Menlo Park, CA: Stanford Research Institute International. Retrieved from www.nlts2.org/reports/2010_09/nlts2_report_2010_09_complete.pdf

12.

Sigafoos

O’Reilly

Cannella

Edrisinha

de la Cruz

Upadhyaya

. . . Young

(2007). Evaluation of a video prompting and fading procedure for teaching dish washing skills to adults with developmental disabilities. Journal of Behavioral Education, 16, 93–109.

13.

Sparrow

S. S.

Balla

D. A.

Cicchetti

D. V.

(1984). Vineland Adaptive Behavior Scales: Interview edition survey form. Circle Pines, MN: American Guidance Service.

14.

Van Laarhoven

Johnson

J. W.

Van Laarhoven-Myers

Grider

K. L.

Grider

K. M

. (2009). The effectiveness of using a video iPod as a prompting device in employment settings. Journal of Behavioral Education, 18, 119–141.

15.

P. F.

Cannella-Malone

H. I.

Wheaton

J. E.

Tullis

C. A.

(2014). Using video prompting with different fading procedures to teach daily living skills: A preliminary examination. Focus on Autism and Other Developmental Disabilities, 31, 129–139. doi:10.1177/1088357614533594