Using the best-fitting model, which included session as a random slope effect, we also found a significant Group × Session interaction effect on positive affect, as measured by the PANAS positive subscale, F(3, 323) = 3.78, p < .05 (see Figure 4). We found that abstainers (simple slope = 0.53) showed a significant increase in positive affect across the pre- and postcessation sessions as compared with relapsers, simple slope = —0.84; t(323) = —2.13, p < .05, and never-quitters, simple slope = —1.29; t(323) = —3.13, p < .01. No difference in positive affect changes was found between control participants and the other three groups of smokers.
