Executive Perspective

Freeing Schools of Inappropriate Assessments

by DANIEL A. DOMENECH

Teachers always have assessed student perform­ance as the basis for the grades they assign. As someone who taught graduate-level test and measurement courses, I also am aware that most teachers are not well-informed regarding the science of test-item construction and the importance of test validity and reliability.

 

Daniel DomenechDaniel Domenech

Teacher-developed tests are not scientific, but they do provide feedback as to what a student has learned. Coupled with the firsthand knowledge that a teacher acquires after daily contact with a student, you have a fairly accurate assessment of student performance.

Comparability is another matter. Is the B grade in physics given by Mr. Smith comparable to the B in physics given by Ms. Jones? Is a student’s 3.5 GPA at Lincoln High comparable to a student’s 3.5 GPA at McKinley High? Once you move beyond the classroom and the school, the ability to compare student performance becomes more complex. For students wanting to go on to higher education, the colleges and universities are concerned about the comparability of the 3.5 GPA from two or more high schools. The SAT and the ACT help admissions directors answer that question. Higher education institutions use student performance on those tests to sort and rank applicants with similar GPAs.

Problematic Purpose
Beyond comparability, tests also have been designed to help teachers develop proper instructional strategies for their students. Technological advances have led to the development of online formative assessment instruments that allow for the ongoing monitoring of student performance wherein the teacher is provided with insights as to the student’s needs and prescriptive activities to meet them. Indeed, there are online tutorial programs that perform the same function without the teacher.

It is testing for the purpose of accountability that has become problematic. We have the science and technology to develop valid and reliable comprehensive assessments with which to gauge student performance. Unfortunately, creating these assessments is an expensive and time-consuming process. Instead, we have resorted to the paper-and-pencil, fill-in-the-bubble, standardized tests that are relatively cheap and quick and easy to administer. These tests provide little value to individual students because they are not diagnostic, they fail to inform instruction, and they are a more valid measure of a group’s performance than an individual’s.

The push for accountability has forced every state to develop a test to comply with the requirements of No Child Left Behind. Students are tested annually to determine whether their school is making adequate yearly progress toward fulfilling the performance standards established by the state. The goal is that by 2014 every child will be performing at the established standard.

The problem with this national goal is that there is no basis for comparability. Each state has established its own set of standards, and each state has developed its own test. There is no basis for comparing Maryland with Montana other than the percentage of students who meet the state standard, but how would that matter if the standards are not the same? It’s the same issue as comparing the B in science given by Mr. Smith with the B in science given by Ms. Jones. Even to the point that one teacher may be a more lenient grader than the other.

Some states are suspected of lowering the cut points on their tests to increase the number of students passing. This is evidenced by the significant discrepancy between state results and results on the National Assessment of Educational Progress, the closest assessment we have to a national test.

A Better Way
Since the inception of NCLB, a teaching-to-the-test culture has developed that has resulted in the diminution of the comprehensive curriculum and an overemphasis on the subjects being tested. We also have been plagued by a rash of cheating scandals that have diminished educators in the eyes of the public. Such unethical behavior cannot be condoned, but there is something terribly wrong with a testing system that hurts students and pressures adults into inappropriate behavior out of fear they will lose their job, status and reputation.

There is a better way. The move toward a set of common core standards is the first step. There can be no comparability as long as we continue to assess using 50 different sets of standards. Once there is a set of common core standards, a common core test will not be far behind. As a matter of fact, two are being developed with federal funding.

The next step would be to separate assessment for the purpose of accountability from assessment for the purpose of instruction. The NAEP can serve as the model. It is administered randomly, not to every student, no student takes the entire test, and not every subject is tested every year. Yet the NAEP is respected as one of the more valid and reliable measures of educational progress.

A NAEP-like common core test would provide comparability while significantly reducing the number of children tested, the frequency of the tests and thus the intrusion of testing on instructional time, but schools, districts and states still could be held accountable. It is time to free our schools from the bonds of oppressive assessments.

Daniel Domenech is AASA executive director. E-mail: ddomenech@aasa.org