Last year, we were discouraged about the findings of a federal Institute for Education Sciences study which compared comprehensive induction services modeled after (but not identical to) those of the highly respected New Teacher Center and the Educational Testing Service with the standard district induction fare. In one year of study, researchers found no differences in the outcomes most of us care about, namely teacher attrition rates and student progress.

The second year results appear to be equally discouraging. There still is no impact on teacher retention and student achievement, relative to routine induction programs, even if teachers get two full years of high quality induction. Is it time to write-off "Cadillac" induction programs? Our conclusion: not yet.

One issue that deserves a closer look is how the mentors get selected, the attention given to mentors' credentials, but not necessarily their track record in the classroom. The report notes that almost anyone who volunteered to be a mentor got the job, never a good sign. They all had master's degrees, which means nothing in terms of their quality. They all mentored previously, which also indicates nothing. There was reportedly a large share of National Board-certified mentors, but how many is unclear and the research on National Board doesn't lead us to think they were necessarily the highest performers. In sum, not enough is known about the level of performance these mentors delivered in their own classrooms.

Moreover, there is no consistency between year one and year two in terms of the nature of support provided to treatment teachers relative to control teachers.

Here's where this research leaves us: Mentor programs that offer uneven support provided by full-time mentors who have good credentials don't seem to improve student achievement or teacher retention relative to run-of-the-mill induction programs offering uneven support by mentors of uncertain qualifications who are not working full time at it. Got it?