Table 2.

Posterior estimates for the association between store count and racial/socioeconomic covariates (N=97)

Fixed effect	Model 1		Model 2		Model 3		Model 4
	Adjusted effect		Adjusted effect		Adjusted effect		Adjusted effect
	Mean (RR)	95% CI	Mean (RR)	95% CI	Mean (RR)	95% CI	Mean (RR)	95% CI
Intercept	1.06	0.87–1.30	0.78	0.42–1.47	1.01	0.48–2.06	0.99	0.39–2.49
Entropy index			0.61	0.18–2.06			0.55	0.15–2.01
High White clustering (Ref=low)			1.92	0.97–3.85			1.86	0.86–4.08
High Black clustering (Ref=low)			0.87*	0.55–0.98			0.87	0.44–1.74
Percentage — Asian population			1.02	0.99–1.05			1.01	0.97–1.05
Percentage — Hispanic			1.02*	1.00–1.03			1.02*	1.00–1.04
Percentage — population below poverty					1.01	0.99–1.03	1.01	0.99–1.03
Percentage — household SNAP recipient					0.98*	0.96–0.99	0.98	0.96–1.01
Percentage — vacant housing					0.99	0.98–1.02	0.98	0.96–1.02
Percentage — household with no vehicle					1.01	0.99–1.02	1.01	0.99–1.03
Percentage — college educated					0.99	0.96–1.02	0.99	0.96–1.02
Random effect
Spatially structured effect	4.56	1.10–2.07	0	0.00–3.88	1.02	0.00–2.04	4.37	0.09–5.35
Spatially unstructured effect	0	0.00–1.67	0	0.00–1.79	1.72	0.00–1.50	5.26	0.00–8.35

The table shows the posterior estimates of the adjusted effects of the covariates on the store count, using four different Bayesian spatial models. The mean (RR) column shows the mean of the posterior distribution of the rate ratio (RR) for each covariate, and the 95% CI column shows the 95% credible interval of the RR. An RR above 1 indicates a positive association between the covariate and the store count, while an RR below 1 indicates a negative association. An asterisk (*) indicates that the 95% credible interval does not include 1, which suggests a statistically significant association. The random effect shows the posterior estimates of the spatially structured and unstructured effects, which capture the spatial dependence in the store count due to the census tracts neighboring structure