The data sgp is a powerful tool that can be used by teachers to improve their students’ learning experiences. It can help them identify their students’ strengths and weaknesses and make changes that will increase their chances of success in the future. It can also be used to evaluate the effectiveness of their teaching methods.
The classes, functions and data within the sgp package are used to calculate student growth percentiles and growth projections/trajectories using large scale, longitudinal education assessment data. This is accomplished by performing quantile regression on the student’s assessment history to estimate their conditional density and then projecting/estimating future achievement using these derived coefficient matrices. SGP analyses provide more accurate estimates of student achievement than VAMs and could serve as a viable alternative to current accountability systems that focus on test score measures.
SGPs are more appropriate for systems that emphasize a growth-oriented approach to accountability. However, they are still not ready for widespread use given their complexity and lack of fit with current accountability systems that emphasize proficiency targets. This is an area where we are actively working to improve and extend the SGPs.
To use the SGP package, you will need access to a long dataset of state-level education assessment data (e.g. sgptData_LONG). In addition to the raw scores, this should contain the percentiles calculated by the SGP package as well as state level student aggregates based on the meta-data contained in sgpstateData.
Using wide format data like sgpData with the SGP package is, in general, straight forward. The lower level functions such as studentGrowthPercentiles and studentGrowthProjections utilize the WIDE data format but all of the higher wrapper functions are designed to work with LONG format data sets. In fact, if you plan on running your SGP analyses operationally year after year, you will likely want to format in the LONG data format which provides numerous preparation and storage benefits over WIDE.
In sgpData, the first column, ID, provides the unique student identifier while the next 5 columns, GRADE_2013, GRADE_2014, GRADE_2015, GRADE_2016, and GRADE_2017, provide the student’s grade level assessment score in each of these years. If a student does not have 5 years of test data, the value is missing (NA).
To prepare sgpData for SGP calculations, it must be sorted by teacher. To do this, the data set contains an anonymized, student-teacher lookup table (sgpData_INSTRUCTOR_NUMBER) that provides the instructor for each of a student’s test records. This information is then used to synchronize teacher assignments for all future analysis. Using this synchronized data will ensure that all student growth analyses are performed under the same conditions, regardless of which instructors were involved in the testing. This will ensure consistency in the interpretation of student growth trends and the accuracy of future projections/projections. Alternatively, you can also synchronize the data by teacher. However, this will not provide the same degree of consistency because each teacher may have different teaching styles. This will result in different student growth trends and projections/projections.