COVID-19 has upset the traditional ways of teaching and looks set to continue to have an impact in the coming years. The situation in Taiwan--partial disruption and then an early return to classes--allows for a comparison of student performance in a conventional teaching and testing environment with the non-contact teaching/testing environment caused by COVID-19. For the part of the general physics class delivered remotely, test marks dropped uniformly across the whole class from 69% to 46% from the previous year. Poorer knowledge acquisition was not just limited to poor students but surprisingly equally affected those at the top of the class as well. In contrast, remote testing resulted in a higher overall grade--due to ~5% of the ~70 students enrolled scoring significantly better on unsupervised tests (>93% vs ~40% on three supervised tests). The marks for the remaining 95% of students were consistent with their overall performance. An advantage of moving to online lectures is the ability to archive lectures. Our research indicates that students take advantage of these online lectures: both in reviewing in the days immediately after the class and immediately prior to testing.