Table 5.

Multiple regression model for built environment factors and social interaction of older people.

Models		B	S. E.	Sig.	95% CI	q (FDR)	VIF	R	R2	ANOVA
										F	Sig.
1	Interpersonal interaction	←				BE Factors
	Constant	−0.478	0.514	0.355				0.281	0.079	4.975	0.008
	BE6 Functional facilities	0.328	0.118	0.006	[0.097, 0.559]	0.012	1.151
	BE8 Privacy	−0.224	0.093	0.018	[−0.406, −0.042]	0.018	1.151
2	Activity engagement	←				BE Factors
	Constant	−1.749	0.380	0.000				0.400	0.160	22.228	0.000
	BE7 Accessibility	0.428	0.091	0.000	[0.250, 0.606]	q < 0.001	1.000
3	Caregiver relationship	←				BE Factors
	Constant	−1.265	0.397	0.002				0.289	0.083	10.659	0.001
	BE7 Accessibility	0.310	0.095	0.001	[0.124, 0.496]	q = 0.001	1.000
4	Conflict	←				BE Factors
	Constant	3.210	0.585	0.000				0.459	0.211	15.469	0.000
	BE2 Recreational area	−0.406	0.150	0.008	[−0.700, −0.112]	0.016	1.657
	BE4 Lighting	−0.313	0.151	0.041	[−0.609, −0.017]	0.041	1.657

S. E. = standard error; Sig. = significance; VIF = variance inflation factor. Multiplicity and effect sizes. For each model, p-values are accompanied by Benjamini–Hochberg FDR–adjusted q-values. Effect sizes are shown as unstandardized coefficients with 95% confidence intervals. Coefficient–significance matrix of all predictor was shown in Supplementary Table 3.