Value-Added Technical Webinars

These presentations are intended to provide greater detail to viewers seeking additional technical information about the value-added model. Model development, a stylized example of the model, and results interpretation are included. For statistically savvy users seeking a full mathematical description of the model, please see the Value-Added Technical Report (coming November 2011).

Section 1

This presentation is an overview of how Tulsa Public Schools worked in conjunction with VARC to create the TPS value-added model. It also includes a description of data quality requirements, the statistical model, control variables, and how both the OMAAP and OCCT were used.


Section 2

This presentation covers a stylized example of how the value-added model works. District-average growth trends are determined, district-wide demographic adjustments are determined, students at a school are compared to their predicted growth, and results are aggregated and scaled to find a value-added estimate.
Note: It is assumed that viewers have already watched "The Oak Tree Analogy" prior to this content.


Section 3

This presentation covers interpretation of value-added results. It explores point estimates, confidence intervals, basic TPS report interpretation, and district-level demographic coefficients.


Shrinkage

This presentation explains why VARC uses statistical shrinkage in its value-added estimates. Estimation error adds to the total variance of school effects, so shrinkage is applied to bring estimates back to the true variance.