labour force (such as students, homemakers, or average, 7.5 percentage points higher among retirees) show slightly lower, but not statistically heavy users compared to others. Although modest, this coefficient indicates a meaningful significant, levels of depressive symptoms gap; it is more than twice the difference in compared to employed people. Richer, higher- educated, and religious individuals have a lower propensity to report depressive symptoms. unemployed individuals. Results also indicate that Panel B provides similar information for stress. symptoms than men, whereas people outside the probability models on the association between heavy social media use and our two variables of interest. Depressive during the past six months, and 0 otherwise. Stress (2018-2019) is a binary indicator equal to 1 if the respondent felt stressed often or most of the time during the past six months, and O otherwise. All regressions control for genera-