Stanford’s Center for Research on Education Outcomes (CREDO) report on Pennsylvania charter schools continued many themes from its national report: average learning gains for charters are smaller, charters’ performance improves as students are enrolled longer, and importantly, high variability among charters, with significant numbers of both high-performing and low-performing charter schools.
But, there’s one very interesting finding. In the CREDO sample, taken from schools that enrolled students from 2007 – 2010, there are eight full-time, online “cyber” charter schools. These eight schools enrolled 18,705 students — 30% of the study population — and contrast sharply with the typical “bricks and mortar” charter student. Online students tend to be white, ineligible for free and reduced lunches, and entered the online schools with significantly higher test scores. And, online students are more likely to be repeating a grade.
Overall, the results for the cyber schools were terrible. None of the eight online schools did better than their “virtual twins” in traditional district schools. And, the cyber school performance was so poor that it dragged down the results for charters in general.
It’s hard to know what’s causing this poor performance. While the report breaks charter performance down along multiple factors (demographics, years of enrollment, grade level), we don’t see cyber schools isolated within these breakdowns. So, we can’t see the impact of online schools on the overall charter performance in each of these sub-areas, nor can we determine if there are other variables linked to cyber charters that pull their results down. For instance, the recent National Association of Charter School Authorizers brief on authorizing for online charter schools noted that high mobility among online school populations is a challenge when trying to understand these schools’ performance. And, while all of the eight schools performed poorly, it would be interesting to know about the variation of student performance. Did all students tend to perform poorly, or where there sub-populations that did well?
Another interesting angle on the Pennsylvania performance is that the states’ online schools are run by an interesting mix of providers, including those actually sponsored and run by traditional school districts (or consortia of districts, see for example, PA Learners Online). Only three are run by national for-profit education management companies (see full list). I don’t know which eight of the current twelve cyber schools listed are included in the CREDO study, but again, it would be interesting to see whether there were significant differences among types of providers.
Regardless, this pattern of performance is not new. Two years ago, in response to an article noting that at the time, only 3 of 11 Pennsylvania cyber schools met AYP, many of the schools offered familiar excuses. While families engage in cyber schools for a number of reasons, including medical necessity, it’s increasingly clear that Pennsylvania’s cyber schools are falling short on academic measures.
Finally, it’s important to remember that there are an increasingly wide number of online learning models, with full-time online cyber charter schools representing an estimated 20-25% of total online learning students (most are in district or supplemental programs). And, each state offers a unique policy environment for these types of schools. While this study raises serious performance questions about Pennsylvania’s cyber schools, it would be inaccurate to extrapolate broad conclusions about online learning in general.
Disclosure Note: Macke Raymond, director of CREDO, is chair of Education Sector’s board of directors