MVLRI Research in Review: K-12 Online Special Learner Populations

Published on September 29, 2020
Written By: 

Kristen DeBrulerMichigan Virtual Learning Research Institute


Christa GreenMichigan Virtual Learning Research Institute

Research suggests that online learners with disabilities, those at risk of dropping out, and those taking courses for credit recovery benefit from additional assistance and instructional support. These learners can benefit from online courses however those courses and the accompanying instruction need to be responsive to the unique needs of these learners.

Suggested Citation

DeBruler, K. & Green, C. (2020). MVLRI research in review: K-12 online special learner populations. Michigan Virtual.


Since its creation in 2013 through 2020, the Michigan Virtual Learning Research Institute (MVLRI) at Michigan Virtual published approximately 20 research blogs and 75 research reports. This total does not represent everything published by MVLRI but rather only those publications including original research on K-12 blended and online learning. The nearly 100 resources represent research conducted internally by MVLRI staff, research conducted by partners at universities, colleges, and educational organizations, and covers a vast range of topics including, but not limited to, K-12 online best practices, online student motivation, K-12 blended teaching and professional development, and K-12 special populations. 

This body of work is extensive, and while there is tremendous value in each individual publication, there is also value in how that work fits with other similar research and the narrative that emerges from the collective understanding. Toward this end, MVLRI sought to identify, review, and synthesize the original research published in the past 6 years. Again, not every blog or report published via the website was included, only those containing original research. 

Out of the synthesis of resources, 10 main themes emerged. Each theme is presented individually in the interest of brevity. A full reference list is provided at the end of this document noting the resources that contributed to this report. 


Resources for inclusion in the synthesis were identified through the website in the “Publications” and “Blogs” sections. All published blogs and reports were assessed to determine if they included original research. Those that did were included for synthesis. Once the approximately 100 resources containing original research were identified, each blog or report was reviewed and given up to three keyword tags. The following fields were also completed for each of the 100 resources: what we already know about the topic of research, what the resource adds, and implications for policy and practice. Resources were then thematically grouped and keywords were refined and combined. For example, K-12 online program evaluation and quality was combined with K-12 online program policy because although distinct, the themes were related and spoke to many of the same concepts and conclusions. 

Once the 10 thematic categories were identified, the resources within that category were reviewed again, both for accuracy in interpretation and to determine its relationship to other resources in the same category. Out of this process, the core findings and practical implications were identified. What is presented below is the synthesized understanding from the original research included. Because of the process, not every finding of every resource could be included, rather resources were reviewed to form a broad understanding of each theme and to identify what MVLRI has contributed and learned in the 6 years since it was formed.

K-12 Online Special Learner Populations Core Findings 

  • Overall, credit recovery learners did more poorly in their online courses than learners who take the course for other reasons. 
  • The number of learners taking the same mathematics course for credit recovery increased as the school year progressed.
  • A majority of credit recovery learners exhibited course behaviors, such as persistent work, similar to the majority of learners. A small number, however, showed less active engagement followed by final spikes in course work or procrastination. 
  • Failing credit recovery learners tended to exhibit extremely poor course pacing.
  • Teacher type did not significantly impact credit recovery learner outcomes.   
  • K-12 online program standards need to better support learners with disabilities and include specific considerations.
  • K-12 online teachers may require additional training and support to better serve learners with disabilities who often need help learning the content and understanding how to learn online. 
  • K-12 online courses should be designed to adhere to principles of Universal Design for Learning and be accessible for all learners. 
  • K-12 online learners with disabilities should be encouraged to enroll in online courses appropriate with their educational goals and supported in such courses. 
  • Learners drop out of school for a variety of reasons; K-12 online programs and courses may offer solutions to some of these reasons. 
  • Learners at-risk of dropping out need additional instructional and personal support, including re-engagement efforts and individual education plans. 
  • Relevant data are critical in identifying and supporting learners at risk of dropping out. Some models can accurately predict failure as early as 8 weeks in the semester.

K-12 Online Credit Recovery

Following general enrollment trends, most credit recovery course enrollments were at the high school level. The most frequent course subject for credit recovery was mathematics, and most credit recovery enrollments came from suburban or rural areas. Overall, credit recovery learners did more poorly in their online courses than learners taking the course for other reasons. The pass rate for credit recovery enrollments was 62% compared to the general pass rate of 85%. 

