Data SGP is an open-source package used for statistical computing in R’s environment. Compatible with Windows, OSX and Linux operating systems, in order to utilize its statistical computing features it must first have R installed on one’s computer by visiting CRAN website and downloading their installation installer corresponding to your operating system. There are various resources available online which will assist beginners getting acquainted with R.
The data sgp package allows users to calculate student growth percentiles and projections. To run these analyses, users need access to certain data: VALID_CASE, CONTENT_AREA, YEAR, ID, SCALE_SCORE GRADE ACHIEVEMENT_LEVEL and ACHIEVEMENT_LEVEL (only required if running student growth projections). In an ideal world, this data should conform with what SGP uses as its format; additionally RID must have an unique value set on it so as not to throw off its calculations.
Students can be organized based on achievement level in many ways, with results being compared with state or district averages for comparison purposes. This data can help schools and districts target efforts at improving the performance of low performing students more effectively.
To grasp how this method of performance comparison works, let’s use an example involving low scoring students compared to their academic peers. A student’s SGP score would then be determined by dividing his or her SGP score by all academic peers with similar backgrounds whose SGP scores have also been calculated; the resultant number indicates how far below expected average this student is performing on his or her achievement level.
The Progress Measure allows educators to accurately track student advancement. Even students who achieve lower-than-average scores on tests can demonstrate significant increases in achievement over time. Improved academic outcomes are particularly essential for struggling students and can serve as an incentive to work hard in school. Furthermore, high achieving schools may find this a stimulating challenge that drives them forward towards continued excellence. The two most frequently employed methods for calculating student growth percentiles are standard normal distribution and bell curve methods. While standard normal distributions offer more mathematical accuracy, bell curves tend to be easier for non-mathematicians to comprehend and understand. Both approaches produce similar results; therefore it is essential to select an approach which ensures representative data of a population being studied as well as accurate and valid analysis. The optimal distribution will depend on the specific circumstances and goals of a research project, but in general should aim for one which is statistically robust so as to reduce overestimation or underestimation of effects from treatments.