Students’ performance on state assessments and the subsequent data they generate provide educators with invaluable information about student growth. This information is used to identify areas of concern, inform classroom practices, support research initiatives and evaluate schools/districts. This information is also used to develop measurable goals for teachers.

Data sgp is an analytical tool for longitudinal student assessment data that creates statistical growth percentiles (SGPs) and percentageile growth projections/trajectories using students’ standardized test scores with covariate information from their previous testing history. SGPs are more accurate measures of growth than traditional percentile scores and offer a clearer picture of students’ progress relative to their academic peers.

Unlike standard assessment metrics such as mean, median and mode, SGPs are reported on a 100-point scale where higher numbers indicate greater relative growth than lower ones. This enables educators and parents to easily determine whether a student’s test score has grown more, less or about as much as their academic peers – this is particularly useful for identifying underperforming students who require additional support while differentiating instruction for high-performing students and evaluating the effectiveness of existing teaching methods.

SGP also stands out from standard growth models and other methodologies in that it allows for comparison of student/teacher performance against official state achievement targets/goals. This provides districts with an ideal means of communicating to stakeholders that proficiency must be reached within a defined timeframe while serving as a motivating factor for teachers by linking their performance against measurable goals. SGP is also the only method currently available to accomplish this and serves as a critical component of Michigan’s educator evaluation system.

While SGP analysis is powerful, it does require a significant investment of time to prepare and run analyses. OSPI staff are ready to offer training and help throughout this process.

SGP analyses are most effective when performed on longitudinal student assessment data – this requires a minimum of five years of testing to accurately represent a students’ performance and progression. The SGP Package offers a data set called sgpData that simplifies this work by providing a pre-formatted longitudinal student assessment dataset for use with SGP functions. This dataset includes the identifier, year, content area, test date and scale score for each of the five most recent testing windows and a unique index to link these data points to a student’s longitudinal record.

This dataset can be downloaded from the SGP GitHub repository here. In addition, the SGP Package includes a WIDE format data set that simulates time dependent data used with functions like studentGrowthPercentiles and studentGrowthProjections as well as a LONG format data set to assist in converting existing assessment data into the proper SGPdata format. The sgpData vignette provides more detailed instructions on using this data set.