The number of learners taking the same mathematics course for credit recovery increased as the school year progressed. That is, more learners took the course for credit recovery in the spring than fall, and more in the summer than in the spring (Kwon, 2017a). Learners seemed to be enrolling in courses to recover credit lost in previous semesters. 

A majority of credit recovery learners exhibited similar learning behaviors, such as persistent work throughout the semester. There were some credit recovery learners who did display procrastination or an overall lack of course work throughout with a final spike of activity near the end of the semester (Kwon, 2017a). The number of learners in this group increased as the number of credit recovery learners overall increased each semester. 

Credit recovery learners who ultimately failed the course unsurprisingly exhibited extremely poor pacing throughout (Kwon, 2017b). While the converse is not true, that all learners with poor pacing will fail the course, it is an indicator to the online teacher that the learner may be at risk of failing and may need additional interventions or support. 

Credit recovery enrollments taught by full-time Michigan Virtual teachers had a higher overall group mean score than other credit recovery enrollments. However, there was no statistically significant difference in learner outcomes based on teacher type (Kwon, 2017b). This suggests that while the group of credit recovery learners in courses led by full-time teachers had higher scores, their outcomes (passing or failing the course) were the same as credit recovery learners with other teacher types. 

K-12 Online Learners with Disabilities

Research on learners with disabilities conducted or sponsored by MVLRI tended to focus on four themes through the lens of learners with disabilities: online programs, online teaching, online courses, and online learners with disabilities. 

In a review of the iNACOL (now the Aurora Institute) National Standards for Quality Programs, researchers concluded that there was room for considerably more language dedicated to programmatic needs of learners with disabilities (Pace, Rice, Mellard, & Carter, 2016). There was also the assertion that many learners with disabilities may be counseled out or dissuaded from enrolling in online courses. Should those learners with disabilities be admitted into online courses, they may hide their disability status either intentionally or unintentionally, often with the understanding that support may be limited to the typical supports available to all learners (Pace et al., 2016). 

One of the typical supports available to all learners is an on-site mentor, a requirement in Michigan. The quality of support that on-site mentors provide varies across schools and more research is needed that examines the practices of successful on-site mentors (Rice, 2018).

With respect to teaching, there was an assertion that online teachers not only need to understand the overall complexities of teaching learners with disabilities, but also with the online setting, including the needs to different sub-groups of learners with disabilities (Rice, Pace, & Mellard, 2016). In addition, teachers and course designers may need to consult with professionals to support learners who are deaf, hard of hearing, or visually impaired (Deschaine, 2018). 

Overall, there was general consensus that online teachers require training and support to meet the needs of learners with disabilities. Technology may allow teachers to personalize their online classes and meet the academic and non-academic needs of learners with disabilities (Rice, Pace, & Mellard, 2016). Online teachers need to be aware that high learner load and physical separation, however, may make it more difficult to provide each learner with the types and levels of support they need (Rice, 2018). One area of critical support for all learners, but particularly learners with disabilities, is developing and supporting social-emotional learning and self-regulation (Rice et al. 2016). 

With respect to online courses, first and foremost, it is important to note that online courses are held to the same Federal laws protecting the rights of learners with disabilities (Rice, Mellard, Pace, & Carter, 2016). With this in mind, courses should be designed to adhere to Universal Design Principles (Rice et al., 2016). Such well-designed courses benefit not only learners with disabilities but all learners by allowing them to easily access and thoughtfully engage with the course content. 

Course designers must also understand how individual educational needs affect the educational expectation and progression through the course (Deschaine, 2018) and that learners with disabilities may benefit uniquely from engaging with the content through multiple means and working with peers in their online courses (Rice et al., 2016). 

Finally, it is important to note that online learners with disabilities face unique challenges in addition to those faced by all learners, such as not only learning the content but learning to learn online. Because of this, learner attrition rates are higher in online courses than in face-to-face courses (Rice, 2018). Online programs, teachers, and course designers need to understand that learners with disabilities may need additional peer or instructional support (Deschaine, 2018). Most importantly, learners and parents need to be advocates for the needs of the learner in the online or blended class (Deschaine, 2018). 

K-12 Online At-Risk Learners 

There are a number of individual and institutional factors contributing to learner dropout. Many learners drop out of school for academic reasons such as they were failing or were unprepared by elementary and middle school, or for personal reasons such as they became a parent, to get a job, or to care for a family member. Dropout solutions that directly address the reasons learners drop out are most effective (Ferdig, 2010). For many of these reasons, online programs and the features of online courses (anytime, anywhere, any pace) may offer a solution to some challenges of high school completion (Ferdig, 2010). 

