Put yourselves in the shoes of a middle-to-low-income parent in a dysfunctional urban or rural school system. You want your kids in the best schools possible. You want your kids to have the opportunity to choose college if it’s right for them, but all of the available schools you are zoned for are awful and your child may not be able to test into the selective magnet schools that many districts offer. Urban school districts have a long and illustrious track record of mismanagement. The chances of reform while your child is in school are extraordinarily slim.
When faced with this people normally only have two options: either send their kids to a charter school, or try to move somewhere with a better school system. Both options come with their issues. Charter schools typically only perform marginally better than traditional public schools, and many families can’t afford to uproot their family to move to a better school district which likely has housing policies that restrict any low-income immigration to the city.
One possible future option which hasn’t quite come into fruition is online schooling. Non-profit charter networks could build a system of online modules paired with teacher assistants to allow students to get an accredited and personalized education at home. It could be modeled on existing structures like Coursera, Udacity, and Khan Academy, except it would have to have the internal infrastructure and flexibility to provide a full accredited education to more students across different states with different education standards. The general model is to provide students with a series of online modules that are responsive to their questions and confusions. Students could move through the material at their own pace and reach out to teacher assistants when questions arise.
On an operations side this could be very efficient since it cuts down on the significant operations and building costs associated with running a school. It could allow schools to pour more money into providing high pay to the curriculum developers and teaching assistants while also making sure that superstar teachers are the ones creating the curriculum for all students. In addition, it cuts out the significant costs associated with commuting and it may allow for a greater student-to-teacher ratio which cuts down on labor costs.
The most important part, however, is the efficiencies it brings to the learning side. Our education system is set up such that it is very difficult to teach students the material that they haven’t yet mastered. Students are instead pushed through the education system, being presented content that they are not ready for. They are almost never taught information based on what they already know. Students who never passed algebra 1 are thrown into geometry. Students with 5th grade reading levels are frequently asked to succeed in high school English classes. Teaching students at the right level (TaRL) has consistently been one of the most effective interventions in global education but there is almost no adherence to it in the United States. It would take a massive and costly restructuring of our current system to implement teaching at the right level. An online virtual school may be one of the easiest ways that TaRL could be tested and implemented in the United States.
On top of the ability to teach students at the right level, greater access to online learning as an alternative to traditional public schools also improves on the amount of time that students will spend actually learning. A huge amount of time every day is spent disciplining students, preventing fights, transitioning between classes, going to lunch, getting students on task, and commuting to and from school. Students spend a tremendous amount of time not learning. Unfortunately there aren’t any formal studies on this question. At least anecdotally I have homeschooled family members who are grades ahead of their age but only spend 4–5 hours a day working on lessons. The hope is that an efficient and widespread implementation of online school could allow students to complete the same amount of work in a fraction of the time.
By far the most important aspect is the possibility of massive, cheap scalability . This could be provided at low cost to cash strapped school districts in both rural and urban areas. It could be especially effective in areas with a lot of difficulty with teacher retention. Millions wouldn’t have to be spent on paying for land and buildings, and, once again, the best teaching procedures, curricula, and teachers could be democratized and shared across regions that would not otherwise have access.
Now I know that this may sound naive and optimistic. This sort of idea would get many eyerolls by most people in the education establishment. But that shouldn’t alone be a reason to not take it seriously as an option for some students. The technology needed to accomplish this task is starting to catch up. For instance, a recent randomized study on an all-online statistics course found that students performed as well as students who attended lectures in-class. That study followed typical online classes where students are just given lecture videos and course materials without much access to direct tutoring or software based interventions. This study is helpful for evaluating the effectiveness of MOOC-models. It looks like colleges can plausibly MOOCify their primary courses without measurable impact on student understanding. This is a good reason for why universities at least could provide many of their degree options all online. But we can do better than that.
We now have different attempts at making intelligent software that responds directly to the students. Think of it as an AI tutor. This meta-analysis of intelligent tutoring systems found that the median effect “was to raise test scores 0.66 standard deviations over conventional levels, or from the 50th to the 75th percentile.” This 2011 review found an effect size of .76 for intelligent tutoring systems as opposed to an effect size of only .79 for human tutoring. In other words, these intelligent tutoring systems are improving continually and a 9-year old paper that reviewed even older studies found that intelligent tutoring systems could meet the ability of human tutors. There of course needs to be more research on this and the usefulness of this technology for K12 education will depend entirely on the effectiveness of the logistical systems needed to get students to consistently work with and learn from the systems, however the current state of the research points to this being a very real and near possibility with the right implementation.
Finally, I should say that this is not meant in any way to be presented as a global solution to all of the US’s education problems. It clearly won’t work for many students and many ages. At the very least, there would have to be minimum parent/guardian presence and resources would have to be put into providing public wifi, laptops, and other tools that will make this viable for a broad range of students. It would also most likely work best for students who are already self-starters and are conscientious enough to do their daily lessons. In a way this could be seen as a form of gifted education for children with a lot of potential and curiosity who are held behind by otherwise dysfunctional school environments.
The main issue, though, is that public education isn’t just about learning. In some ways its main function is as professional babysitting. In a lot of unfortunate ways school is state funded daycare from 7 am till 4 pm from August to June. There is no way that online schooling alone could meet the needs of even a majority of students, but there are families who know that public education will mean trauma and developmental stagnation for their children. This could be a solution for them and it should get more support as a viable and widespread school choice option.