Table 3.

Estimates, posterior standard deviations (SDs), and 95% credibility intervals (CI) from multilevel structural equation model of stress, mutliple precipitants, and smoking lapse

Variable		Est.	SDa	p	[95% CI]b
Stressi(j-1→j) → Smoking Lapsei(j→j+1)	(c’ )	−.043	.029	.076	[−.104, .012]
Stressi(j-1→j) → Precipitants	(a paths)
Negative Affectij	(a1 )	.199	.015	.000	[.170, .229] *
Smoking Urgeij	(a2 )	.160	.017	.000	[.132, .198] *
Abstinence Self-efficacyij	(a3 )	−.056	.017	.000	[−.089, −.025] *
Motivation to Quitij	(a4 )	−.061	.013	.000	[−.089, −.037] *
Difficulty Concentratingij	(a5 )	.170	.018	.000	[.134, .202] *
Coping Expectanciesij	(a6 )	−.076	.018	.000	[−.109, −.042] *
Smoking Expectanciesij	(a7 )	.108	.018	.000	[.070, .139] *
Precipitants → Smoking Lapsei(j→j+1)	(b paths)
Negative Affectij	(b1 )	.248	.036	.000	[.175, .319] *
Smoking Urgeij	(b2 )	.430	.028	.000	[.376, .485] *
Abstinence Self-efficacyij	(b3 )	−.188	.026	.000	[−.239, −.134] *
Motivation to Quitij	(b4 )	−.173	.036	.000	[−.247, −.109] *
Difficulty Concentratingij	(b5 )	.086	.022	.000	[.042, .131] *
Coping Expectanciesij	(b6 )	−.099	.023	.000	[−.143, −.052] *
Smoking Expectanciesij	(b7 )	.114	.021	.000	[.073, .158] *
Smoking Lapsei(j-1→j) → Smoking Lapsei(j→j+1)	(ar1 )	.255	.015	.000	[.228, .284] *
Indirect Estimatesc
Negative Affectij	(a1 × b1 )	.047	.010	.000	[.026, .067] *
Smoking Urgeij	(a2 × b2 )	.070	.011	.000	[.050, .094] *
Abstinence Self-efficacyij	(a3 × b3 )	.018	.007	.008	[.003, .032] *
Motivation to Quitij	(a4 × b4 )	.004	.007	.280	[−.011, .018]
Difficulty Concentratingij	(a5 × b5 )	.010	.008	.091	[−.003, .026]
Coping Expectanciesij	(a6 × b6 )	.020	.006	.000	[.008, .033] *
Smoking Expectanciesij	(a7 × b7 )	.016	.007	.010	[.001, .029] *

Notes. Est. = undstandardized estimate (probit); p = Bayesian one-tailed p -value; i indexes participants (N=370) and j indexes EMAs (N=32,561); paths control for the association between previous lapse and precipitants, the passage of time, and the length of time between intervals

^a= for positive values, the one-tailed p -value is the the proportion of posterior distribution above 0

^b= significance

^*is determined by 95% asymmetric posterior distributions that do not include 0.

^c= cov(a, b ) added to each indirect estimate.