Data SGP is a software package for analyzing student assessment data. Teachers and administrators can use Data SGP to gain information on how well students have performed compared to their academic peers; using this information, teachers can determine if a student has met or exceeded state growth standards set out for growth standards set by each state. Furthermore, Data SGP allows teachers and administrators to compare performance among different schools, districts or teachers using it.
Data SGP is founded on the idea that standardised test scores are imperfect measures of their latent achievement traits (Akram, Erickson & Meyer 2013), so all prior and current test score estimates contain measurement errors. When calculating SGP, the goal is to create an estimate of true latent achievement trait using covariates associated with test score – however this process can be complex and error-prone.
To reduce errors, we utilize an iterative process combining least squares modeling and Bayesian inference. The iterative process uses student test scores from their most recent and one prior tests in order to estimate latent achievement traits, then compare this estimated SGP against growth standards established based on teacher evaluation criteria and student covariates.
Iterative process repeats itself until estimated SGP falls within a given confidence interval around its growth standard. SGPs falling outside this interval are considered exceeding its requirements while those that fall within its boundaries are considered meeting them.
SGPs range from 1-99 and measure relative growth. A score of 75 indicates that a student has demonstrated growth comparable to 75% of his or her academic peers.
Although a high SGP can serve as an indicator of student success, it’s essential to recognize that teacher growth standards must be set individually. SGPs can be read both as indicators of educator effectiveness and of educator progress. SGPs at both teacher and school levels allow educators to differentiate students’ achievement outcomes and enhance classroom practice, while school level SGPs may help inform decisions regarding student groupings and educator allocation. Care must be taken when using SGPs at both the school and district levels, taking into account policy goals, intended outcomes and any possible limitations that might come into play. SGP analyses may take up significant amounts of time and are more challenging in larger systems with multiple educators or test administrations. Time spent on SGP analysis typically goes towards its preparation and execution, which can be time-consuming. The sgpData package simplifies this task with an intuitive user interface for LONG formatted data sets. Please refer to the sgpData vignette for more details on using sgpData or any wide format data sources.
