Abstract
During the COVID-19 pandemic, there were alarming reports of children missing out on online special educational activities due to a lack of access to those resources. We evaluated a simple online intervention using a concurrent multiple baseline design for three second-grade students with disabilities who unreliably accessed the remote curriculum. The dependent variable was the number of daily assignments completed. During baseline, the teacher provided students and parents with educational activities via Google Classroom, and the teacher contacted parents when a student missed educational activities. For intervention, the teacher implemented a task analysis that listed five daily assignments. Students earned preferred rewards contingent on completing all activities. Results showed that the intervention was effective in increasing engagement in online learning.
The COVID-19 global pandemic has profoundly altered the education system, especially for students with disabilities receiving special education services. The World Health Organization stated that individuals with disabilities may be disproportionately affected by the pandemic due to interruptions to the services they rely on and recommended children with disability “to continue playing, reading, learning, and connecting with friends” (World Health Organization, 2020, p. 4). To ensure that special needs children continue learning and receive public education and special services required by law (U.S. Department of Education, Individuals with Disabilities Education Act, 2020), school districts continued to provide the educational services remotely and online.
While online learning opens the possibility of delivering services to individuals in remote areas, access to online learning may vary depending on resources and proficiency with technology. There may be issues with structural inequities that affect the ability of educators and school administrators to deliver high-quality instruction. Limited access to online learning negatively affects how students learn and acquire new skills. Access to learning, especially for students with disabilities, has been a problem during the pandemic. Nationally, only half of U.S. school districts track students’ engagement in learning through attendance or one-on-one check-ins (Gross & Opalka, 2020). If students’ engagement and attendance are not checked, we cannot ensure that the students access remote learning opportunities.
Student attendance is related to student achievement. Students who attend school regularly showed higher achievement levels compared to students who do not have regular attendance, and the results have been shown in grade levels as early as kindergarten (National Center for Education Statistics, 2009). Also, poor attendance has long-term implications. Absenteeism throughout childhood, even as early as kindergarten and first grade, was more often found for high school student dropouts compared to high school graduates (National Center for Education Statistics, 2009). Although teacher effectiveness is the strongest determinant of student success (Adelman, 2006), student absence decreases the teacher’s ability to provide learning opportunities. Thus, limited access due to lower level of attendance and instruction would limit learning and educational outcomes. This highlights the need for interventions that promote access and engagement in online learning to obtain effective educational outcomes.
Access to education during a pandemic includes at least two levels. First, students may not have the available resources to get online and access educational resources. And indeed, for many students—those with and without disabilities—this level of access was problematic during the pandemic (Gross & Opalka, 2020). A second level of access is for students who have the resources to engage with online education but still do not access those resources for other environmental reasons, such as parents working and not having time to assist with online learning, lack of technology skills to appropriately use resources to engage with remote learning, or a lack of specific expectations for online learning, to name a few possible explanations. This study deals with the second level of access because the participants were fortunate to attend a school district that provided laptops and internet to students who did not already have these resources.
This study was designed to evaluate an intervention for children who had the appropriate technology resources but engaged with the online learning curriculum to a lesser degree than would be expected. All participants had a laptop and internet access that was purchased by the family or provided by the school district. The researchers investigated a simple method using task analysis and virtual reward to examine whether this intervention would increase participants accessing the full online curriculum. As a form of explicit instruction, a task analysis promotes successful learning by setting clear expectations and contingencies for consequences (Hughes et al., 2017). It also supports the well-being of children during the unprecedented time. There is a correlation between individuals’ well-being and contingencies for a positive future outcome (Luo et al., 2018) and long-term benefits to having predictable daily routines (Malatras et al., 2016). Thus, a task analysis with virtual reward contingencies can positively affect the well-being of students. The online curriculum was provided from March 2020 until June 2020 in response to the COVID-19 pandemic when the school shuttered its doors in accordance with a state mandate. To further the literature on responses to the global pandemic and to achieve social justice for individuals not reliably accessing education during uncertain times, the researchers aimed to develop a strategy that would increase student engagement for students with limited access to learning opportunities.
