This study was conducted to examine the problem of attrition from a public middle school foreign language enrichment program by students who were admitted on the basis of superior grades, test scores, and recommendations from their teachers, counselors, and parents. The study took place in inner-city San Antonio and involves Latino sixth and seventh graders from mostly low-income families. Literature pertaining to school choice options, education of Latino students, and student attrition was reviewed. Research questions pertained to the differences in characteristics of students staying in the program and leaving it and in the reasons students gave for their decisions to stay or leave. In addition, the efficacy of an existing student attrition model, modified for this study, was tested for organizing data. Data sources included surveys of students and teachers, interviews with administrators and counselors, and school records. Logit regression analysis revealed two factors linked to student persistence in the program to be significant to the .01 level: student involvement in the initial decision to apply to the program, and the presence of a student's best friend at the school the student attended. A third variable approached significance (at the . 10 level): the student's score on the math …