The data sgp package offers tools for analyzing longitudinal student assessment data to produce statistical growth percentiles (SGP). SGP analyses are often employed as an alternative to Value-added Models when assessing teacher effectiveness; SGPs rank students with similar starting academic performance relative to each other based on the rate of achievement progress over time, and can allow comparisons of student achievements between state standards and objectives while accounting for differences in the difficulty of assessments.

Users wishing to conduct SGP analyses require longitudinal student assessment data and suitable software or hardware for calculations. Most of the time spent conducting an SGP analysis will likely be spent preparing data; once complete, running analyses itself becomes relatively straightforward.

Starting off SGP analyses can be daunting, so this package offers four sample data sets to get them going. The first, sgpData, specifies data in WIDE format required by lower level functions like studentGrowthPercentiles and studentGrowthProjections; two other data sets (sgpData_LONG and sgptData_LONG) meet this need as do higher-level SGP functions such as abcSGP, prepareSGP and analyzeSGP while the fourth data set (sgpData_INSTRUCTOR_NUMBER) acts like an anonymized teacher-student lookup table used to produce teacher level aggregates.

Data in the sgpData package is organized into eight windows (3 windows annually) of student assessment data in three content areas (Early Literacy, Mathematics and Reading). Each window of student data includes unique student identifiers, scale scores, grade levels and achievement levels for every student in that window of data. Furthermore, demographic/student categorization variables used for creating teacher aggregates are included within sgpData_LONG for use when producing aggregate teacher reports.

Every year, sgpData_LONG is processed through the studentGrowthPercentiles function to generate student growth percentiles at each grade level and content area. These percentiles are then aggregated to produce a mean student growth percentage for every school/district/subgroup.

As seen on the right, the mean SGP for fifth grade students is 57; this indicates that most have achieved grades above 55 statewide average; however, some may have SGPs significantly below 55, signaling less academic progress compared to others in their class.

SGPs provide more understandable results than VAMs as they compare students of similar starting academic performance based on how much progress each one has made rather than just an absolute level of performance. SGPs can make the SGP process easier to comprehend for teachers, administrators, and parents. SGPs can be used to communicate how far a student needs to improve in order to achieve target achievement and help develop plans to do so. SGPs can be particularly helpful for schools and districts seeking the attention of state and local leadership when striving to meet ambitious educational goals. SGPs can effectively communicate the need for high levels of student achievement as part of a comprehensive plan to reduce achievement gaps.