Linking Teacher to Student Scores - education

School Administrator, Oct, 2000 by SCOTT LaFEE

* When judging educator effectiveness, consider the context in which teachers and administrators work.

In some circumstances, teachers and administrators have done everything feasible to improve student performance, but conditions beyond their control, such as unreasonably large class sizes or classes taught in the cafeteria, prevent maximum benefit by children. Thus, we recommend that consideration be given for variables beyond the control of the teacher or administrator.

* Use measures of student growth versus a fixed achievement standard or goal.

In the world beyond schools, few human endeavors are judged in terms of fixed goals. More typically, they are based on growth and progress toward stated goals. Even the hard-nosed world of business judges performance based on a variety of economic indicators and on comparisons to projected growth.

We propose the same paradigm be used in education with an acknowledgment of possible learning inhibitors and comparisons to projected learning growth. This approach requires the use of pre- and post-testing to determine progress, not just the attainment of predetermined pass rates or proficiency levels.

* Compare learning gains from one point in time to another for the same students, not different groups of students.

Implicit in the concept of gain scores is the assumption that similar tests will be used to measure student learning across time on an individual basis. Teachers and administrators are not gauged fairly, however, when the absolute achievement level of one class of students is compared to the absolute achievement of a different class of students. This type of comparison invites a bias in measuring gain scores that should be eliminated, not perpetuated.

Recognize that test score gains have pitfalls must be avoided.

Even when measures of student growth are used, it is critical to properly interpret gain scores. In particular, statistical artifacts known as the regression effect and ceiling effect need to be considered. These artifacts result in a tendency for students starting with low performance levels to show greater gains than warranted. Conversely, students who start with high performance may show lower gains, or even declines, if the measure of student achievement is not adequately difficult to measure what those high-scoring students know.

* Use a time frame for evaluation that allows for patterns of student learning to be documented.

If teachers and administrators are to be held accountable for student learning, patterns of student learning must be established and not be based on single snapshots of performance. Repeated measures of student learning over time enhance reliability from a statistical point of view and credibility from a decision-making perspective.

* Use fair and valid measures of student learning.

Reliability, validity, freedom from bias and fairness are obvious concerns and conditions for connecting student assessment to teacher evaluation.

* Select student assessment measures that are most closely aligned with existing curriculum.