Results of a study of the efficacy of foreign language instruction at the elementary school level are presented. The study provides new information permitting comparison of three program types--Foreign Language in the Elementary School (FLES), partial immersion, and immersion--currently used in schools in the United States. The results are based on achievement in listening, speaking, reading, and writing in French and Spanish as measured by the Modern Language Association (MLA) Cooperative Foreign Language Proficiency Tests. The following research questions are addressed in the study: (1) How do different foreign language program types affect student performance on a standardized language achievement test in terms of total time or intensity of instruction and use or non-use of the foreign language to teach core curriculum subjects? (2) How does variation among schools within a given language program affect performance on the MLA test in terms of (a) articulation or continuity at the individual school level and within the school system, (b) socioeconomic status of students, and (c) number of years of program operation at the individual school level? (3) What is the correlation between self-assessment of language skills and performance on the MLA test? Overall, the results show immersion programs to result in student test performance significantly higher for all skills than FLES programs, and partial immersion programs to result in test performance at a level between total immersion and FLES programs. Policy implications based on the variety of interactions found are discussed. (MSE)