Failed testing policy cannot be justified
The evidence that achievement testing resulted in improved test scores is, to be charitable, very weak. The most recent justification, that test results will enable teachers and administrators to identify and correct weaknesses in teaching and curriculum, is seriously flawed.
In the New York State description of the Common Core, there are about 90 standards and subheadings at each grade level from K-12. Others have identified 274 specific math skills and 116 specific language arts skills at the third-grade level and a similar number of specific skills at each subsequent grade level. How can a test with limited reliability and with 50 to 75 test items be used to evaluate mastery of 90 standards, much less 274 math skills and 116 language skills?
At best, an average of less than one test item will be used to evaluate mastery of each standard. Statistical reliability theory, to say nothing of common sense, says that a diagnostic determination based on less than one item per standard is necessarily insufficient. This latest justification is a feeble pretext to rationalize a failed testing policy.
Murray Levine, J.D., Ph.D.
Distinguished Service Professor emeritus