Data SGP is a collection of tools and resources used by districts and schools to analyze student growth percentiles and their implications. These include:

Student Growth Percentiles (SGPs) provide a relative performance measurement that compares current test scores of academic peers from prior years’ assessments in the same subject area. Teachers and schools use SGPs as an effective way of identifying students who are outperforming or underperforming their academic peers in school assessments and can focus on helping these pupils with appropriate interventions that support their learning in classroom settings.

An SGP may be calculated using one year or multiple years’ of assessment data, with trend analyses being performed on that latter grouping to allow users to compare student performance over time with that of peers and identify patterns which could signal intervention needs.

An SGP model typically compares students based on their prior test scores and grade levels to identify comparable students with similar score histories, in order to assess growth potential. Therefore, using similar grades for all participants is vitally important when using multiple prior test scores as benchmarks for growth comparisons. Furthermore, most models include high performing students being compared to peers with similar high scores so as to detect maximum potential gains.

Use of SGP models can be challenging and requires an understanding of the mathematics involved with comparing scaled scores over time. Therefore, it is vitally important that educators receive proper training on these tools and their implications for educators – an excellent place to begin would be OSPI’s SGP vignette and videos as starting points.

SGP analysis can be conducted in various ways with numerous variations to its source code. Generally speaking, lower level functions like studentGrowthPercentiles and studentGrowthProjections require WIDE formatted data whereas wrappers for these lower level functions often assume LONG formatted data is available; to make operational SGP analyses simpler and reduce code complexity more effectively we suggest utilizing the sgpData long formatted data set as part of this analysis process.

The sgpData long formatted data contains five years worth of MCAS scores for every student in every subject area over five years, separated into individual ID fields for easy retrieval. The next five columns (SS_2013, SS_2014, SS_2015 and SS_2016) provide each student’s scaled MCAS assessments scores from previous exams administered through MCAS tests administered between 2013-2016 in each subject area. Furthermore, the embedded function “sgpData_stateData” within sgpData also stores state specific meta-data. Please refer to the sgpData documentation for more details on how to utilize this data for SGP calculations. When properly prepared, SGP analyses tend to be straightforward with any errors usually stemming from poor data preparation practices. If any questions about using sgpData arise, don’t hesitate to reach out – our SGP team would be happy to assist!