Data SGP is the collective set of aggregated student performance measures collected over time that teachers and administrators use to understand students’ progress. It includes individual-level measures like test scores and growth percentiles as well as aggregated measures at the school and district level such as class size and graduation rates. Data SGP is critical for identifying areas for improvement, informing classroom practices, evaluating schools/districts and supporting broader research initiatives.
DESE calculates growth percentiles for ELA and math for students in grades 4 through 8 and grade 10. For these students, growth percentiles compare their current year’s MCAS test score with their MCAS test scores from the previous grade level.
A key element in the growth percentile calculations is the ability to identify students with missing data. These missing students can be represented by an NA value in the data. A number of analyses can be performed with missing data.
The higher level SGP functions (studentGrowthPercentiles and studentGrowthProjections) rely on data in the long format for all but the simplest, one-off, analyses. This is primarily because managing the long format for an operational analysis is far easier than the wide format that the lower level functions require. Additionally, the longer data format supports a streamlined process for updating analyses with each new year of data. The new data is appended to the bottom of the currently existing long data set.
For example, if you want to compare a school’s 2017 growth in math to its 2013 growth, simply select the data containing those students and click on “SGP Calculator”. The results will include not only the averages for each of the stat category columns but also the associated standard deviations. The smaller standard deviations indicate that the school-level mean SGP values vary less from year to year than do the medians that sum up these same groupings.
This same approach is used for all other content area SGPs as well as the overall SGP distribution. By calculating knots and boundaries using the entire distribution of multiple years of compiled test data the estimation errors associated with a single year are smoothed out.
It is important to note that the sgpData_INSTRUCTOR_NUMBER table is an anonymous student-instructor lookup database. This means that each student may have more than one teacher assigned to them. In these cases, the teacher’s SGP will be a weighted combination of all of the students in their classroom who have been assigned to that teacher. This approach is designed to provide more accurate comparisons of SGPs between teachers. In the future we hope to expand our efforts to make more comparisons between teachers and between schools based on data sgp. We will be in contact with the appropriate state offices to explore additional options for providing this type of access to the data. Ultimately the goal will be to migrate these data into permanent research databases. The availability of these databases will enable researchers to accession considerable legacy and unpublished meta-data and data that would otherwise not be available.