Data sgp is an analytical tool that allows users to examine the relative performance of students enrolled in the same course. It provides important information for teachers and administrators as they work with their students to improve student achievement, ensuring that each one is on the path to success. It also provides valuable information for students and parents as they evaluate the effectiveness of their schools and teachers.
As with all analytics, the bulk of time in conducting SGP analyses is spent on data preparation. Most errors that come up in analysis generally revert back to issues with data preparation and thus most of our support time is spent helping users get their data into the proper format for running SGP analyses. Once the data is prepared properly, the analyses themselves are relatively simple and straightforward.
SGP analyses use up to two years of historical MCAS data in order to create a percentile rank for each student. This ranking is determined by comparing the student’s performance with that of academic peers who scored similarly on the two years of state assessments. The rankings are derived using a statistical process called quantile regression.
In addition to the student rankings, the SGP data set contains a number of other useful information. For example, the variable sgpData_INSTRUCTOR_NUMBER is an anonymized lookup table that indicates which teacher is associated with each student’s test record. This information is useful in interpreting the meaning of the teacher practice score and SGO score.
The SGP data set also includes a number of windows in which to compare student growth. This is particularly useful when a district wants to compare student growth for a specific window of time. This is accomplished by selecting a prior or current school year when customizing a report (using the Timeframe drop-down list). It is also used to produce student growth projections, which are displayed on Star Growth Reports.
Several functions in the sgpdata package use SGP data in both WIDE and LONG formatted sets. The lower level functions that do the calculations, such as studentGrowthPercentiles and studentGrowthProjections, require WIDE formatted data while the higher level functions (wrappers for the lower level functions) utilize the LONG formatted sgpdata meta-data.
It is also possible to run a more detailed analysis of the students within each school by creating individual student aggregates. This can be done by combining the student rankings and teacher aggregates with the additional demographic/student categorization variables. These variables can be found in the sgpData_LONG data set. These include VALID_CASE, YEAR, ID, GENDER, CONTENT_AREA, SCALE_SCORE, GRADE and ACHIEVEMENT_LEVEL (required for running student growth projections). The remaining variables are optional but useful. See the sgpdata vignette for more detail on how to create and analyze this data set.