Source: Authors’ analysis of data from U.S. Census Household Pulse Survey
Note: **p < 0.01, *p<0.05. Coefficients are plotted as point estimates (boxes) with 95% confidence intervals (whiskers). Coefficients are derived from models in which the primary exposure is a triple interaction term between an indicator for whether the interview occurred after (versus before) the CTC expansion, a binary variable representing adults with (versus without) children, and a binary variable for whether the interviewee belonged to a given racial/ethnic group (reference category: White). All regressions adjust for gender, race/ethnicity, income, marital status, number of children, and level of education as well as fixed effects for bi-weekly waves. Depressive symptoms were captured using the Patient Health Questionnaire-2 scale, and anxiety symptoms were captured using the Generalized Anxiety Disorder-2 scale; both were dichotomized at the standard cut-off of 3 or more to indicate high risk of mental health problems.