Based on a theoretical framework describing culturally sensitive (CS) health communication, this experiment tested the relative contributions of surface structure and deep structure in the recall of oral health information from pamphlets varied in written message and images. Using a 2 × 2 factorial design, Spanish-speaking Mexican heritage mothers of children under six (n = 160) were randomly assigned to read one of four 12-page pamphlets containing the same oral health information in Spanish: (1) standard written message/standard images; (2) standard written message/CS images; (3) CS written message/standard images; and (4) CS written message/CS images. Participants completed a 22-item oral health knowledge questionnaire before and after reading the pamphlet. Controlling for the effects of pretest scores, acculturation, and educational level on information recall, findings showed significant positive main effects for CS images (F(1, 152) = 5.03, p = .026, partial ŋ2 = .032) and CS written message (F(1, 152) = 5.21, p = .024, partial ŋ2 = .033). There was no interaction. These results support the two dimensions of CS and their independent effects. They should be applicable to a variety of health communication channels. Further research is needed to investigate the causal mechanism behind the observed effects.