In four experiments, a change detection task using emotional (i.e., angry and happy) faces as stimuli was implemented to investigate the effects of evaluative congruency on working-memory performance and to replicate the angry-face benefit (i.e., better performance for angry compared to happy and neutral faces) found in former studies. Although results of the single experiments were heterogeneous, an overall analysis revealed better performance in trials with evaluatively congruent compared to evaluatively incongruent displays and an angry-face benefit. The congruency effect is in line with recent assumptions that evaluative-priming effects might arise from a mutual facilitation of simultaneously active evaluatively congruent concepts. Research on the angry-face benefit is enriched by the finding that the benefit was also found in control experiments using inverted faces. This result suggests that the effect is based on perceptual features of angry faces.