for the measurement of student achievement. It suggests, for example, that recall of factual information is inadequate as a measure of deep understanding or as an indicator of the ability to transfer learning to new situations or problems.
Conventional psychological and educational testing is an outgrowth of theories of ability and intelligence that were current at the beginning of the century. Psychometrics has become increasingly sophisticated in its measurements, yet it does not attempt to look inside the “black box” of the mind. Now that the newer sciences of cognition and development have transformed our understanding of learning and the development of expertise, measurement theory and practice need fundamental rethinking. There is much in the traditional methods that is valuable, including a focus on objectivity and reliability of measurement. There is a problem, however, with what is being measured.
As a first step in the process of rethinking educational testing, the committee recommends that assessment tools be designed and tested with the goal of measuring deep understanding, as well as the acquisition of factual knowledge. This is both a modest beginning and a challenging task. To be useful for policy purposes, these assessments should be in a form that can be administered widely and scored objectively and that meets reasonable standards of validity and reliability. These requirements can be at odds with the measurement of deep understanding, at least in the current state of the art. But it is important to begin finding solutions that, for example, minimize the trade-off between assessing for understanding and scoring objectively. A variety of experiments is needed, both with new forms of standardized tests (including computer-based instruments that permit “virtual” experiments), and with alternative assessments (such as portfolios) that have become more popular in recent years.
Research on assessment tools of different types is recommended to determine:
Whether alternative assessments yield significantly different measures of student achievement or highly correlated results.
How alternative assessment measures might be combined to offer a balanced view of achievement.
19. Review teacher certification and recertification requirements.
Currently, 42 of the 50 states assess teachers as part of the certification and licensure process. But states vary enormously in the criteria used and the amount and type of assessment they require. The federal government also has provided support for an assessment process for advanced certification that is developed and administered by the National Board for Professional Teaching Standards. It is recommended that research be conducted to re-