Student Growth Percentiles (SGPs) are increasingly being used in evaluations of teacher effectiveness. This study investigates two properties of SGPs: 1) SGP sensitivity to reference group characteristics such as sample size, free and reduced lunch (FRL) status, and English language learner (ELL) status; and 2) variation in score changes across students with the same SGP. Math and reading score changes from grade 4 to grade 5 were considered. Results indicate that SGPs are sensitive to sample size and FRL and ELL status. Similarly, significant variation exists in average score change for students with the same calculated SGP, particularly if the SGP is near the extremes. The implications these findings have on teacher accountability are discussed.