In three experimentswe examined whether songs that are widely used in times of crisis (crisis songs, CS) could resume their unifying effect when they are played offcontext. In the first experiment, two conflictual groups, religious and secular Jews, were exposed t o CS, t o love songs (LS), or t o no songs and were then asked t o express their attitudes towards their outgroups. It was found that CS positively affected respondents' attitudes: stigmas and prejudice were lessened. In a second experiment, the CS effect was examined under more restrictive conditions. Instead of listening t o CS, participantswere asked t o recall them from memory. In addition, attitudes towards ingroups and outgroups were collected more systematically. Results showed again that CS reduced intergroup bias. In a third experiment we tried t o understand the mechanism underlying the CS effect by examining the thoughts and associations that people had while listening t o CS. Analysis of the associations showed that unifying themes such as "nationalism," "sorrow and grief," and "unity" were most prominent when religious and secular respondents listened t o CS. The CS effect and its underlying mechanism are explained in light of the "common ingroup identity model." CS had the power to bring into awareness that the conflicting groups belong t o one superordinate social group which, in turn, reduced stigmas, prejudice, and intergroupbias. Possibleimplications of these findings are discussed.