5.4.4 Robustness checks We perform several robustness checks in the appendix to validate our main results. We perform a robustness check with a sample excluding school year 2019-2020, when the COVID-19 pandemic and its measures were at their peak. Next, we consider day-of-the week effects, which are known to affect student performance (e.g., Gue´rin et al., 1993; Hannon & Dunlop, 2016)). In addition, we perform a robustness check by only considering the months January-February and May-June, which are the most common time periods in which the test is taken. Next, we perform a robustness check with a sample in which tests in all grades are more equally distributed. This sample consists of all students that we follow from first to sixth grade, and shows robust estimated effects. Last, we also consider polynomial weather controls. All these robustness checks can be found in the appendix and show that our analyses yield similar results as our main results. 131
RkJQdWJsaXNoZXIy MjY0ODMw