Rational emotive behavior therapy (REBT) demonstrates good results in evaluation therapy researches. However, some of its basic concepts, as well as theory as a whole itself, did not receive satisfactory empirical support so far, in comparison to other cognitive models (Beck, Lazarus etc.). Quasiexperimental study was designed to test the role that (1) negative feedback (A) and (2) irrational beliefs (B) both play in formation of dysfunctional negative emotions, in the context of significant personal goals (in our case value of potential award - G). ABC theoretical model received limited support: statistically significant three-times interaction A x B x G was found in predicting general negative emotional state, as well as anger. In contrast with that, ANOVA showed only main effect of irrational beliefs (as continuous variable) to be significant in predicting emotions of anxiety and depression. Findings are discussed in the context of REBT theory of emotions, as well as their possible practical applications. Limitations of the study were also mentioned. .