Data Sgp – Comprehensive and trustworthy data available for measuring student growth.
This dataset comprises student assessment records from multiple years, including the most recently collected year (2014-2017). This data is then analyzed to create percentile rank tables that show how students compare with their academic peers.
Percentile rank tables are an invaluable resource for educators, parents and community members, providing quick and simple way of conveying vital information on students’ learning progress. Percentile ranks measure how closely a test score fits within a range that represents “average” performance for their grade level and content area; to calculate one for an individual student it compares their test score against an average of similar tests taken by other students who began school in similar grades.
The sgpData package features several percentile rank tables that are easily interpretable by teachers and parents, providing insight into students’ progress in ELA and math in comparison with peers within similar content areas.
In addition to percentile rank tables, sgpData includes two additional sets of summary information about student growth: school-level mean and median SGPs. The difference between means and medians indicates how volatile or stable school SGP estimates may be over time; school-level mean SGPs tend to be more stable due to being less affected by students who may lag significantly behind or ahead of their classmates.
As much as it is possible to use aggregated SGPs as an indicator of teacher effectiveness, one must take note of potential sources of variance in SGP estimates. These sources include correlations between true SGPs and student background characteristics taught, including correlations with current and prior test scores – creating potential sources of bias which cannot easily be eliminated with aggregated SGPs alone.
SGPdata’s analytic functions use longitudinal, panel and long data formats to more precisely evaluate student progress. WIDE data format is utilized by lower level functions studentGrowthPercentiles and studentGrowthProjections, while LONG format is more commonly utilized by higher level functions like summarizeSGP. The SGPdata package offers two exemplary WIDE and LONG data sets (sgpData_WIDE and sgpData_LONG), to assist users in setting up longitudinal data for analytic functions. Typically, data should be presented in LONG format if one wishes to run these analytic functions consistently from year to year. Formatting data for these higher level functions requires several data variables: VALID_CASE, CONTENT_AREA, YEAR, ID, SCALE_SCORE and GRADE. In addition, a teacher-instructor lookup table known as sgpData_INSTRUCTOR_NUMBER must also be created so as to associate each student test record with its appropriate teacher instructor.