6 281 EXPLICIT MESSAGING PROMOTES GAME SELECTION however, there was no effect of game choice on the change in affect [F(1, 124) < 0.01, p = .98, η2 p < .01]. This indicates that participants experienced more positive affect after gameplay regardless of playing a game promoted for entertainment or mental health. Exploratory Analyses In order to address the second research aim, we explored the effects of mental health symptom severity and type on all dependent variables. For all groups, we first tried to predict game choice with a logistic regression using symptoms and the interaction between symptoms and trailer design. Next, the remaining dependent variables were predicted using symptoms and the interaction between symptoms and game choice as predictors. All descriptives and statistics can be found in Table 5, with the exception of the three-way interactions for fun and attractiveness and the RM-ANOVAs for change in affect. Symptom Severity For game choice, symptom severity was not found to be related either directly [X2 (1, n = 129) = .86, p = .35] or in interaction with trailer design [X2 (1, n = 129) = 0.51, p = .48]. Thus, participants without severe symptoms were equally likely to select the mental health game (58.0%) as participants with severe symptoms (63.4%). However, further analyses did show effects of symptom severity on competence and autonomy. A direct effect was found for competence, demonstrating that participants with severe symptoms experienced more competence in the game than participants without severe symptoms regardless of game choice. Additionally, both a direct and an interaction effect was found for autonomy. Together these effects show that for those who selected the entertainment game, no difference was found on autonomy for participants with or without severe symptoms [F(1, 125) = 0.02, p = .90, η2 p < .01]. However, when participants selected the mental health game, participants with severe symptoms experienced more autonomy in the game than participants without severe symptoms [F(1, 125) = 11.43, p < .001, η2 p = .08]. Moreover, we observed that affect was lower prior to, but not after, gameplay in participants with severe symptoms than in participants without severe symptoms. Further analyses showed an interaction of time × symptom severity [F(1, 122) = 4.16, p < .05, η2 p = .03], but no three-way interaction: time × game choice × symptom severity [F(1, 122) = 0.83, p = .36, η2 p = .01]. Post hoc tests suggested that there was an increase in affect for participants with
RkJQdWJsaXNoZXIy MTk4NDMw