Skip to main content

Table 2 Zero Inflated Negative Binominal Models of Alcohol Use

From: Distraction towards contextual alcohol cues and craving are associated with levels of alcohol use among youth

 

Logistic Component

Count Component

β [SE]

Exp(β)

Z

P

β [SE]

Exp(β)

z

p

Model 1: Change R2: 0.17 LR χ2(6) = 19.14** (AIC: 514.62)

 Age

−0.248[0.204]

0.780

− 1.217

0.224

0.559[0.036]

1.057

1.540

0.124

 Gender

0.652[0.792]

1.919

0.823

0.411

−0.469[0.192] *

0.626

− 2.434

0.015

 Years of education

−0.037[0.280]

0.964

−0.133

0.894

0.021[0.050]

1.021

0.411

0.681

Model 2: Change R2: 0.11 LR χ2(2) = 12.25** (AIC: 506.36)

 Age

−0.224[0.199]

0.799

−1.128

0.259

0.051[0.034]

1.052

1.510

0.131

 Gender

0.685[0.796]

1.983

0.860

0.390

−0.460[0.182] *

0.631

−2.534

0.011

 Years of education

−0.079 [0.280]

0.924

−0.281

0.779

0.031[0.047]

1.031

0.660

0.509

 Distraction Bias

−0.024[0.016]

0.976

−1.474

0.141

0.010[0.003] **

1.009

2.910

0.004

Model 3: Change R2: 0.30 LR χ2(2) = 37.32*** (AIC: 473.05)

 Age

−0.878[0.412] *

0.416

−2.130

0.032

0.065[0.031] *

1.067

2.086

0.037

 Gender

−1.668[1.564]

0.188

−1.066

0.286

−0.375[0.160] *

0.687

−2.340

0.019

 Years of education

0.543[0.526]

1.721

1.032

0.302

0.033[0.043]

1.034

0.771

0.441

 Distraction Bias

−0.054[0.038]

0.948

−1.426

0.154

0.109[0.028] **

1.008

2.811

0.005

 Craving Scores

−4.099[2.365]

0.017

−1.733

0..083

0.109[0.029] ***

1.115

3.802

< 0.001

  1. Note. N = 106, Gender was coded as male = 0, female = 1; Significant coefficients are in boldface. *p < 0.05, **p < 0.01, p < 0.001***. Final model R2: 0.58, adjusted R2: 0.48. The R2 is a log-likelihood-based coefficient of determination (see method for details). LR χ2(df) Log Likelihood Ratio Chi-squared test, df degrees of freedom, β coefficients SE standard error, Exp(β) exponentiated coefficient. Exponentiated coefficient represent the odds of a structural zero score in the logistic component of the model and levels of use in the count component of the models
Back to article page

BMC Psychiatry

ISSN: 1471-244X

Contact us