District and state education leaders will learn a lot from today's Education Trust-Midwest report on Michigan's roll-out of its teacher evaluation system.

A key aspect of Michigan's 2011 legislation was that districts and charters could either use the evaluation framework that an expert state committee (chaired by the dean of University of Michigan's School of Education, Deborah Ball) would develop, or "opt out" and come up with one on their own, albeit one that met state guidelines.

300 school districts and charter schools have already announced their intention to opt out. From Education Trust-Midwest's intensive examination of 28 of the evaluation frameworks developed by those choosing to do so, it's clear that many are going to have to go back to the drawing board.

• Almost 50 percent don't require that tenured teachers get in-class observations.
• More than half do not lay out how to combine the use of student achievement data with in-class observation data to determine teacher ratings.
• Most troubling: None of the districts or charters in the study are using a student-growth or value-added model sound enough to determine teachers' impact on student learning. One district is actually assessing teachers based on how many of their students test "above the state average." Given that formula, what teacher in that district would want to teach kids who scored significantly below average the year before?