A recent report from
the Wisconsin Institute for Law and Liberty attempts to compare student
test score performance for the 2015-16 school year across Wisconsin’s
public schools, charter schools, and private schools participating in
one of the state’s voucher programs. Though it highlights important
patterns in student test score performance, the report’s limited
analyses fail to provide answers as to the relative effectiveness of
school choice policies.
Apples to Apples: The Definitive Look at School Test Scores in Milwaukee and Wisconsin was reviewed by Benjamin Shear of the University of Colorado Boulder.
Comparing a single year’s test scores across school sectors that serve different student populations is inherently problematic. One fundamental problem of isolating variations in scores that might be attributed to school differences is that the analyses must adequately control for dissimilar student characteristics among those enrolled in the different schools.
The report uses linear regression models that use school-level characteristics to attempt to adjust for these differences and make what the authors claim are “apples to apples” comparisons. Based on these analyses, the report concludes that choice and charter schools in Wisconsin are more effective than traditional public schools.
Unfortunately, the limited nature of available data undermines any such causal conclusions. The inadequate and small number of school-level variables included in the regression models are not able to control for important confounding variables, most notably prior student achievement. Further, the use of aggregate percent-proficient metrics masks variation in performance across grade levels and makes the results sensitive to the (arbitrary) location of the proficiency cut scores.
The report’s description of methods and results also includes some troubling inconsistencies. For example the report attempts to use a methodology known as “fixed effects” to analyze test score data in districts outside Milwaukee, but such a methodology is not possible with the data described in the report.
Thus, concludes Professor Shear, while the report does present important descriptive statistics about test score performance in Wisconsin, it wrongly claims to provide answers for those interested in determining which schools or school choice policies in Wisconsin are most effective.
Apples to Apples: The Definitive Look at School Test Scores in Milwaukee and Wisconsin was reviewed by Benjamin Shear of the University of Colorado Boulder.
Comparing a single year’s test scores across school sectors that serve different student populations is inherently problematic. One fundamental problem of isolating variations in scores that might be attributed to school differences is that the analyses must adequately control for dissimilar student characteristics among those enrolled in the different schools.
The report uses linear regression models that use school-level characteristics to attempt to adjust for these differences and make what the authors claim are “apples to apples” comparisons. Based on these analyses, the report concludes that choice and charter schools in Wisconsin are more effective than traditional public schools.
Unfortunately, the limited nature of available data undermines any such causal conclusions. The inadequate and small number of school-level variables included in the regression models are not able to control for important confounding variables, most notably prior student achievement. Further, the use of aggregate percent-proficient metrics masks variation in performance across grade levels and makes the results sensitive to the (arbitrary) location of the proficiency cut scores.
The report’s description of methods and results also includes some troubling inconsistencies. For example the report attempts to use a methodology known as “fixed effects” to analyze test score data in districts outside Milwaukee, but such a methodology is not possible with the data described in the report.
Thus, concludes Professor Shear, while the report does present important descriptive statistics about test score performance in Wisconsin, it wrongly claims to provide answers for those interested in determining which schools or school choice policies in Wisconsin are most effective.
No comments:
Post a Comment