Changes in mathematics achievement at grade 11 before and after the implementation of the Texas Educational Assessment of Minimum Skills (TEAMS) were studied in an urban school district in Texas. High stakes tests are used by state education agencies and school districts to monitor and increase educational achievement. The TEAMS is a minimum skills testing program; mastery of its exit level became a graduation requirement in 1987. Performance of students in the district was examined over a span of 5 years with 12,404 11th-graders (about 2,400 each year). For each subject, a percent correct score was determined for basic skills and high level skills categories of the Tests of Achievement and Proficiency. The grade point average of each student was included as a covariate to control for differential effects of instruction on students with varying previous achievement. Multivariate analysis of covariance was used. Over the period, basic skills scores did improve. It had been hypothesized that high level scores would decrease, as less attention was paid to higher achieving students when teachers concentrated on minimum performance. High level scores did not decrease, but they did not improve significantly. In addition, in the final year considered, after four years of concentration on the minimum competencies, both basic and high level scores declined. Implications for testing are discussed. One table and one graph present study data. (SLD)