Data from the 1980 and 1982 High School and Beyond Surveys are used to compare the achievement of immigrant and native students in high school. The study examines effects of generation and duration of residence in the United States on improvement in students' performance on standardized tests over a 2-year period beginning with the year the students were sophomores in high school. The results of the multivariate analyses suggest that even controlling for family background and ethnic group, recent immigrants have the lowest test scores (in 1980). However, when the trajectory of achievement is examined, as measured by increases in test scores between 1980 and 1982, generation status is no longer significant. Thus, the best predictors of achievement during the high school years are not those based on the nativity of the student, but rather those that reflect the social environment experienced in the United States, such as ethnicity, family socioeconomic status, and previous school performance. Recent immigrants, whether or not they are old for their grade, "catch up" with peers of similar age over the 2 years when other background characteristics are controlled. (Contains 5 tables, 4 figures, and 33 references.) (SLD)