Performance-Based Assessment

Since the late 1980s, many states, districts, and schools have developed performance assessments (PAs) to measure students' achievement of educational standards. PA comprises assessment techniques in which students must, independently or in small groups, construct responses to complex tasks that emulate realistic problems faced within an academic domain. For example, PA in mathematics might require students to estimate the amount of paint needed to cover a house by applying concepts of geometry, algebra, proportions, and other mathematical principles.

Although PA, authentic assessment, and portfolio assessment share the use of constructed responses and realistic problems, standards-based PA is unique because its content and outcomes are judged relative to specific academic standards. Educators develop and select PA tasks to assess students' mastery of specific academic standards, and judge students' performances with respect to their achievement of standards (e.g., progressing, partially proficient, proficient, advanced). Judges typically use rubrics that describe the proficiency categories and the criteria for rating student performance within each category. Also, standards-based PA typically restricts itself to on-demand assessment tasks (i.e., those that the student receives and completes in one setting), in part to ensure fairness (i.e., so that all students complete the task under the same conditions).

Because standards-based PA tasks are aligned with academic standards, PA intends to encourage teachers to improve instructional practices, educational equity, and otherwise increase their efforts toward educational reform (Khattri & colleagues, 1995). PA is also intended to challenge students to engage in higher-order thinking, to integrate and synthesize ideas, and to motivate students in ways multiple-choice tests cannot. Although some agencies use PA primarily to inform educators of a school's or system's effectiveness (e.g., reporting only group results), other agencies may use PA (usually in combination with other assessment methods) to describe individual student proficiency, make retention and promotion decisions, or even make graduation decisions.

Although PA may be popular among educators, research is equivocal in supporting its benefits. Psychometrically, PA often suffers from inadequate reliability (e.g., poor agreement among judges) and from task specificity (Shavelson & colleagues, 1999). These problems caution educators to be wary of rating students' proficiency unless judges are rigorously trained and students are assessed on a wide variety of PA tasks. PA is also expensive to develop, administer, and score, and may maintain or even exacerbate equity concerns for ethnically, culturally, and linguistically different students (Braden, 1999). Although some argue group-based PA (which requires students to work together in small groups to complete a task) is more equitable than individual PA (Neuberger, 1993), research suggests that student grouping influences individual outcomes (Webb & colleagues, 1998). Standards-based PA is most likely to benefit students and educators when it is used in conjunction with other methods of assessment (e.g., multiple-choice tests, grades, portfolios) to inform students, educators, and other stakeholders about students' mastery of educational standards.