Skip to main content
. 2020 Jun 22;20:977. doi: 10.1186/s12889-020-09063-z

Table 4.

Multilevel mixed-effects logit estimation: prescription drug abuse recognition

Sig. Exp (B)
Respondent Variables:
 Own Prescription Drug Abuse 0.000 0.345
 Friend/family Prescription Drug Abuse 0.169 1.411
 Male 0.006 0.541
 White 0.085 1.582
[Age base = 35–54]
 Age 18–34 0.060 0.649
 Age 55+ 0.000 2.909
[Respondent Income: base = $50–75)]
 $25 K or less 0.466 1.299
 $25-50 K 0.492 1.250
 $75-100 K 0.976 1.010
 $100 K+ 0.561 1.206
[Respondent Education: base = HS or less]
 Some college 0.018 2.019
 College degree 0.024 2.034
 More than College 0.787 1.099
[Sufferer in Vignette: base = Michael]
 Michelle 0.302 1.236
Regional Level Variables:
 Northeast 0.279 0.644
 Midwest 0.637 1.190
 West 0.655 1.251
County Level Variables:
 Rural 0.390 1.408
 Unemployment 0.833 0.986
 Association Rate 0.415 0.789
 Percent fair or poor health 0.829 0.982
 Percent access to exercise 0.328 1.008
 Percent uninsured 0.642 0.986
 Percent frequent physical distress 0.522 0.852
 Percent frequent mental distress 0.098 1.353
 Percent lack of sleep 0.238 0.939
 Percent physically inactive 0.069 1.083
[County Quartile Mental Health Providers Per Capita: base = lowest]
 2nd Quartile 0.309 0.749
 3rd Quartile 0.216 0.656
 4th Quartile 0.027 0.421
 Percent excessive drinking 0.458 0.952
 Percent alcohol driving deaths 0.169 1.015
 Violent crime rate 0.179 0.999
 Percent child poverty 0.739 0.990
 Legalized Marijuana State Law 0.900 1.043
 Constant 0.960 1.155
Number of observations 590
Number of groups (counties with observations) 392
Log Pseudolikelihood − 292.352