What Is Data Sgp?

Data Sgp leverages longitudinal student assessment data to produce statistical growth plots (SGPs). SGPs are useful because they provide a quick and easy way for teachers to gauge whether or not students are making sufficient academic progress, relative to their peers. SGPs are calculated using a growth standard established via prior test scores and covariates, so they are accurate measurements of students’ learning. Unfortunately, the process of creating SGPs from students’ standardized test score histories involves complex calculations that introduce large estimation errors into these measures.

Rather than attempting to minimize these error terms, the current approach to SGP analyses tries to reduce the resulting uncertainty by comparing SGPs to those produced by an identical cohort of students in a previous year’s assessments. However, this approach is inherently recursive and requires at least three years of stable assessment data in order to produce a model that can be compared with a baseline SGP. Further, correlations between the baseline and prior year SGPs will likely be non-zero, introducing bias into the interpretation of SGP results.

In addition to allowing teachers to compare their student’s current SGP with that of their peers, SGP analysis also provides estimates of students’ projected growth in future years. This is achieved by converting each students’ raw assessment scores into scaled scores, and then comparing those scaled scores with the average of all other scaled scores in that grade and subject area for a given year. The resulting data indicate how much each student has grown above, below, or at the same level as their peers on a percentage basis.

While this methodology does produce a set of valid and reliable projections, it has significant limitations. These limitations stem from the complexity of the regression modeling techniques involved and the fact that latent achievement traits cannot be observed with any degree of accuracy. Furthermore, attempting to directly compare an SGP with a cohort-referenced model introduces substantial uncertainty into the measurement of student learning.

As a result, most SGP analyses conducted by educators are either recursive or use a baseline-referenced methodology. Despite the significant limitations of these approaches, they remain an important tool for measuring student growth and providing feedback to educators about their students’ progress.

Luckily, with the advent of online analytics and data science tools, it is now possible to automate these types of analyses. By utilizing the tools available in R, it is now possible to perform SGP analyses quickly and accurately with a relatively low amount of programming code. Having the ability to analyze SGP data in-house is vital for districts and their stakeholders in order to make informed instructional decisions. We are excited to be able to provide this capability to educators through data sgp.