A fundamental premise in assessment is that the underlying construct is equivalent across different groups of students and that this structure does not vary over years. The pandemic has potentially impacted opportunity to learn and disrupted the internal structure of assessments in various ways. Past research has suggested that students tended to have greater "COVID slide" in mathematics than in reading. Accordingly, this study evaluated the impact of the 2020 pandemic on the measurement invariance of a high-stakes assessment in mathematics. Multigroup confirmatory factor analysis was conducted to explore whether measurement invariance held across the years prior to and immediately after the 2020 pandemic year: 2019 and 2021. In addition, the stability of measurement invariance was evaluated across different ethnic groups as well as groups of students from different district locales such as city and rural areas. Across groups of interest, configural invariance was confirmed as the first step, followed by examination of measurement invariance, including metric, scalar, and residual invariance. Results from the study provided supporting evidence that measurement invariance held reasonably well across the years (2019 vs. 2021) for the assessment under study, despite potential impact of the 2020 pandemic on student learning. The study contributes to a growing body of literature about the degree of impact from the pandemic on test score interpretations.