TABLE A2.
Robustness check for Difference‐in‐differences Estimates: Excluding a 3 week window pre and post week 11
| Main effects | Call resolution | |||||
|---|---|---|---|---|---|---|
| Calls | First‐time callers | Resolved | Prescription | Follow‐ up | Referral | |
| Post × Year2020 | 2.465*** | 2.106*** | 2.504*** | 3.337*** | 3.217*** | 1.930*** |
| (0.074) | (0.080) | (0.078) | (0.122) | (0.128) | (0.117) | |
| Week F.E | Yes | Yes | Yes | Yes | Yes | Yes |
| Day of week F.E | Yes | Yes | Yes | Yes | Yes | Yes |
| Observations | 616 | 616 | 616 | 616 | 616 | 616 |
| Adjusted R 2 | 0.971 | 0.952 | 0.967 | 0.958 | 0.944 | 0.912 |
| Average before week 11 2020 | 30.58 | 18.42 | 19.28 | 5.77 | 2.52 | 3.56 |
Note: As a robustness test by excluding 3 weeks before and after week 11. Each column presents the results of the difference‐in‐differences specification for a different dependent variable, estimating θ in Equation (2) using ordinary least squares. The dependent variables used in these models are (from left to right): log (number of calls), log (number of first‐time callers), log (number of resolved calls), log (number of calls resulting in prescription), log (number of follow‐up calls + 1), and log (number of referrals + 1). All models include week fixed effects (F.E.) and day‐of‐the‐week fixed effects (F.E.). The last line shows the average of the dependent variable in levels (i.e., not in logs) before the implementation of the mobility restrictions. * statistically significant at 10%, ** at 5%, *** and at 1%.