What is a Student Growth Percentile (SGP)?

Student Growth Percentiles (SGPs) provide a measure of student progress relative to previous test scores (their academic peers). While calculations for SGPs are complex, the results can easily be communicated to teachers and parents.

DPI releases School Growth Plans based on student assessment scores collected over five years with an overall scale of 0-100. ID is used as the unique student identifier while the following columns, SS_2013, SS_2014, SS_2015, SS_2016 and SS_2017 provide assessment scores from these years.

To calculate a teacher’s SGP, all individual student growth percentiles for that teacher must be averaged together and divided by the total number of students with that score – this gives us their average growth percentile and we will use this score when evaluating that teacher.

Note that DPI does not include 2014-15 school year data when producing its Student Growth Profiles as this was the only year when Badger Exam was given and thus including that year of data would present numerous difficulties when evaluating educators given that its growth calculations and results differ significantly from WKCE and Forward Exam results prior.

MDE strongly suggests that SGPs not be used for educator evaluations until 2018/19 to give time for new data to stabilize and become more reliable. Although using SGPs for educator evaluations is legal, each district will ultimately decide if this practice should occur.

SGP analyses should be fairly straightforward provided the data has been properly prepared. Any discrepancies often relate back to improper preparation of data for analysis; thus, there can be some back and forth between gathering the needed information and running actual analyses.

No matter which format you use, calculations made using studentGrowthPercentiles and studentGrowthProjections require WIDE formatted data while higher level wrapper functions like studentProfile use LONG formatted data to perform analyses. If running anything beyond simple one off analyses it is advised that data are prepared in LONG format as much of the functionality of SGP package depends on its users supplying this way.

SGPs are relatively small data sets compared to larger sets, like global Facebook interactions, making them easy to manage. Although “big data” has become the trendiest topic of discussion these days, SGPs don’t fall under that umbrella either.