Characteristics of Hispanic students referred for special education services during 1983-84 in several large urban school districts with large minority populations were examined. Also examined was the predictive ability of various statistical models with respect to the eventual diagnostic outcomes of the students. File data were collected for 1,319 Hispanic students in grades K-12, with the analyses focusing on the 1,154 elementary students. Findings showed that the majority of the referrals were in the early elementary grades, were male, and were born in the United States, while the majority of parents were born in Mexico. Most frequent reasons for referral were low academic achievement and reading problems, followed by poor oral skills. The most frequent eventual diagnostic classifications were learning disabled (63% of the sample) and language impaired (20%). A predictive analysis using path analytic procedures examined factors leading to eventual classification, and accounted for about 40% of the variance in the dependent variable. In addition, there appeared to be two "tracks" leading to the eventual classification, one for diagnosis of learning disabilities and one for diagnosis of language impairments. Appendices include the data collection instrument, a codebook for variables, and descriptive profiles of participating school districts. (JDD)