Learning engagement varies in different modes of instructions. With such consideration, the authors of this article conducted comparative analyses on student learning outcomes measured by their final mean grade point averages (GPAs) during three timeperiods (before, at the onset of, and during pandemic) when inperson, online, and hybrid instructions were delivered. The records of 600 middle school students in a school district from a western state, across three academic years, were used as the sample (2018-2019 before pandemic when they were in sixth grade with in-person instruction, 2019-2020 at the onset of the COVID-19 pandemic in seventh grade with online instruction, and 2020-2021 during pandemic in eighth grade with both online and hybrid instructions). Results showed that no difference was found (from paired t-test) in student final GPAs when they transferred from in-person instruction before pandemic to online instruction at the onset of pandemic, but differences were found (a) between in-person instruction and hybrid instruction (from one-way repeated ANOVA), and (b) between online instruction and hybrid instruction (from both one-way repeated ANOVA and an independent samples t-test) at and during pandemic. Recommendations for practice and future research focusing on students' needs were presented.