Table 4.

Treatment effect heterogeneity

	Feel unsafe walking at night	Stay at home because worried about safety	Stay at home because worried about safety or take taxi	# of Nights Outdoors (per week)
Age	0.004	0.000	0.003	− 0.012
	(0.004)	(0.003)	(0.003)	(0.013)
Male	− 0.008	0.059	0.047	− 0.084
	(0.067)	(0.050)	(0.066)	(0.303)
Black	− 0.047	0.073	− 0.047	− 0.028
	(0.109)	(0.077)	(0.143)	(0.436)
Income > $75,000	− 0.010	0.029	− 0.002	0.114
	(0.062)	(0.043)	(0.064)	(0.264)
Crime victim	0.139	0.033	0.072	− 0.84**
	(0.076)	(0.089)	(0.108)	(0.299)

Table presents coefficients along with clustered standard errors in parentheses from the following least squares regression model: Y_i^j = α + β^jD_i + γX_i + ρ^jX_iD_i + ε_i. In the model, D_i is the treatment indicator for the LED version of the photo and X_i is a covariate of interest. The table reports ρ^j along with its standard error for each of five selected covariates and four outcome variables

Significance: *p < 0.05, **p < 0.01