Data SGP – How to Calculate Student Growth Percentiles (SGPs) and Percentile Growth Projections (Trajectory)

The data sgp package offers classes, functions and data to calculate student growth percentiles (SGPs) and percentile growth projections/trajectories using large scale longitudinal education assessment data. An SGP describes student progress over time as it compares against others with similar prior test scores; teachers and parents can easily understand this method of evaluation of progress over time.

SGPs allow the establishment of official achievement targets based on what students should reach in a set period of time, with this information translated into growth standards per student for all stakeholders involved in education. This differs from other methodologies which only estimate improvement needed in order to meet an intended level of performance and may not take account of what level this will eventually reach.

SGPs enable educators to make more informed decisions regarding how best to support their students’ learning by ranking students against those with similar prior test scores – providing an more equitable and relevant measure of progress than simply looking at unadjusted achievement levels. They may also help identify subgroups who may require targeted supports.

At Star Growth Reports, we use a distribution-based model to estimate each student’s SGPs conditional density estimate, then generate percentile growth projections/trajectories that can then be displayed as star reports and updated with test results as they become available. SGPs are calculated for students in grades K-12; to view your window-specific SGP, use the Timeframe drop down list on your Star Report page.

SGPs can be useful in evaluating teacher effectiveness, but their interpretation and transparency benefits must be balanced against the potential source of bias in these evaluations. One way to prevent bias would be using a value-added model that regresses teacher fixed effects, prior test scores, and student background variables into one model.

Data SGP seeks to provide an efficient means of organizing longitudinal (time dependent) student assessment data into statistical growth plots. It supports two common formats of this type of information – WIDE and LONG – where each row and column represent unique students at different times; examples for both are provided with example data sets called sgpData_WIDE and sgpData_LONG respectively.