Table 3.

Predictors of bicycle crashes, multivariable conditional logistic regression, Iowa, 2007-2010, N=294.

	Model 1a		Model 2b
Characteristic	Any Facility		Facility Type
	Adjusted OR	95% CI	Adjusted OR	95% CI
On-road bicycle facility present (ref=none)	0.48	0.18-1.36
On-road bicycle facility type (ref=none)
Pavement markings (bicycle lane or shared lane arrows)			0.40	0.09-1.82
Bicycle-specific signage			0.62	0.15-2.58
Pavement markings & signage			0.36	0.03-4.32
Sidewalks, index street (ref=none) c
Full	2.60	0.95-7.10	2.65	0.96-7.29
Partial	1.66	0.61-4.49	1.66	0.61-4.54
Traffic controls present, non-index street (ref=no) d	1.97	0.80-4.84	1.91	0.77-4.73
Bicycle volume (per 5 bicycles)	1.10	0.73-1.66	1.10	0.73-1.67
Motor vehicle volume (per 10 vehicles)	1.02	0.99-1.05	1.02	0.99-1.05
Curb-to-curb width, index street (per 10 feet) c	1.38	1.06-1.79	1.37	1.05-1.79

Ref=reference

^aModel 1 includes: on-road bicycle facility present, sidewalks, traffic controls, bicycle volume, motor vehicle volume, and curb-to-curb width. Likelihood ratio test: Model 1 vs. Univariate model (χ²=23.42, df=6, p<0.001)

^bModel 2 includes: on-road bicycle facility type, sidewalks, traffic controls, bicycle volume, motor vehicle volume, and curb-to-curb width. Likelihood ratio test: Model 2 vs. Univariate model (χ²=21.46, df=6, p=0.002)

^cIndex street = street motor vehicle was traveling on when crash occurred

^dNon-index street = street motor vehicle was not traveling on when crash occurred