Do NBCTs Still Make A Difference? Yes

In the May 9th issue of Education Week, Bess Keller reported, in a thoughtful manner, the release of information regarding another study on the effects of National Board Certified Teachers on student achievement.  The study, conducted by Bill Sanders of value-added method assessment fame, suggested that in this particular analysis, students of NBCTs did not perform any better than students of teachers who took the exam and did not pass as well as those who did not take the test as all.  Sanders’ study involved test records from two large school districts in North Carolina, Charlotte-Mecklenburg and Wake County, and drew upon data from 35,000 student records and 800 teachers.  Four other studies revealed that NBCTs did make a difference and two other recent investigations showed how teachers who achieved the standards of the National Board improved their teaching (in science) and their assessment practices.  Another study by The Finance Project, released at the 2006 meeting of the American Educational Research Association, shows that the National Board assessment process is a cost-effective form of professional development.

The recent Sanders' study was not released by the National Board for Professional Teaching Standards like it should have been.  Instead, Andy Rotherham, or the “EduWonk” as he calls himself, reported on it in his blog. Mr. Rotherham claimed that the National Board was "sitting on the results because they were not favorable.”  He may very well be right, and if he is, it is unfortunate for the National Board.

A highly respected labor economist informed me that the reason Mr. Sanders’ results were different from other more favorable studies was the way he scaled the standardized tests.  For example, Dan Goldhaber’s 2004 study of NBCTs in North Carolina – which drew upon almost 20 times more student records — found that NBCTs did make a substantial difference in student achievement.  Researchers like Sanders and Goldhaber, when using tests at different grade levels, must equate them so to arrive at comparable scores for them across time.  A researcher’s scaling decision will determine whether or not a student is seen as improving and, in this case, whether or not a teacher is seen as effective.  Unfortunately, Sanders’ methodology is proprietary and held in secret, and thus, not available for other researchers so that they can conduct the typical peer review of statistical procedures and models used (for example, see Kupermintz, H. (2004). Value-Added Assessment of Teachers.)

Additionally Sanders seems to ignore in his study that NBCTs outperform non-NBCTs on 27 out of 30 measures. Grant these differences were not statistically significant, but given the pattern it seems something positive might be suggested. And perhaps most importantly Sanders does not go to great lengths to show methodologically the differences in his control and treatment groups, other than National Board status. In any case, no one study should be used to make an “up” or “down” decision about public policy. The weight of the evidence suggests that the National Board for Professional Teaching Standards is good for the profession.  Almost 50,000 teachers are certified by a process many claim is the best professional development they have ever experienced and a few studies are revealing that when utilized effectively by administrators, NBCTs can help turn around a low performing school (abstract of Dec 05 Phi Delta Kappan article - registration required).

The National Board for Professional Teaching Standards should get its act together and fully disclose what went on with the Sanders’ study, and if there is something to be learned from it, then use the data to improve the certification process.  This nation needs to develop its teaching profession, and the National Board assessment process, with its rigorous focus on both teachers’ content and teaching knowledge, is critical.

Comments

[Ed Fuller]

The differences in the results between Goldhaber and Sanders using very similar data sets underscores the fact that value-added methodology (VAM) is dependent on the skill and assumptions of the person doing the methodology. Such findings should be a wake-up call to those who blindly tout VAM as some type of value-neutral methodology that accurately identifies the best and worst teachers among a group of teachers.