Whether you’re a student, parent, or teacher, knowing about data sgp is essential to understanding how to assess student progress. But with so many methods and metrics available, it can be difficult to keep up with all the different terminology and jargon. This article will help clarify some of the main concepts about student growth percentiles (SGP) and provide a simple overview of how they are calculated.
A Student Growth Percentile (SGP) is the percentile rank of a student’s current achievement relative to students matched to them based on their prior achievement. For example, a student with the 70th percentile rank in math would have an SGP score of 70. Two features of this definition make SGPs popular: the percentile rank scale is familiar and interpretable, and the percentage ranks remain well-defined even when test scores are not vertically or intervally scaled.
When a student takes a test, their raw score is compared to the average raw score of all students who took that same assessment. The raw score is converted to a percentile rank, and the student’s SGP is determined by their current percentile rank and the number of years that they have been assessed. SGPs are most useful for evaluating students who have been in the same classroom and the same subject area.
The relationship between SGPs and student covariates can be explored by creating a scatterplot of the student’s SGP and their math and ELA scores. While the scatterplot is not perfect, it shows that the covariates explain a small amount of the variance in both the math and ELA scores. This is important to remember when interpreting the results of an SGP because it means that it’s not just the SGP score itself that is explaining the variation in student performance; it’s also the relationship between the student and their previous performances.
As a result, it is generally not recommended to use SGPs without considering the covariates. The scatterplot also illustrates that conditional SGP estimates based on the math scores only (curve with triangles), the math and ELA scores plus the covariates (curve with +), and the math and ELA scores plus both the covariates and the math scores, have similar magnitudes of error. This is consistent with McCaffrey’s (2015) finding that conditioning on additional information does not improve the accuracy of SGP estimations.
Lastly, the choice to format in WIDE or LONG formats has implications for what analyses can be performed. The lower level SGP functions studentGrowthPercentiles and studentGrowthProjections utilize WIDE format while the higher level wrapper functions use LONG. For most analyses, LONG format is recommended as it provides numerous preparation and storage benefits over WIDE. However, for operationally-oriented analyses, it is likely that you’ll be primarily using the studentGrowthPercentiles and studentsGrowthProjections functions. If you plan to run the same analyses operationally year after year, then the WIDE format may be more appropriate for your needs.