Table 2

Logistic regression: odds of 4+ days of parent-reported children’s neighborhood activity as predicted by parent-reported demographics and neighborhood characteristics
Univariate model Full model Final model
OR P OR P OR [95% CI] P
Demographics (reference category)
Child race(White) 1.06 .75 1.19 .41
Child age 0.92 .09 0.91 .08 0.91 [0.82 – 1.01] .08
Child sex (M) 0.73 .04 0.67 .01 0.68 [0.49 – 0.93] .02
Hispanic (N) 0.81 .30 0.85 .48
Child’s BMI % 1.00 .69 1.00 .41
Household income < $50,000 (N) 0.76 .21 0.90 .70 0.87 [0.54 – 1.41] .57
Household income $50-$100,000 (N) 0.87 .36 0.90 .56 0.91 [0.65 – 1.30] .62
Neighborhood environment
Physical activity environment (GIS) 0.83 .23 0.79 .20 0.84 [0.60 – 1.16] .29
Safety from crime 1.17 .17 0.90 .47
Traffic safety 1.77 <.001 1.46 .06
Street connectivity 0.82 .05 0.76 .02 0.77 [0.62 – 0.96] .03
Neighborhood aesthetics 1.68 <.001 1.47 .01 1.57 [1.21 – 2.03] <.01
Walk/cycle facilities 1.05 .60 0.99 .94
Proximity to stores 1.00 .96 0.94 .63
Proximity to play areas 1.19 .07 1.17 .29
Barriers to walking/biking: Logistics 0.68 <.01 0.85 .40
Barriers to walking/biking: Route 0.75 <.01 1.11 .55
Barriers to activity: Perceived lack of appropriate play areas 0.64 <.001 0.79 .12 0.74 [0.58 – 0.95] .02
Barriers to activity: Crime 0.81 .04 1.10 .48
Constant 2.00 .54 4.35 .05

N = 673, Seattle and San Diego, 2007–2009.

Percent of cases correctly classified with this model: 63.5%.

OR Odds ratio.

GIS Geographic Information Systems data.

Tappe et al.

Tappe et al. International Journal of Behavioral Nutrition and Physical Activity 2013 10:39   doi:10.1186/1479-5868-10-39

Open Data