In these days of Twitter, Facebook, Tumblr and other services that provide daily, hourly and even minute-by-minute updates, I get lots of information about reports, studies and education events. This week , on my Twitter and Facebook accounts, I highlighted two of the more interesting reports.
* a terrific new report on long-term benefits of early childhood education: http://obs.rc.fas.harvard.edu/chetty/STAR_slides.pdf
* an interesting report from the U.S. Department of Education’s Institute of Education Sciences (IES), which reinforces Kentucky’s caution on using value-added criteria for high-stakes decisions: http://ies.ed.gov/ncee/pubs/20104004/
Readers of these reports may find conflicting information. The early childhood report was mentioned in a New York Times article entitled “The Case for $320,000 Kindergarten Teachers.” The report came from Harvard economist Raj Chetty. Many critics of government spending on early childhood education allege that the early gains for children often wash out over the course of the school years. I really liked a quote attributed to Mr. Chetty – “We don’t really care about test scores. We care about adult outcomes.”
Mr. Chetty and his team reviewed the adult outcomes for students who were part of the Tennessee Star Project from the ‘80s. They found students who were in the most effective kindergarten classrooms were more likely as adults to go to college, less likely to be single parents, more likely to be saving for retirement and more likely to earn more than their peers in the same study who were in less effective classrooms. So, the article and study seem to support the big idea of teacher effectiveness having a major role in student learning and adult outcomes.
In a somewhat conflicting report from the National Center for Education and Evaluation, researchers raised many cautions about using teacher effectiveness data for high-stakes decisions like merit pay and tenure. The report, Error Rates in Measuring Teacher and School Performance Based on Test Score Gains, highlighted concerns about error rates when using this data. The report states that three years of teacher data that classifies teachers as low performing could be erroneous as much as one out of four times. The report shows that data classifying a teacher as high-performing also could be erroneous as much as one out of four times. With five years of data, the results become a little more stable, and if the data are used mainly for school classification, then the error rates decline. Regardless, this research does raise much concern about the use of one, two or even three years of data to make high-stakes decisions.
So, the question becomes, “how do we resolve the apparent conflict in these two reports with regard to outcomes and the impact of teachers on the learning outcomes?” The answer is in multiple measures. I have been clear with policy makers, superintendents, teachers and other stakeholders. Kentucky will not depend solely on test scores for high-stakes decisions for teachers and principals. Both of these studies point out concerns with just using test scores. In Kentucky, we will continue to work with teachers, principals and other stakeholders to develop evaluation systems that utilize multiple measures of effectiveness.