Method
Participants
Researchers selected three second-grade participants who showed low engagement and participation during online learning. This means that the students completed fewer than an average number of educational activities. These students were selected because verbal and email prompts to the student’s parents were ineffective at increasing engagement. Eddie was a 7.8-year-old Hispanic male who received an Individualized Education Program (IEP) with a classification of specific learning disability. He received free lunch and was classified as an English Language Learner (ELL). Jack was an 8.3-year-old Hispanic male who received an IEP with a classification of communication impaired. He received free lunch and was classified as an ELL. Riley was an 8.1-year-old African American female who received an IEP with a classification of other health impaired and medical diagnoses of Attention-Deficit/Hyperactivity Disorder (ADHD) and dyslexia. She received free lunch. All three participants read on at below grade-level. Based on the mid-year Developmental Reading Assessment Second Edition® (Pearson Education, 2006) assessed in January, Eddie read at an early-second-grade level (Level 18), Jack read at a mid–first-grade level (Level 10), and Riley read at an early first-grade level (Level 6). Based on the mid-year i-Ready® K-12 Adaptive Reading Diagnostic (Curriculum Associates, LLC, 2017) assessed in January, Eddie read at a first-grade level (score = 447), Jack read at a first-grade level (score = 438), and Riley at a kindergarten level (score = 405).
The classroom was housed in a Title I public elementary school located in a suburb outside of a large metropolitan area with pre-school to second-grade classrooms. The specific class was an inclusive second-grade classroom with 18 students, of whom five students were considered to be special education students with IEPs, seven students received free/reduced-priced lunch, and five students were classified as ELLs. The classroom implemented evidence-based behavioral and academic interventions and implemented the Comprehensive Application for Behavior Analysis to Schooling (CABAS, Accelerated Independent Learner program; Greer, 1994) model. This model included one head teacher and two teaching assistants who implemented a scientific approach to pedagogy, learning, curriculum, and classroom management, and all instruction was individualized to each student’s repertoire.
All 18 students had the prerequisite technology skills required to complete online daily assignments. Each student had a laptop and internet access purchased by the family or provided by the school district. The students could type and/or use a QR code badge containing the login information to their school accounts, submit assignments on Google Classroom, record videos on Flipgrid, and attend a Google Meet link at a designated time. Before the school closing, the classroom teacher taught all students how to use Google Classroom and Google Meet, and all students regularly used the QR code badge and/or typed login information. During online learning, the teacher created tutorial videos and written directions on how to navigate each platform and provided them to the students and parents via email and Google Classroom posts.
In the same inclusion class, there were two more second-grade students who had disabilities and did not require the intervention due to high engagement and participation rates. They completed an average of 4.5 out of 5 online educational activities (range, 4 to 5) each day across 22 school days. There were 13 general education (i.e., no IEP) second-grade students who did not require the intervention due to high engagement and participation rates. Five students identified as Caucasian, five students identified as multi-racial, two students identified as Hispanic, and one student identified as African American. These 13 students completed an average of 3 of 4 online educational activities (range: 1.7–4.0) each day across 22 school days. Students without IEPs did not have individual instruction as per the school district’s policy. Thus, these students were required to complete a total of four activities each day (Activities 1 through 4 listed in the Dependent Variable section). There were five students who engaged in fewer than the average number of educational activities. These students and parents systematically received verbal prompts and email prompts (described in Baseline), which increased the students’ engagement levels to the average without an intervention. The researchers did not count the days students were absent to calculate the average. Absent days were defined as days the teacher received a formal absent notice from the parent. There were two cases of parents communicating with the teacher when kids would be out (e.g., visiting grandparents, death in family), and the remainder responded to teacher e-mails (see Baseline section under Procedure). The first author was the head teacher of the classroom and had access to students’ daily login data.
Setting and Materials
The study took place across online platforms Google Classroom, Google Meet, and Flipgrid provided by the school district. All participants used their district login information to access the platforms. All participants had engaged in online learning along with their classmates since the school closure due to the COVID-19 pandemic. Online learning lasted approximately 3 months from March 2020 to end of the school year in June 2020.
