The Educational Testing Service (ETS) was contracted by the National Center for Education Statistics to investigate changes in tested achievement using cross-sectional comparisons between the 1972 National Longitudinal Study (NLS) seniors and the 1980 High School and Beyond (HSB) senior cohort. The ETS is also responsible for estimating the extent and correlates of cognitive growth that takes place between the sophomore and senior year in high school; this latter analysis involves testing the 1980 HSB sophomore cohort and following them in 1982 with the same test battery. To discharge these responsibilities, an extensive psychometric analysis, reported here, was undertaken. The analysis was designed to: (1) estimate possible effects of differing administrative procedures from one cohort to another on tested performance; (2) evaluate whether the tests were measuring the same things across cohorts or within different administrations within the same cohort; (3) estimate item statistics, reliabilities, and indices of precision of measurement for separate tests and determine whether these measures are invariant across sex and ethnic group; and (4) develop and apply latent trait theory procedures where appropriate. There is little empirical evidence for the notion that the tests or test parcels measure different things for different ethnic or sex groups. The biggest gains over the 2-year schooling period were in vocabulary and reading. White students showed greater gains that did either of the other ethnic groups. There was a consistent tendency for Hispanic women to show greater gains than Hispanic men, and for the test score variance to increase from the sophomore to senior year. Student test booklet samples, 25 data tables, and 22 figures are included. (TJH)