Thursday, January 27, 2011
New Study of Value-Added Modeling Reveals Problems
University of California at Berkeley researcher, Jesse Rothstein, has taken another look at a Gates Foundation study released in late 2010 that purports to show that value added scores are correlated with broader, higher level teaching and learning competencies. Rothstein's new analysis reveals that the data in the Gates research study actually shows the opposite of what their conclusions state. The fact that there is surprisingly little correlation between value added scores and broader teaching competencies measured by other tests means that teachers with low VAM scores might actually be excellent teachers when other, richer measures are used. The Rothstein correction calls into serious question the use of value added modeling in teacher evaluation, such as that used in the DCPS IMPACT system.
Posted by Real Ed Reform DC at 10:33 AM