African American women (N = 100) recruited from an urban clinic were randomly assigned to view 1 of 3 experimental videotapes promoting HIV testing: (a) an ethnicity-matched information control videotape; (b) the same ethnicity-matched videotape presented by an African American woman (gender-ethnicity-matched control condition); or (c) an experimental videotape with a culturally relevant context that embedded HIV-testing information within a frame of reference emphasizing personal loss. Consistent with D. Kahneman and A. Tversky's (1979) prospect theory, women who viewed the context-framing videotape were most likely to have been tested during a 2-week follow-up interval. Among women who expressed intentions to get tested after viewing the videotapes, 63% of those in the message-framing condition were tested for HIV during a 2-week period compared with 23% in the gender-ethnicity-matched condition, and none in the ethnicity-matched condition.