Day: October 11, 2023

Data SGP – Calculating Student Growth Percentiles and Other Measures of Student Progress

The data sgp package, a Python library for analyzing educational assessment data, provides a number of functions that allow users to calculate growth percentiles and other measures of student progress. These calculations are useful in comparing students’ performance relative to their peers, and can help identify areas for improvement. In addition, data sgp can provide information about student achievement by subject, grade, and demographic characteristics.

Data sgp is an invaluable tool for educators who want to understand how their students are performing in each subject and grade. It can help teachers determine what growth is needed for a student to reach or maintain proficiency, and it can also help educators evaluate their own teaching skills. It can even help teachers identify specific areas that need the most improvement, and it can help them develop strategies for addressing those issues.

Student growth percentiles (SGP) are calculated for students in grades 4 through 8, and in grades 10 (because there are no MCAS assessments administered in grade 9). SGPs compare a student’s current test score to the average of the two previous tests taken in different testing windows. The SGPs are calculated using a single, unique student identifier and five test scale scores: the first column, ID, gives the student’s unique identifier; the next 5 columns, SS_2013, SS_2014, SS_2015, and SS_2016, give the student’s score for each of the previous testing windows; and the last 2 columns, TS_2014 and TS_2015, report the students’ current TS scores against their prior TS scores.

Typically, a student’s SGP is reported as an index value between 0 and 100, with higher values indicating greater relative growth. The SGPs are also reported as a decile, allowing the user to see how their students’ SGPs compare to other student groups. For example, the negative differences seen for students in the FRL row of the table below for math suggest that these students are experiencing more difficulty reaching proficiency than their non-FRL peers.

A common mistake is to interpret aggregated SGPs for a teacher or school as a measure of educator effectiveness. However, SGPs are not designed for high-stakes uses, and aggregation introduces additional biases that can distort the interpretation of individual student SGPs.

Almost all of the lower level functions that do the actual SGP calculations, such as studentGrowthPercentiles and studentGrowthProjections, require WIDE formatted data. The higher level functions (wrappers for these lower level functions) require LONG formatted data. Using long formatted data for operational analyses is simpler and more manageable, and most errors that arise during SGP analysis revert back to problems with data preparation. For more information, consult the SGP data analysis vignette.