Skip to main content

Table 3 Robustness tests

From: The linkage between the perception of neighbourhood and physical activity in Guangzhou, China: using street view imagery with deep learning techniques

 

Model 2

Model 3

Model 4

Coef. (SE)

Coef. (SE)

Coef. (SE)

Wealthy

0.854 (1.025)

0.966 (1.033)

1.189 (0.977)

Safety

1.343** (0.558)

1.483** (0.663)

1.585** (0.647)

Lively

1.445** (0.689)

1.372** (0.660)

1.688** (0.815)

Depressing

− 1.181** (0.523)

− 1.239** (0.528)

− 1.231** (0.608)

Boring

− 1.277** (0.608)

− 1.209** (0.601)

− 1.496** (0.747)

Beautiful

1.229** (0.497)

1.297** (0.481)

1.166** (0.499)

Number of observations

769

763

808

  1. Coef. coefficient, SE standard error
  2. Significance levels: “*” p < 0.100, “**”, p < 0.050, “***” p < 0.010. All models were adjusted for individual-level covariates. Model 2 excluded people who had functional restrictions, Model 3 excluded people aged > 70. Model 4 used a circular buffer with a radius of 1.5 km
Back to article page

International Journal of Health Geographics

ISSN: 1476-072X