Interpreting Data SGP

Data sgp is a collection of tables, files and functions allowing the user to conduct student growth percentile (SGP) analyses. The sgpdata package contains four exemplar data sets sgpData, sgpData_LONG, sgptData_LONG and sgpData_INSTRUCTOR_NUMBER that can be used to perform a variety of SGP analyses. The lower level functions studentGrowthPercentiles and studentGrowthProjections utilize the WIDE formatted sgpData set while the higher level wrapper function studentGrowthPlots uses the LONG formatted data sets. For operational analyses requiring the production of student reports year after year it is recommended that users use the long formatted data sets since they offer numerous preparation and storage advantages over the wide data formats.

The sgpData sets contain many helpful tables that can be used to produce SGP analyses. One key table is the sgpData_STUDENTS_PERCENTILE_TABLE which provides teachers with a handy way to compare their students’ performance to that of academic peers nationwide. The sgpData_INSTRUCTOR_NUMBER table is another useful tool for educators since it allows instructors to identify students who were taught by multiple instructors within a content area over the course of one year.

One of the most important factors to consider when interpreting a SGP is the number of prior assessments that have been taken. This is important because the SGP metric compares a student’s current assessment score to their previous assessment scores. To get the most accurate picture of a student’s growth, it is essential that their SGP is calculated using a minimum of two previous assessment scores. This ensures that the prior scores are accurately reflected in the current SGP score and also minimizes the chances of misinterpreting the results of a low or high SGP.

Another important factor to consider is the method that was used to calculate a student’s growth. SGP is determined through a statistical technique called quantile regression. Quantile regression estimates the probability that a student’s score will fall in a specific range. It is this probability that is then compared to the probabilities of each academic peer group and an adjustment is made to estimate the likelihood of the student’s score falling in a given percentage of their academic peer group. The higher the SGP, the greater a student’s growth relative to their academic peers.

To compute a student’s SGP, the prior assessment scores are scaled up or down to account for differences in student test taking ability. In addition to the scaled up or down scores, the SGP is adjusted to take into account the student’s grade level. This allows the SGP to be calculated for different groups of students and enables teachers to focus on improving the performance of specific groups. SGP is most effective when it is applied to a large and diverse group of academic peers. This is why it is important that the academic peers selected for the SGP be representative of the state’s population and that multiple testing windows are used to measure student growth.