Table 3.

Relationship between perceived safety from crime and walking inside the neighbourhood (min/week)

Variable	Model 1 Demographics		Model 2 Built environment		Model 3 Social cohesion		Model 4 Perceptions
	Estimate (SE)	p	Estimate (SE)	p	Estimate (SE)	p	Estimate (SE)	p
Total walking	18.04 (3.50)	0.0001	18.77 (3.50)	0.0001	13.53 (3.61)	0.0002	10.54 (3.97)	0.0079
Walking for recreation	13.51 (2.36)	0.0001	13.70 (2.37)	0.0001	10.25 (2.46)	0.0001	7.01 (2.70)	0.0096
Walking for transport	3.18 (1.27)	0.0127	3.56 (1.27)	0.0051	1.38 (1.31)	0.2922	0.68 (1.44)	0.6378

Proc Mixed marginal model with unrestricted variance pattern

Model 1 adjusts for age, gender, income, education, marital status, importance of safety from crime to neighbourhood selection and time

Model 2: Model 1 + residential density, street connectivity and local destinations

Model 3: Model 2 + perceptions of neighbourhood social cohesion

Model 4: Model 3 + perceptions of aesthetics, traffic and street lighting

Bold denotes significant p-value