. Author manuscript; available in PMC: 2013 Dec 1.

Published in final edited form as: Psychol Addict Behav. 2012 May 28;26(4):850–858. doi: 10.1037/a0028375

Table 3.

Analysis of GEE parameter estimates based on empirical standard error estimates, using an exchangeable working correlation structure, with smoking abstinence as outcome variable, and task persistence, time, diagnosis, ability, FTND score, and the interaction between disorder and persistence as explanatory variables.

		Standard	Wald χ²
Parameter	Estimate	Error	(df=1)^a	p-value	OR	95% CI
Intercept	−4.319	1.207
Mirror Tracing Persistence^*	0.014	0.006	8.69	0.003	1.014	1.014 – 1.017
Time^**	2.083	0.327	52.77	<0.001	8.029	2.114 – 237.81
Diagnostic Group	0.188	0.557	0.11	0.740	1.207	0.991 – 1.019
Confidence in ability to quit	0.051	0.053	0.93	0.334	1.052	1.047 – 1.176
FTND Score	0.077	0.092	0.70	0.401	1.080	1.065 – 1.332
Persistence X Diagnostic Group	−0.009	0.006	2.15	0.143	0.991	0.991 – 1.019

p < .05,

^**

p < .0001,

df=3 for variable Time