Skip to main content
Fig. 1 | International Journal of Behavioral Nutrition and Physical Activity

Fig. 1

From: Municipal investment in off-road trails and changes in bicycle commuting in Minneapolis, Minnesota over 10 years: a longitudinal repeated cross-sectional study

Fig. 1

Multivariable-adjusted regression estimatesa for the differenceb in the percentage of workers commuting by bicycle in 2000 and 2010 according to joint levelsc of the distance (km) between the tract and the trail systemc and proportion of commuting trips that cross the trail systemd. aRegression models included: time-varying and tract-level variables for distance to trail system, proportion to work-related trips that cross the trail system, total work-related trips, intersection density, population density, median household income, professional workforce, workforce aged 13–34 years, total length of bicycle lanes, maximum reach of bicycle lane network, maximum reach of network comprising both bicycle lanes and off-road trails, and the time-invariant variable for commuting by bicycle in 1990. Estimated effects for changes for all time-varying variables were modeled by including a year* variable interaction term. bDifferences were obtained using the ‘margins’ post-estimation command following repeated-measures random effects linear regression models (-xtreg-) in Stata. c P-value for interaction = 0.06. dLevels of predictor variables reflect the percentiles of the variable distribution for combined 2000 and 2010 data. For distance: 25th = 1.08 km; 50th = 2.83 km; 75th = 5.91 km. For proportion of commuting trips that cross the trail: 25th = 0.11; 50th = 0.29; 75th = 0.42

Back to article page