Skip to main content
. Author manuscript; available in PMC: 2016 Oct 1.
Published in final edited form as: Addict Behav. 2015 May 24;49:33–39. doi: 10.1016/j.addbeh.2015.05.014

Table 3.

Longitudinal Multilevel Regression Model with Frequency of Cigarette Use as the Dependent Variable

Predictor Frequency of Cigarette Use (days used in last month)
Est. (SE) df t p-value
Intercept 1.25 (0.80) 1093.60 1.55 .12
Time (linear) 0.99 (0.06) 1109.96 15.48 <.001
Gender+ 0.54 (0.45) 1106.02 1.21 .23
Ethnicity/Race 0.81 (0.19) 1071.98 4.35 <.001
Time-varying predictors
 Depression (CES-D Total)+ 0.01 (0.02) 5086.99 0.26 .61
 Marijuana use (frequency of
 use in last month)
0.32 (0.02) 5078.23 13.51 <.001
Interactions
 Time*Depression+ 0.01 (0.01) 4751.40 1.88 .06
 Time*Marijuana Use −0.02 (0.01) 4784.57 −2.62 .01
 Gender*Depression+ 0.01 (0.03) 5058.98 0.21 .83
 Gender*Marijuana Use+ −0.05 (0.05) 5122.14 −0.90 .37
 Depression*Marijuana Use+ 0.00 (0.00) 5213.23 −0.04 .97
 Time*Gender*Depression+ 0.00 (0.01) 4751.02 −0.08 .94
 Time*Gender*Marijuana Use+ −0.01 (0.01) 4831.76 −0.46 .65
 Gender*Depression*Marijuana Use+ 0.00 (0.00) 5207.02 0.35 .72

Note. CES-D, Center for Epidemiological Scale-Depression;

+

variable removed from final reduced model.