For analysis of the postquit in-session ratings data, we ran models containing postquit WSWS Craving and its interactions with EOT group status, slide valence (of the slide immediately preceding the in-session ratings), and session, with baseline in-session ratings as covariates, separately for each of the in-session rating DVs. We found significant WSWS Craving × EOT Group Status interactions for in-session craving, F(2, 1009) = 31.27, p < .0001; negative mood, (2,1009) = 6.99, p < .001; and positive mood ratings, F(2, 1009) = 9.21, p < .0001 (see Table 5). Post hoc comparisons indicated that the slope for in-session craving ratings was larger for relapsers compared to abstainers and controls, and the slope for abstainers was significantly larger than that of controls. For in-session negative mood ratings, each unit increase in WSWS Craving score led to significant slope increases in relapsers and controls compared to abstainers. Conversely, for in-session positive mood ratings, each unit increase in WSWS Craving score let to significant slope decreases in relapsers and controls compared to abstainers. There were no other significant effects involving postquit WSWS
