This study was designed to find out whether a person’s relationship with his voices and the negative affect he suffers from are mediated by beliefs about the voices. Research done to date shows contradictory results (Sorrell, Hayward, & Meddings, 2010, Vaughan & Fowler, 2004). A cross-sectional study was done to study the associations among variables, and a multiple mediation model (Preacher & Hayes, 2008) in which the beliefs about voices were the mediating variables was tested. Sixty subjects who heard voices participated. The VAY (Hayward, Denney, Vaughan, & Fowler, 2008), BAVQ (Chadwick & Birchwood, 1995), BAI (Beck & Steer, 1993) and BDI-II (Beck, Steer & Brown, 1996) were given. We found a significant positive correlation between perception of voices as dominant and intrusive and maintaining a position of distance from them on one hand, and negative affect [anxiety (r = .57, p < .001; r = .40, p < .001; r = .34, p < .01 respectively) and depression (r = .58, p < .001; r = .37, p < .01; r = .38, p < .001 respectively)] on the other. We also found that beliefs of malevolence and omnipotence mediated between relating style and negative affect (anxiety and depression). The theoretical implications of the results and clinical implications of the mediating relationships found are discussed.