Does testing environment matter for virtual school students?

• Students in virtual schools perform better on tests when allowed to test from home. • At-home testing enhances the concurrent validity of tests. • Evidence is inconclusive as to whether better outcomes are explained by cheating. Standardized test performance among students in full-time virtual schools generally lags behind that of students in traditional, face-to-face schools. One largely unexplored potential cause is the requirement to complete tests at unfamiliar facilities rather than from home, the normal learning environment for these students. The COVID-19 pandemic enabled some full-time virtual school students to test from home, providing a unique opportunity to assess how testing environment might influence outcomes. Using data from an educational management organization (EMO) running virtual schools in 21 U.S. states, we test whether full-time virtual school students perform better on standardized exams administered at home ( n = 524) relative to those testing at an unfamiliar testing site ( n = 1,443) in 2020–21. Home testing is associated with nearly one fifth of a standard deviation improvement in math and ELA performance even after controlling for prior test performance. Moreover, tests taken from home more strongly correlate with other measures of academic performance than do tests taken at designated facilities, indicating they better capture academic progress. Analyses investigating whether home testing enables cheating are inconclusive.