Using DATA SGP to Analyze Student Growth Data

Using DATA SGP to Analyze Student Growth Data

DATA SGP is a collection of tools to facilitate the analysis and interpretation of student growth data. It is a software package that runs on the open source R computing environment. It is recommended that users familiarize themselves with the use of R prior to running operational analyses using DATA SGP.

What is student growth percentile?

A student’s SGP is a measure of their performance relative to academic peers. It is calculated using up to two years of historical MCAS assessment results. Using a statistical procedure called quantile regression, the student’s performance is placed on a normative scale and used to identify their relative rank among students in their grade. Academic peers are defined in a variety of ways such as by demographic (e.g., gender, income) and educational programs (e.g., sheltered English immersion, special education) and are identified statewide across all grades in the same academic year.

SGP is measured on a standardized scale so that scores are comparable from one state to another. Similarly, the average statewide student growth percentile (mSGP) is also standardized so that it can be compared from year to year. Statewide median mSGP is 50. Half of the students have mSGPs above 50 and half below.

Can I compare the growth of students in my school to that of other schools?

To view aggregation SGP graphs, select a student group from the drop down list at the top of the screen. Each bubble on the graph represents a group of students and the size of the bubble indicates the number of students in that group. The location of a bubble on the graph is based on both the average mSGP and the average student growth percentile within the group.

The SGPdata package installs exemplar WIDE and LONG formatted data sets (sgpData_WIDE and sgpData_LONG) to assist with setting up your data for SGP calculations. Both formats are longitudinal and time dependent, but there are some differences in how they are stored. The sgpData_LONG set is particularly suited for generating SGP analyses because it contains the same variables as the other data sets but in an anonymized, LONG format.

The sgpData_INSTRUCTOR_NUMBER set provides an anonymized lookup table that associates teachers with their students’ test records. This can be useful for identifying teachers that are not reaching or maintaining proficient for all of their students, and is used by the teacher evaluation system to inform educator performance ratings. It is important to remember that mSGPs are based on growth relative to academic peers and are therefore sensitive to the mean achievement level of the academic peers from which it is calculated. For this reason, mSGPs should only be reported at the classroom, district and state levels for educational planning purposes. They should not be used as a part of an educator evaluation system. This is the position of the MDE and is reflected in the language of the state’s Every Student Succeeds Act (ESSA) and New Jersey Quality Single Accountability Continuum (NJQSAC) district evaluation systems.