This study was undertaken to evaluate models that could be used to set single-year individual student academic growth targets. Multiple terms of individual student reading and mathematics test results were analyzed to predict each student's final status score in each subject. Test records from more than 5,300 students in 3 cohorts were used; 2 cohorts of roughly 670 to 750 students and 1 cohort of roughly 4,000 students. The two smaller cohorts were from the same school districts; the larger one was from the 2002 Northwest Evaluation Association norming study and represented nine school districts. Three terms of spring data were used to predict scores in a fourth spring term for the two smaller cohorts. For the larger cohort, four terms of fall data and four terms of spring data were used to predict scores in a fifth spring term. Also, the four terms of spring data were used independently to predict scores in the fifth spring term. The 12 models used to make predictions varied in data treatment, nature of the last score, and estimate of rate of growth used. Five of the models emerged as yielding the most parsimonious set of predictions. The most parsimonious model predicted student status slightly less than 50% of the time. These results suggest that all the models are less accurate than would be liked, but this does not preclude the possibility that some of the models or their components may be useful under certain conditions. (SLD)