Table 2 Regression-estimated impact of being selected as a treatment zip code in the Philly Vax Sweepstakes on weekly first-dose COVID-19 vaccinations per 100,000 people

	Model 1		Model 2		Model 3		Model 4
	β	P	β	P	β	P	β	P
Treatment Zip Code 1 during treatment (19126)	61
	(20)	0.008
	[38]	0.133
	{65}	0.350
Treatment Zip Code 2 during treatment (19133)			19
			(17)	0.288
			[48]	0.705
			{54}	0.728
Treatment Zip Code 3 during treatment (19142)					−102
					(35)	0.010
					[66]	0.148
					{89}	0.253
Treatment zip codes during treatment (pooled)							−4
							(40)	0.929
							[45]	0.936
							{45}	0.936
Observations	270		270		270		300
R²	0.93		0.93		0.93		0.93

This table reports a series of difference-in-differences models relying on ordinary least squares regressions to predict a zip code’s weekly first-dose COVID-19 vaccinations per 100,000 adult residents. The predictor variables in each regression include zip code fixed effects, week fixed effects and an indicator that takes on a value of 1 during the weeks when a treatment zip code of interest was eligible for rewards and 0 otherwise. Standard errors have been estimated clustered by zip code (first, in parentheses), clustered by week (second, in brackets) and robustly without clustering (third, in braces) for all four models. Models 1–3 include 18 zip code clusters and 15 week clusters. Model 4 includes 20 zip code clusters and 15 week clusters. All t-tests are two-sided.

Quick links

Search