This study examined the effects of ethnic matching on two process variables - Working alliance and perceived therapist credibility - And therapy outcome in 100 ethnic minority clients evenly split between matched and nonmatched therapeutic dyads. All clients and therapists belonged to ethnic minorities in London (South Asian, Black African, Black Caribbean, and Middle Eastern). Clients in the matched dyads had expressed a preference for matching. Outcome and process variables were all significantly better in matched than in nonmatched dyads, and the process variables mediated the relationship between ethnic matching and outcome. Age, gender, and length of therapy did not predict outcome or process. The validity of these results is discussed, and implications for further research, practice, and training are briefly considered.