As the idea of including limited English proficient (LEP) students in large-scale assessments is becoming popular among researchers, the practical question of how to do so meaningfully so that the needs and performances of LEP students are recognized while their scores are interpreted and used correctly becomes more important. This paper presents the experience of the Dallas Public Schools (Texas) in finding a solution to these problems. A program was designed with the help of a commercially developed test, the Woodcock-Munoz Language Survey. All enrolled LEP students in Dallas schools are first required to take this language assessment and then are assigned to the state-required criterion test, the Texas Assessment of Academic Skills, and/or the district testing program. In this way, in the 1995-96 school year, 96% of LEP students participated in at least one of the academic testing programs. The process of setting up guidelines and the initial implementation of the approach are described, and the district's current policies in assessing LEP students are outlined. The district's administrators feel that conditional inclusion is the key to success with the LEP testing program. Baseline results have indicated that the tests provide good measures of student skills and can be used to evaluate the effectiveness of bilingual education and English-as-a-Second Language programs. (Contains 4 figures, 7 tables, and 13 references.) (SLD)