The data sgp website is an online resource for educators analyzing student growth percentiles (SGP) and student growth projections. It provides an easy-to-use, customizable set of tools for interpreting and reporting student growth. It is available to all educators and districts in Michigan.
SGP is a metric that measures how much a student has improved relative to his or her academic peers in comparison to a prior assessment. SGPs are based on a statistical technique called quantile regression that compares students’ current test scores to their previous test scores using the same assessment subject and grade level. SGPs allow teachers and administrators to compare a student’s progress to that of their academic peers across the state. SGPs can be viewed by individual student, group, school and district and are intended to provide insight into a student’s achievement path.
Unlike traditional standardized test results that are reported in percentages, SGPs are reported as growth percentiles ranging from 1 to 99 with higher numbers indicating greater relative improvement. The SGP distribution is not expected to be a bell curve as seen in the chart below, instead it should have a number of students at each growth percentile or decile. For example, the middle 50% of SGPs is expected to have approximately 10 students at each point on the curve.
There are two types of SGPs: Window Specific and Current. Window Specific SGPs are a snapshot of a student’s SGP at a given time period and are available in Star Growth Reports when the report customization is set to “Winter”, “Spring”, or “Summer”. The Current SGP is the most recent SGP for a student based on the last assessment that has been completed.
For SGP analyses to work properly it is important that the data be prepared correctly. The majority of errors that occur when analyzing SGPs revert back to problems with the initial data preparation. Having the correct format and structure for the data is the key to making accurate SGP calculations. The data sgp package, installed with the Star Toolkit, contains four exemplar WIDE and LONG formatted data sets to assist in this process: sgpData, sgpData_LONG, sgptData_LONG, and sgpData_INSTRUCTOR_NUMBER. The lower level functions studentGrowthPercentiles and studentGrowthProjections require WIDE formatted data while the higher level functions studentGrowthTrends and studentGrowthTrendSummaries require LONG formatted data.
The data sgp package is written in the open source software environment R which is available for Windows, OSX and Linux. Having a basic understanding of how to use R is helpful when working with the SGP data package. The data sgp package has an extensive help manual that covers the various aspects of the SGP analysis process. In addition to the manual, there are a number of online resources that can be utilized to help guide the process of creating and running SGP analyses. Please feel free to reach out to us with any questions regarding the data sgp package. We are happy to help!