As the other articles in this issue make clear, professional practice regarding the assessment of teacher performance is in a state of transition. In the past, most teacher evaluation practices had been held in low regard. Researchers concluded that, generally speaking, traditional evaluation practices were simplistic, purposeless, and overIy subjective (Cruickshank & Haefele, 1990). A number of significant current initiatives, such as the projects of the National Board for Professional Teaching Standards, the Stanford Teacher Assessment Program, and the Connecticut Department of Education, are trying to overcome this legacy by identifying characteristics of effective teaching and assessing them through sophisticated performance-based techniques such as semistructured interviews, portfolios, "inbasket" activities, and videotaped demonstration lessons.
Perhaps not surprisingly, the state of the federal law relevant to teacher performance assessment is currently also in a state of transition. Although education is primarily a state responsibility within the American federal system, the major class action litigations that have had the greatest impact on psychometric practices have arisen in the federal courts. State statutes and case law have provided a modest degree of oversight of teacher evaluation standards (Rebell, 1990), but most of the legal standards that have influenced psychometric thinking have resulted from the enforcement of Title VII of the 1964 federal Civil Rights Act in cases involving claims of employment discrimination against African-Americans, IIispanics, or women.
Title VII class actions traditionally focused almost exchrsively on "objective," standardized employment selection practices such as multiple-choice, paper-andpencil tests or diploma requirements. Only a relatively few cases involved performance assessment issues. Typically these involved supervisory judgments that were impressionistic or based on rating systems that involved no explicit assessment criteria or rater training (see, for example, Rowe v. General Motors Corp., 1972). The courts invalidated these practices with littIe hesitation and with little discussion. Judicial skepticism was particularly aroused in these cases when no minorities served as evaluators in the process (see, for example, Fisher v. Proctor & Gamble Manufacturing Co., 1980).