Materials included a computer, camera, task analysis, Microsoft PowerPoint, and access to the online platforms. The district provided a Chromebook laptop to families that requested one. All three participants used Chromebooks distributed by the school district, which had a camera and access to internet. All students and teachers were provided access to the platforms using the district account. The teachers used Microsoft PowerPoint to create the task analysis with five components listed on a picture of a clipboard and the virtual reward written on top (see Supplemental Materials, Appendix A, for an example).
Dependent Variable
The dependent variable was the number of completed daily instructional activities. Every day, the participants could access five instructional activities: (a) attending live morning announcements on Google Meet, (b) recording a video on Flipgrid to answer morning meeting questions, (c) completing assignment for Subject 1 on Google Classroom, (d) completing assignment for Subject 2 on Google Classroom, and (e) attending individual instruction on Google Meet.
The researchers counted the components completed from 12:01 a.m. to 11:59 p.m. on the respective day. The researchers marked the live morning announcement component complete if the participant was present at the meeting using the Google Meet link and incomplete if the participant did not attend. The researchers marked the morning meeting video component complete if the participant recorded a video on the respective day’s Flipgrid link and incomplete if a recording was absent. The researchers marked the assignment components complete if the participant attempted answering at least one question on the worksheet and incomplete if there was no attempt. The researchers marked the individual instruction component complete if the participant was present at the meeting using the Google Meet link and incomplete if the participant did not attend. The researchers did not collect data on Fridays because there was no assignment as per the district’s policy of providing non-academic, project-based activities.
Procedure
Baseline
Baseline began after 1 month of remote learning when all students who needed computers and internet access received them from the district. During baseline, each student had five activities to complete each day. Every day at 8 a.m., the head teacher posted the Google Meet links, assignments, and Flipgrid link on Google Classroom under the assignment tab. The morning announcement Google Meet link was posted on the homepage of Google Classroom. Morning announcements were at 9 a.m. every day using the same link.
To increase engagement and participation rate, the head teacher posted a class-wide daily schedule on the Google Classroom homepage along with the Google Meet link for morning announcements. Morning meeting questions changed every day and consisted of questions asking students’ opinion (e.g., What is your favorite animal?). Students were asked to record a 90 s video with a response to the question. Assignments for each subject consisted of a teacher-created or sourced instructional video and corresponding worksheets. Each assignment took approximately 30 to 40 min to complete. The worksheets contained approximately 10 opportunities to respond, and the teacher provided written feedback for correct and incorrect responses and returned the work throughout the day. The daily individual instruction session for Eddie was at 9:30 a.m. and consisted of Direct Instruction (DI; Engelmann & Carnine, 1982) on reading, Jack’s session was at 10:00 a.m. and consisted of DI (Engelmann & Carnine, 1982) on reading, and Riley’s session was at 1:30 p.m. and consisted of individualized programs that targeted IEP goals for reading. Each individual instruction session lasted 30 min.
If a participant did not complete all the components from one day, on the next day the teacher provided a verbal prompt to the participant during the individual instruction sessions. Verbal prompts included the teacher saying, “We need to finish (components the participant did not complete).” If a participant missed 3 or more components for four consecutive days, the head teacher provided email prompts to parents. Email prompts included a gentle reminder that the student is strongly encouraged to complete the daily assignments listed on the daily schedules. The three participants continued to show low engagement despite these measures. Note that the parent reminder system was different for students without IEPs because there were no individual sessions and, thus, repeated non-engagement resulted in email prompts to parents only.
Intervention
Preference assessment
Prior to the intervention for the first participant, the researchers assessed preferred virtual rewards for each participant. During the individual instruction sessions, the teacher presented the participant with four options and asked, “Which one would you like the most?” The options included (a) playing a game with the teacher (e.g., tic-tac-toe, hangman, 20 questions), (b) acting as the teacher for morning announcement, (c) having lunch with the teacher on Google Meet, and (d) receiving a certificate with the participant’s name on it. After a participant chose one option, the teacher continued to ask for the next most preferred reward using the same procedure (multiple stimuli without replacement preference assessment; DeLeon & Iwata, 1996). As their first choices, Eddie and Riley chose playing a game with the teacher and Jack chose acting as the teacher for morning announcements.
