In the Louisiana Basic Skills Testing (BST) program, initiated in 1982, criterion-referenced tests of language arts and mathematics were administered to all second grade students. Longitudinal data over three years tracked four cohorts of students to address when retention appears to be most effective and whether a combination of repeated classroom instruction and remedial services can correct such deficiencies. The issue of relative effects of early or later retention was addressed by comparing the performance of promoted-then-retained and retained-then-promoted students using the Statistical Analysis System general linear models procedure to compare 1984 grade 3 performance of the two groups with initial 1982 grade 2 performance differences controlled. The issues of promotion and then mastery of subsequent grade level skills, as well as the prevention of later academic difficulties were addressed by computing the numbers and percents of students meeting the performance standards of the BST and the mean BST scores of promoted and then retained group of students. The results indicate that a combination of remedial instruction and a repeated year of classroom instruction appears adequate for correcting skill deficiencies among the majority of students. (PN)