Assessing urban and rural neighborhood characteristics using audit and GIS data: derivation and reliability of constructs

Table 4 Standardized Cronbach's alpha and confirmatory factor analysis for constructs derived from EFA and Pearson correlation coefficients, exploratory sample (n = 6,388)

		Standardized Cronbach's Alpha^§		Goodness-of-fit
Model	# items	Urban (n = 4,533)	Rural (n = 1,846)	Urban				Rural
				# parameters	CFI > 0.95	TLI > 0.95	RMSEA < 0.05	# parameters	CFI > 0.95	TLI > 0.95	RMSEA < 0.05
1-factor
Arterial or thoroughfare	13	0.73	0.64	34	0.922	0.927	0.003	33	0.910	0.904	0.003
Walkable neighborhood	11	0.67	0.60	32	0.881	0.909	0.003	31	0.924	0.930	0.002
Physical incivilities	9	0.61	0.69	21	0.980	0.980	0.002	22	0.934	0.922	0.002
Decoration	3	0.43	0.48	9	1	1	0	9	1	1	0
4-factor	32	-	-	93	0.77	0.79	0.003	91	0.63	0.63	0.004
Pearson correlations				Arterial or toroughfare	Walkable nighborhood	Physical incivilities	Decoration	Arterial or thoroughfare	Walkable nighborhood	Physical incivilities	Decoration
Arterial or thoroughfare				1				1
Walkable neighborhood				-0.07	1			-0.26	1
Physical incivilities				0.25	-0.12	1		0.00	-0.55	1
Decoration				-0.19	-0.24	-0.38	1	0.16	-0.07	-0.35	1

§Data are weighted by the road segment's length.
In the 4-factor model, there were 32 items instead of 36 because 4 items (#31 sidewalk buffer, #30/32 sidewalk condition, #36 pedestrian oriented lighting, and #43_5 presence of pedestrian yield signs) loaded on several factors.
Abbreviations: RMSEA root mean square error of approximation, CFI comparative fit index, and TLI Tucker-Lewis
*These are "rules of thumb" guidelines to interpret these indices (see paper for more information).

ISSN: 1479-5868