Educating AP Students Across Locales

In 2016, Michigan Virtual, in conjunction with partner schools in the Virtual School Leadership Alliance, published a report on educating students across locales. This report investigated how students in locales across a variety of states were using K-12 online courses to supplement their education, and in particular focused on enrollment and success trends for rural students. This focus on rural students was based on the premise that rural students had less access to specialty (e.g. Japanese, or photography courses) or advanced courses (i.e. AP level courses).

For five of the seven virtual schools included in the initial report, rural students were over-represented in online student populations. Specifically, the percent of rural students enrolled in the state virtual school was larger than the percent of rural students statewide, see Table 1 for specific enrollment percentages. Based on some of our original assumptions, that were not specifically addressed in the report, we wondered if the same pattern would hold true for rural students in advanced level courses, or might be more or less pronounced. Would rural students continue to be over-represented in advanced level courses, and would we see even larger discrepancies?

For the 2014-19 school years, we found rural students overrepresented at about the same amount in advanced level courses that we did for all courses in the 2013-14 school year.
Given that the data from the 2013-14 school year is now 5 years old we wanted to assess whether the percentage of rural students remained the same or changed over this time. Using data provided by the State of Michigan through the Center for Educational Performance and Information (CEPI) Educational Entity Master and mischooldata.org website we calculated the percent of rural students statewide in the fall of the 2018-19 school year to be 21%, unchanged from the 2013-14 school year. Please also note that this follow-up investigation only included data from the Michigan Virtual School, and as such, no data from the 2014-19 school years are available for the other schools included in the original study.

This trend of over-representation in online courses held true for advanced courses, suggesting that the same proportion of rural students were enrolling in online courses as were enrolling in advanced level online courses. Given this, our team also wondered about some of our other original yet un-tested assumptions. In our original report, we found that for the Michigan Virtual Schools, rural students successfully completed their courses on par with students in other locales for the 2013-14 school year, with students from cities being the least likely to pass their online courses and students in towns the most likely to pass.

While other virtual schools also showed significant results during our analyses, no consistent pattern emerged across schools. Put another way, each virtual school presented a unique pattern of probability of success by locale. It was not the case that rural students performed a certain way across multiple schools. Rather it seemed that factors such as student gender and enrollment reason factored into student success more strongly than locale. The results of this initial report laid the groundwork for investigating the effect of locale on student enrollment and success, however it also left our team wondering if the initial analysis was too broad and perhaps we would find a different pattern if we investigated student success in only advanced courses.

As a reminder, the initial report used the metric of pass rate, a calculation of how many completed (pass or fail) course enrollments earned the minimum required amount of course points. Incomplete or withdrawn enrollments were removed from analysis both in the initial report to maintain consistency across schools (often with varying policies of recording and reporting on those incomplete enrollments) and, in this analysis to maintain consistency across reporting. Again too, most virtual schools do not assign grades or set the threshold for the passing of a course, as such the threshold used here is at least 60% of the total course points, a metric used by the Michigan Virtual School.

It was clear that while the percentage of students passing their advanced level courses was higher overall (due in large part to student self-selection into advanced courses) than for general enrollments in the 2013-14 school year, the same trends held true, see table 2 for pass rate percentages. A statistical analysis to test significance was not done however, the category of ‘city’ again reported the lowest overall pass rate of 87.4% while town reported the highest at 93.2%. The rural and suburban locales were again in the middle and quite close in terms of overall pass rate.

Given these results as well as those regarding enrollments above, it seems that rural students are enrolling in and passing their advanced level courses at approximately the same rate as general education courses. Our initial assumptions regarding rural students enrolling in more advanced level courses was not correct. Further, it seems that rural students fare just as well in the advanced level courses as they do in their general courses, when compared to other locales. For our team, these findings confirm those of the initial report that there are likely school-level (availability and commitment of an on-site mentor, stable internet access, etc.) and student-level (readiness for online learning, reason for enrollment in online course, etc.) factors that play a larger role in determining enrollment and success in online courses than locale.