Online learners at risk of dropping out need instructional support and personal support from their online teacher (and on-site mentor, if applicable). If learners were at risk of dropping out for motivational reasons, any retention effort must attempt to re-engage the learner (Ferdig, 2010). Individual instruction plans, coupled with innovative instruction, can be used to re-engage previously disengaged learners (Ferdig, 2010). Retention efforts must also focus on communication with learners, keeping them on pace with their coursework and supported throughout the online course. 

There is a critical need for the collection and use of data towards drop-out prevention (Ferdig, 2010). Early warning systems can help online teachers identify at-risk learners and prevent learners from dropping out (Ferdig, 2010). One such early warning system was studied for use with Michigan Virtual courses. The 2018 study conducted by Hung and Rice (2018) adopted a Deep Learning algorithm (previous models had used a Machine Learning algorithm to some success) to identify at-risk learners in the eighth week of the semester. 

The predictive model was constructed using learner demographic variables, online behaviors, and online discussions (Hung & Rice, 2018). The predictive model used the constructed learner profile (common indicators for early warning) as well as the learner’s participation level (online behaviors and discussion). The results from the study indicated that the Deep Learning model was slightly better than the previous Machine Learning model and that the inclusion of the online discussions improved the overall predictive accuracy of identifying learners at risk of failing (Hung & Rice, 2018).

K-12 Online Special Learner Populations Practical Implications and Actionable Resources

  • While many credit recovery learners exhibit consistent pacing through the course, there are some that exhibit engagement patterns that are not associated with success in online courses such as low engagement with a final spike in activity near the end of the semester. Given that this population has already tried and failed to earn credit, it is important that credit recovery learners who do not exhibit a consistent pattern of course engagement have early additional teacher intervention and support. 
  • Online programs and providers, particularly those that serve credit-recovery and other at-risk populations, should have re-engagement policies and procedures for learners who are not making significant course progress by a predetermined date.
  • Given the complexities and challenges of teaching online learners with disabilities, K-12 online teachers should be provided training and regular professional development on supporting online learners with disabilities. Michigan Virtual has this as part of their new teacher onboarding process, and it is revisited regularly throughout the year. 
  • Policies and procedures should be in place to support students with disabilities in their online courses. Michigan Virtual works with the learners local schools to develop a plan to provide the necessary accommodations both online and face-to-face students. 
  • To ensure that all learners can access all course material and be successful in their online courses, online courses should be designed in accordance with industry standards regarding accessibility such as, but not limited to, the Universal Design for Learning (UDL), Universal Design Principles, and Web Content Accessibility Guidelines. While Michigan Virtual has not formally adopted the UDL framework, course design adheres to standards (Quality Matters) that are consistent with the UDL framework.


Deschaine, M. (2018). Supporting students with disabilities in k-12 online and blended learning. Michigan Virtual University.    

Ferdig, R. E. (2010). Understanding the role and applicability of K-12 online learning to support student dropout recovery efforts. Michigan Virtual University.   

Hung, A & Rice, K. (2018). Combining data and text mining to develop an early warning system using a deep learning approach. Michigan Virtual University. 

Kwon, J. B. (2017a). Examining credit recovery learning profile from time-series clustering analysis. Michigan Virtual University.  

Kwon, J. B.(2017b). Examining credit recovery experience at a state virtual school. Michigan Virtual University. 

Pace, J., Rice, M., Mellard, D., & Carter, Jr., R. A. (2016). Meeting the needs of students with disabilities in K-12 online learning: An analysis of the iNACOL standards for quality online programs. Michigan Virtual University.  

Rice, M. (2018). Virtual school course design: Accommodating students with disabilities. Michigan Virtual University. 

Rice, M., Pace, J., Mellard, D. (2016). Meeting the needs of students with disabilities in K-12 online learning: An analysis of the iNACOL standards for quality online teaching. Michigan Virtual University. 

Rice, M., Mellard, D., Pace, J., & Carter, Jr., R. A. (2016). Meeting the needs of students with disabilities in K-12 online learning: An analysis of the iNACOL standards for quality online courses. Michigan Virtual University. 

Table of Contents