Data SGP is a student growth metric that tracks longitudinal test score data to provide information about a students progress in comparison to their academic peers. The metric is used to improve instruction, inform teacher evaluation systems and support research initiatives.
SGP is based on the idea that the performance of a given student in one year can be predicted by his or her performance in previous years, using a statistical model based on standardised assessment data and student covariate information. It is designed to reduce estimation error and provide more valid comparisons than conventional percentile scores.
Generally speaking, SGP analyses are straightforward; most errors encountered with the analysis process revert to issues with data preparation rather than the analysis itself. We offer a simple two step process: 1) prepare the data correctly and 2) run the analysis. Our vignette provides detailed instructions for both of these steps and the SGP function set offered in this package makes the process as easy as possible.
The sgpData dataset in this package is a panel data set with 5 years of annual, vertically scaled, assessment data in WIDE format. This exemplar data set models the format required by the lower level SGP functions studentGrowthPercentiles and studentGrowthProjections. It also serves as a template for preparing LONG format data sets which offer additional preparation, storage and retrieval benefits.
To prepare the sgpData dataset for use with the SGP function set, the following steps should be taken:
Create an XREF record for sgpData containing the standardized assessment scores of individual students. This XREF record should contain the student’s unique identifier, the grade level associated with each of the standardized assessments and the associated scale scores for each assessment. The XREF record should also include a sgpIndex for each of the assessed subjects in order to generate SGPs.
Once the sgpIndex for each subject is established, SGPs are calculated by comparing the sgpIndex of the individual to the sgpIndex for all students with the same sgpIndex. The resulting sgpPercentile for each subject shows the percentage of students who scored higher than the individual.
SGPs for a student are then compared with the corresponding sgpPercentiles for all of the other students in the student’s grade to determine how well the student performed in comparison to his or her peers. This information can then be used to identify areas of concern for the student and to inform teaching practices.
In addition to SGP calculations this function set supports the creation of growth achievement plots for each student in the data set. The plots show a student’s relative position on the sgpPercentile curve for each tested subject in the current year. Growth achievement plots can then be used to identify the students who need more help with a particular subject, and the percentage of students who met or exceeded state benchmarks for their grade level and tested subject. The data sgp vignette provides more detail on the creation of growth achievement plots.