Share this article

print logo


Those of us in the profession of measuring and assessing health-care outcomes can only grimace when we read stories of public school teachers being judged on the basis of a single statistic -- their students' Regents scores.

Equating student passing rates and class averages directly with teacher performance constitutes bad statistics. But this can be easily countered.

Class level outcomes, for example, average test scores for Mrs. B's 1999 earth science Regents classes at School C, will likely differ from Mrs. K's classes in the same school. They are also likely to differ from Mrs. B's previous classes in 1998, and from Mrs. Z's 1999 class at School H.

Why? Because no two classes of earth science students are apt to have the exact makeup of all the important factors that contribute to successfully passing the exam.

Reading comprehension, which is likely to determine a student's ability to understand a test question, as well as learning capability and style, personal motivation, self-efficacy, family support, etc., are not identical between students, classes of students and schools.

There is an expectation for each class performance on a standardized test and it is not necessarily 100. It varies with every class. In the year 2000, Mrs. B might have a class with a higher proportion of students lacking a key inherent or acquired ingredient (we call them predictors) necessary to achieve a passing grade on the Regents, much less a high grade.

One should be able early on in the school year to assess a probable class average Regents score for Mrs. B with some confidence interval. That should become her target to meet or exceed -- not last year's score or Mrs. K's target.

Each class' outcomes need to be compared with expectations for that unique group of students. Expectations require sophistication to set. Failure to set them appropriately defies sound science, ignores reality, creates undo pressure and invites the very type of abuse sadly depicted in The News article "Teachers under pressure."

Accountability is good when done right and tends to serve everyone. When done inappropriately, it serves no one.


There are no comments - be the first to comment