Task analysis and virtual reward
As during baseline, during intervention the head teacher continued to post the links and assignments in the same manner every day at 8 a.m. The day before introducing the intervention, the teacher told each student during the individual instruction sessions, “Starting tomorrow, you will have a checklist with five things to do. If you finish everything, we get to earn (virtual reward ranked first).” The teacher also reminded the participant that he or she needed to come to morning announcements the following day to see the checklist for that day.
Every day at 8:30 a.m., the teacher posted the task analysis on Google Classroom only available to each participant and emailed it to each parent. Each day at the end of morning announcements, the teacher asked the participants to stay after the other students logged off. The teacher first reviewed the checklist from the day before in the order of Eddie, Jack, and Riley. This procedure was implemented in a staggered manner. On the first day of intervention, the teacher only asked Eddie to stay on after morning announcements, and on the second day, the teacher asked Eddie and Jack to stay on. On the third and following intervention days, all three participants were asked to stay on. If the participant completed all components from the day before, the teacher provided verbal praise, “Great job completing your checklist!” and told the participant that he or she earned the reward for later that day. Eddie and Riley received their reward in the beginning of the individualized instruction sessions. Playing a game with the teacher lasted about 10 min and the teacher provided 30 min of instruction after the game. No instruction time was removed. Jack, however, received his reward during morning announcements before going over the checklist together. The teacher said, “You finished everything on the checklist and earned your reward Jack!” and he read the morning announcement list with the teacher in front of all students in class. If all components were not complete, the teacher said, “Let’s try it again” and modeled how to complete the incomplete components using the sharescreen function on Google Meet. For example, if a participant did not record a video on Flipgrid, the teacher showed the steps to record a video on Flipgrid. All participants had already mastered how to accurately complete these components, which means that this procedure was implemented to prompt rather than teach the behavior.
After reviewing the checklist from the day before, the teacher went through the task analysis for that day using the sharescreen function on Google Meet in the order of Eddie, Jack, and Riley. The teacher read each component and said that he or she can earn the reward upon completing all components. If the student did not attend morning announcements, the teacher reviewed the task analysis during the individual instruction session. If the student did not attend individual instruction, the teacher called the parent for a gentle reminder of daily assignments. Delivery of the task analysis and parent communication was done by the head teacher.
Experimental Design
The researchers used a concurrent multiple baseline design across participants (Carr, 2005) to assess the effectiveness of the intervention package including task analysis and virtual reward on engagement in online learning. The researchers collected baseline data until each participant showed steady state responding with little variation. During baseline, the participants and parents systematically received verbal and email prompts (described in Baseline) just as all the classmates did. Intervention was introduced when these initial measures were ineffective at increasing the level of engagement. Then, the researchers implemented the intervention package with one participant while others remained in baseline. After one day, the intervention package was introduced to the second participant, and after another day, to the third participant. The researchers staggered the introduction of the intervention in a shorter period of time with less variation in baseline lengths due to the ethical concerns of withholding a potentially-effective intervention.
Interobserver Agreement and Treatment Fidelity
A second observer independently collected data for the purpose of assessing interobserver agreement (IOA) of the engagement rate for each student. The researchers calculated component-by-component IOA by adding the assignment components in agreements, dividing by the total number of assignment components in agreements and disagreements, and multiplying by 100%. The researchers obtained IOA data for 85% of baseline sessions and for 89% of intervention sessions with 100% agreement.
A second independent observer recorded whether the researcher (a) reviewed the task analysis for the respective day and the day before at the end of morning announcements, (b) posted the task analysis on Google Classroom at 8:30 a.m., (c) sent the task analysis to the parent email at 8:30 a.m., and (d) delivered the reward for Jack during morning meeting when he earned it. We were unable to collect fidelity on the delivery of rewards for Eddie and Riley because this was done during individual sessions with the teacher. The independent observer obtained treatment fidelity data for 96% of intervention sessions and fidelity was 100%.
