Table 3.
Event study regressions of self-reported face mask use, SARS-CoV-2 reproduction numbers, and growth rates of COVID-19 cases on face mask recommendations. The table reports percentage point changes in self-reported face mask use and growth rates of COVID-19 cases as well as unit changes in reproduction numbers. Data sources: COVID-19 Trends and Impact Survey, Oxford COVID-19 Government Response Tracker, Our World in Data—COVID-19 data set, and Google COVID-19 Community Mobility Reports.
|
|
Self-reported face mask usea,b (observations=13,159) | Reproduction numbera,b (observations=8500) | Case growth in casesa.b (observations=8325) | |||||
|
|
Percentage points (SE) | P value | Units (SE) | P value | Percentage points (SE) | P value | ||
| ATTc | 5.84 (1.60) | <.001 | −0.06 (0.16) | .70 | −2.45 (4.54) | .59 | ||
| Mask mandate | 11.70 (3.44) | .001 | −0.19 (0.09) | .04 | −2.01 (1.36) | .15 | ||
| Subgroup mandate | 13.60 (4.34) | .002 | −0.18 (0.12) | .16 | −3.18 (2.22) | .16 | ||
| Level 2—mandate | 1.16 (3.20) | .72 | 0.05 (0.08) | .54 | 2.22 (1.38) | .11 | ||
| Level 3—mandate | −1.27 (3.88) | .74 | −0.08 (0.10) | .42 | 0.76 (1.99) | .70 | ||
| School closure | —d | — | 0.07 (0.05) | .12 | 1.13 (0.92) | .22 | ||
| Ban on events | — | — | −0.12 (0.05) | .02 | −0.38 (0.77) | .62 | ||
| Ban on gatherings | — | — | −0.02 (0.05) | .70 | −0.70 (1.02) | .50 | ||
| Curfew | — | — | −0.09 (0.04) | .02 | −0.81 (0.66) | .23 | ||
| International travel restrictions | — | — | −0.09 (0.06) | .15 | −1.58 (1.41) | .27 | ||
| Protection of older adults | — | — | 0.04 (0.05) | .42 | −0.78 (1.50) | .61 | ||
| Mobility | — | — | −0.01 (0.01) | .26 | 0.37 (0.098) | <.001 | ||
| Tests per 100,000 inhabitants | — | — | 0 (0) | .20 | 0 (0) | .02 | ||
aAll models have been estimated with country and time fixed effects.
bSE (in parentheses) are clustered at the country level.
cATT: Average Treatment Effect on the Treated. The ATT is obtained by averaging Sun and Abraham [26] interaction-weighted estimates of postmandate period effects.
dThe model was specified with ATT and subgroup as well as level-2 and level-3 mandate effects, only.