Data Sgp is a new project for the Wisconsin Department of Education (DPI) that provides teacher evaluation support services. It is part of the broader SEEK initiative and is intended to address some of the limitations associated with existing evaluation systems.
In general, SGPs will be more robust than existing evaluation systems, which are often based on student test scores and teacher-assigned grades. These methods are limited by the availability of student information, the time it takes to collect and organize the data, and the computational resources needed to perform analysis. In addition, they are not sensitive to classroom composition or teacher influence. SGPs are based on the historical growth trajectories of students, and they are able to project the student’s expected score in the future.
As a result, SGPs are able to identify the most likely trajectory that will lead to proficiency, and they can help teachers focus on what it will take for students to reach their projected score. SGPs also incorporate the potential impact of interventions on student achievement. This will make them more sensitive to classroom characteristics, and they will be better able to measure the effectiveness of a particular program or intervention.
Unlike traditional percentiles that compare a single student’s performance to the entire population, SGPs rank a student’s relative performance within a class of students with comparable scores and assessment histories. A student’s SGP is determined by the percentile rank of the student’s raw score on a given subject-matter test, and it represents the amount by which the student’s score increased compared to students in that same group.
SGPs will only be calculated for a student once all three test sections in a given year have been administered. These tests include one fall, winter and spring window. The assessment used in each of these windows is based on the grade level in which the student is enrolled and the subject matter for which an SGP is being calculated.
The student must have been a regular enrolled student in the class for at least 60% of the course before the state assessment. If the student was a held-back student, they must have been enrolled in a regular class for two consecutive years before an SGP can be calculated for them.
SGP calculations require a computer with the R software environment installed. This is an open source software application that can be downloaded for free for Windows, OSX and Linux. Running SGP analyses requires the use of a relational database that supports large datasets. This database is known as the sgpdata repository. A sample sgpdata file is available that models the format for data to be stored and manipulated with higher level SGP functions, which are used to calculate student growth percentageiles and projections. The sgpdata repository also includes a set of example data sets in long format, which are the preferred storage format for SGP analyses. This exemplar data set demonstrates how to manage this large, longitudinal data set and how it can be utilized by the lower level SGP functions.