Results
Figure 1 displays the number of completed educational activities online learning for Eddie (top panels), Jack (middle panels), and Riley (bottom panels). Figure 2 displays a component analysis of specific educational activities. During baseline, all three participants completed an average of 1.1 of 5 components (range: 0.5–1.5) each school day. After the implementation of the intervention, all three participants completed an average of 4.7 of 5 components (range: 4.6–4.8) each school day.
Number of completed educational activities for Eddie, Jack, and Riley.
Specific educational activities completed for Eddie, Jack, and Riley.
During baseline, Eddie completed an average of 1.5 components (range; 0 to 3) each day across 15 school days (Figure 1, top panel). Specifically, he attended individual instruction sessions and started to attend morning announcements in the middle of baseline but did not record videos for morning meeting and did not complete both assignments on most days (Figure 2, top panel). After the implementation of the intervention, Eddie completed an average of 4.6 components (range: 4–5) each day across seven school days.
Jack completed an average of 1.4 components (range: 1–2) each day across 16 school days (Figure 1, middle panel) during baseline. Specifically, he attended individual instruction sessions but did not attend morning announcements, did not record videos for morning meeting, and did not complete both assignments on most days (Figure 2, middle panel). After the implementation of the intervention, Jack completed an average of 4.8 components (range: 4–5) each day across six school days.
Riley completed an average of 0.5 components (range: 0–1) each day across 17 school days (Figure 1, bottom panel) during baseline. Specifically, she inconsistently attended individual instruction sessions, did not attend morning announcements, did not record videos for morning meeting, and did not complete both assignments for all school days (Figure 2, bottom panel). After the implementation of the intervention, Riley completed an average of 4.8 components (range: 4–5) each day across five school days.
The two other students with disabilities did not require the intervention because they showed consistently high rate of engagement with the remote learning curriculum. Rob completed an average of 4.2 components (range: 3–5) each day across 21 school days (see Supplemental Materials, Appendix B, top panel). He was absent on Day 20, and the researchers did not count this day to calculate the average components completed. James completed an average of 4.8 components (range: 3–5) each day across 22 school days (top panel, Appendix B).
Effect size was calculated for the three participants based on the number of components completed data displayed in Figure 1. The researchers used improvement rate differences (IRD) to find the magnitude of effect size between baseline and intervention phases (R. I. Parker et al., 2009). The first author hand calculated IRD and checked those numbers against an online calculator for IRD (http://www.singlecaseresearch.org/calculators/ird). IRD scores can be interpreted as following: 0.5 or below = small intervention effect, between 0.5 and 0.7 = moderate sizes of effect, and 0.7 or higher = large effect (R. I. Parker et al., 2009; Rakap, 2015). The maximum score is 1.00, which is obtained when all intervention data points are higher than all baseline data points (Rakap, 2015). Individual IRD results suggest a large effect for all three participants (Eddie, ES = 1.00; Jack, ES = 1.00, and Riley, ES = 1.00).
Discussion
The purpose of this study was to evaluate the effects of a simple intervention to increase access and engagement for students with disabilities who showed low engagement in online educational activities during the COVID-19 pandemic. Specifically, these students had the resources to access online learning but continued to show low rates of engagement due to environmental reasons. We found that three of 18 students (17% of the class) needed intervention, while the remaining students regularly accessed the educational curriculum and notified the teacher in advance when this was not possible. For the participants requiring intervention, the task analysis and virtual reward were effective in increasing engagement during online learning with large effect sizes. This study joins a literature showing the effectiveness of task analysis interventions combined with rewards (e.g., National Autism Center, 2015; Wong et al., 2014). In previous research, task analysis plus rewards was the primary intervention used to achieve outcomes, such as task completion and social behavior (e.g., D. Parker & Kamps, 2011). However, in this study, task analysis and virtual reward was the intervention to help students access other interventions—educational curricula and learning opportunities.
