The aim of this study was to analyze the performance of Spanish–English bilinguals on the Golden Stroop Test. The effects of bilingualism, participants' age, age of acquisition of the second language, and proficiency in each language were analyzed. Participants consisted of 71 Spanish–English bilinguals, 40 English monolinguals, and 11 Spanish monolinguals from South Florida. Proficiency in Spanish and English was established using a self-report questionnaire and the English and Spanish versions of the Boston Naming Test. In bilinguals, the Golden Stroop Test was administered in English and in Spanish. Overall, performance was slower in bilinguals than in monolinguals. No significant differences were observed in color reading but bilinguals performed worse in the naming color condition. Even though bilinguals were 5% to 10% slower in the color–word condition, one-way ANOVAs revealed no statistically significant differences between groups. Within the bilingual group, the Stroop Test scores were similar in both English and Spanish. Age of acquisition of the second language did not predict the Stroop Test performance. Repeated measures ANOVA demonstrated a significant interaction between Language Proficiency × Language (in which the test was administered) in some of the ST conditions. In balanced bilinguals, the language used in the ST did not matter, but in unbalanced subjects, the best-spoken language showed better results. In addition, our results support the presence of both between- and within-language interference in Spanish–English bilinguals. Different conceptualization models of the structure of bilingual memory are disclosed. (JINS, 2002, 8, 819–827.)