Home | Publications | Just in Time

Just in Time

precision-sm_pg_01

Precision of student growth percentiles with small sample sizes

States in the Regional Educational Laboratory (REL) Central region serve a largely rural population. They face a common challenge of identifying appropriate methods for analyzing data with small samples of students. This new REL Central study investigates how precision of SGPs is related to the number of students included in the analysis, as well as how categorizing students coarsely by prior achievement before the analysis can affect the results. The study found that the margins of error in SGP estimates are larger for high and low achieving students when sample sizes are small and exact prior achievement scores are used, and is only slightly smaller when using categories of prior achievement levels. However, results from this study suggest that the sample sizes that would be expected for even the smallest REL Central states are likely sufficient for a reasonable margin of error for the highest and lowest achieving students. States should still caution stakeholders about making comparisons between SGPs that fall within the margin of error and consider carefully how to communicate to stakeholders that the margin of error in estimated SGPs differs for students with different levels of prior achievement.