The findings have implications for practice and literature on access to remote learning and responses to a global pandemic. First, the results inform clinicians and educators on how to increase attendance and participation, and to have students access online learning opportunities for individuals who rely on them. Studies have shown the importance of student attendance on student achievement in grade levels as early as kindergarten (National Center for Education Statistics, 2009). While teacher effectiveness is the strongest determinant of student success (Adelman, 2006), student absence reduces the likelihood of engagement with effective instruction. A step before introducing an online learning curriculum is to ensure that individuals receiving these services reliably attend online learning opportunities and access the remote learning platform. Without appropriate access to online resources, effective instructions cannot be delivered, and student learning cannot occur. With a simple strategy such as task analysis and virtual reward, students will have a greater chance of accessing the online services and obtaining success, establishing equal opportunities to access education during uncertain times.
Furthermore, using a task analysis and virtual reward provides a routine for students, which supports the well-being of children during this unprecedented time. Researchers demonstrated that anticipation of desired future events activated the bilateral medial prefrontal cortex (MPFC) of the brain, which positively correlated to human well-being (Luo et al., 2018). This suggests the correlation between individuals’ well-being and contingencies for a positive future outcome. Thus, a task analysis with virtual reward contingencies can positively affect the well-being of students. Also, growing up with predictable, daily routines allowed individuals to have less problems with time management and attention as adults (Malatras et al., 2016). Although both studies assessed adults and adolescents, they highlighted that daily routines, which can be provided with a task analysis in conjunction with a reinforcer, can help children feel safe and support a positive experience during the stressful time.
The study is not without limitations. One limitation is that the increase in engagement could have been due to increased parent involvement. The researcher posted the task analysis for the student but also sent it to parent emails to inform the parents of the expectations for each school day. The intervention may have been effective in increasing parent involvement, which in turn increased student engagement in online learning. For future studies, researchers should measure parent involvement separate from student engagement to determine the effectiveness of the intervention directly on student-initiated learning. Researchers should also investigate the effectiveness of the intervention on increasing academic skills such as reading and math.
Another limitation is the shorter verification period in the concurrent multiple baseline design. The intervention was introduced to each participant in a shorter period of time with less variation in baseline lengths. The researchers were concerned about withholding a potentially effective intervention and thus introduced the intervention sooner. Furthermore, the baseline phase was 15–17 days and the intervention could have been introduced sooner. Jack and Riley showed stable baseline early on and Eddie showed slight improvement prior to the intervention. However, during baseline, the participants and parents systematically received verbal and email prompts similar to other classmates. The participants received the intervention when these initial measures were ineffective at increasing engagement. Also, the intervention phase was shorter compared to the baseline phase. A longer intervention phase would have shown the long-term effects of the intervention. However, the intervention phase was shorter because the intervention was introduced toward the end of the school year. Nonetheless, our study demonstrated high quality features such as having three demonstrations of experimental effect at three different points in time (Horner et al., 2005; Kratochwill et al., 2013). Future studies should employ other experimental designs such as multiple baseline design across different time periods or alternating treatment design with different types of virtual rewards to avoid such problems. In addition, replications across participants with and without disabilities and studies that systematically fade the intervention should be encouraged for greater generality.
Despite the limitations, the present study demonstrated that a task analysis along with a virtual reward is effective in increasing student access and engagement in online educational activities for students with disabilities. The findings suggest a strategy clinicians and educators can use to increase access to learning and identify individuals who may be at risk for falling behind due to limited access to learning. The study adds to the literature on remote learning and responses to the COVID-19 pandemic and informs us on how social justice outcomes can be achieved for individuals not reliably accessing education during uncertain times.
Supplemental Material
sj-pdf-1-sed-10.1177_0022466921998067 – Supplemental material for Increasing Access to Online Learning for Students With Disabilities During the COVID-19 Pandemic
Supplemental material, sj-pdf-1-sed-10.1177_0022466921998067 for Increasing Access to Online Learning for Students With Disabilities During the COVID-19 Pandemic by Ji Young Kim and Daniel M. Fienup in The Journal of Special Education
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.
Supplemental Material
Supplemental material for this article is available on The Journal of Special Education website with the online version of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
