Table 1

Mixed-effects regression for county-level, female-specific inactivity by region

MidwestNortheastSouthWest
Per cent of exercise-related tweets (logged)−0.205−0.813*−0.421***−0.725***
0.1290.4210.1170.193
Tweet sentiment towards exercise−0.008*−0.019−0.005−0.016**
0.0040.0210.0040.006
Average exercise intensity, via Tweets0.0010.010.0020.0002
0.0020.0070.0020.002
'Fitness centre' Google search index−0.068−0.042*0.0510.015
0.0540.0250.0370.036
'Weight loss' Google search index0.0190.166**0.136***0.055
0.0780.0680.0440.078
Median age−0.019−0.065−0.070***−0.044*
0.020.0590.0180.024
Median household income (in 1000s)−0.094***−0.060***−0.105***−0.044***
0.010.0180.0070.013
Per cent non-Hispanic black0.0110.0320.024***0.274***
0.0240.0440.0060.079
Per cent Hispanic0.041**0.079***−0.028***0.008
0.0160.030.0070.011
Per cent with access to exercise space−0.019***−0.029**−0.027***−0.044***
0.0050.0140.0040.007
Driving death rate0.850***1.099**0.912***1.113***
0.1770.5370.1510.217
Violent crime rate0.0010.0020.00030.002**
0.0010.0020.00040.001
Constant38.536***27.252***30.618***29.222***
5.2677.3464.2015.796
Observations8882141289377
Adjusted R20.6200.5600.6980.629
SD, random intercept2.4041.0751.8172.033
  • Note: *p<0.1; **p<0.05; ***p<0.01.