Over the past few years, there has been a general dissatisfaction over the results of traditional standardized objective tests. Concerted efforts have, therefore, been expended to find alternative assessment mechanisms of measuring educational outcomes and processes and measure more complex processes in education. For example, multiple-choice tests have been criticized because they, purportedly, are unable to measure complex problem-solving skills, are hopeless in measuring processes that occur in daily classroom activities, gauge the processes involved in accomplishing the task performance and examine learners’ application skills rather than superficial learning of the material. Educators have therefore focused their attention to finding alternative assessment methods that would hopefully address these difficulties with the traditional methods of objective assessment.
Performance-based assessment is one alternative assessment technique that has been proposed. Performance-based assessment procedures believe that the best way to gauge a student or pupil competency in a certain task is through observation en situ or on site. Such a belief appears consistent with the constructivist philosophy in education often taught in courses on Philosophy of Education. A performance-based test is designed to assess students on what they know, what they are able to do and the learning strategies they employ in the process of demonstrating it.
Many people have noted serious limitations of performance-based tests and their vulnerability toward subjectivity in scoring and creating or providing the real or closer-to-the task environment for assessment purpose. However, the concerns for subjectivity may be addressed simply by automating the test. The second issue is obviously a bigger problem, and there is no guarantee that ideas from one domain will apply to another.
There are many testing procedures that are classified as performance tests with a generally agreed-upon definition that these tests are assessment procedures that require students to perform a certain task or activity or perhaps, solve complex problems. For example, Bryant suggested assessing portfolios of a student’s work over time, students’ demonstrations, hands-on execution of experiments by students, and a student’s work in simulated environments. Such an approach falls under the category of portfolio assessment (i.e. keeping records of all tasks successfully and skillfully performed by a student). According to Mehrens performance testing is not new. In fact, various types of performance-based tests were used even before the introduction of multiple-choice testing. For instance, the following are considered performance testing procedures: performance tasks, rubrics scoring guides and exemplars of performance.
In performance tasks, students are required to draw on the knowledge and skills they possess and to reflect upon them for use in the particular task at hand. Not only are the students expected to obtain knowledge from a specific subject or subject matter but they are in fact required to draw knowledge and skills from other disciplines in order to fully realize the key ideas needed in doing the task. Normally, the tasks require students to work on projects that yield a definite output or product, or perhaps, following a process which tests their approach to solving a problem. In many instances, the tasks require a combination of the two approaches. Of course, the essential idea in performance tasks is that students or pupils learn optimally by actually doing (Learning by Doing) the task which is a constructivist philosophy.
As in any other test, the tasks need to be consistent with the intended outcomes of the curriculum and the objectives of instruction; and must require students to manifest (a) what they know and (b) the process by which they came to know it. In addition, performance-based tests require that tasks involve examining the processes as well as the products of student learning.