Data SGP (Statistic Growth Plots) analyze longitudinal student assessment data to generate statistical growth curves which provide visual evidence of a student’s performance relative to their academic peers. This information is useful for assessing individual students, informing teacher/student instruction, supporting classroom research initiatives and evaluating schools/districts.

The SGP package requires a computer running the R software environment. This is a free, open source application that can be run on Windows, OSX and Linux. We recommend that you take some time to become familiar with R before diving into running SGP analyses. There are many resources on CRAN to assist with getting started.

Depending on the scope of an analysis it may be necessary to import additional packages and functions into your program. This is not uncommon for analyses involving complex statistical models and can be especially true with the SGP package. We have created a help document to walk users through the process of importing the necessary packages and functions.

Data SGP analyses are complex and require significant computing resources. We have designed our system to be as user friendly as possible given the level of complexity involved. The bulk of the time required for a SGP analysis is spent preparing the data for analysis. Once the data has been prepared correctly, the analysis is a relatively quick and simple process.

To get started with a SGP analysis we have provided a set of example WIDE format data sets and higher level wrapper functions like studentGrowthPercentiles and studentGrowthProjections that facilitate the preparation of longitudinal student assessment data for SGP analyses. We also have a LONG format data set that is particularly well suited to SGP analyses and provides advantages over the use of WIDE data when creating SGP outputs.

The sgpData_LONG data set contains student assessment data for 8 windows (3 windows annually) over 3 content areas. This dataset has been anonymized and has 7 required variables: VALID_CASE, CONTENT_AREA, YEAR, ID, SCALE_SCORE, GRADE and ACHIEVEMENT_LEVEL (required if running student growth projections). This dataset is best used with the SGP functions that produce plots for each student in an individual school/district.

SGP is a unique methodology that differs from standard growth models in its ability to directly compare student/teacher test score progression with official state achievement targets/goals. This feature distinguishes SGP from other methods and serves to motivate teachers by linking their performance against measurable goals. This is an objective that cannot be achieved via standard growth models and which Michigan requires of its educators as part of its educator evaluation system. It is a critical component of SGP that we continue to work to expand and improve.