and learning. Although any measuring instrument is bound to produce an amount of error, performance measures
have the potential for increased validity because the performance tasks are themselves demonstrations of
important learning targets rather than indirect indicators of achievement (Resnick and Resnick, 1992).
Shepard et al (1996) were of the view that performance assessment should enhance the validity of
measurement by (a) representing the complete range of desired learning outcomes, (b) preserving the complexity
of disciplinary knowledge domains and skills, (c) representing the contexts in which knowledge must ultimately
be applied, and (d) adapting the modes of assessments to enable students to show what they know.
Shepard et al (1996) further explained that the expected positive effects of performance assessment on
teaching and learning follow from their substantive validity only when assessments capture learning expectations
fully. In light of this, if teachers provide coaching and practice to improve scores, their actions will directly
improve student learning without corrupting the meaning of the indicator. Fuchs et al (1999) observed that
teachers’ use of classroom-based PA-driven instruction helped improve the problem-solving abilities of their
students.
According to Resnick and Resnick (1992), PA-driven instruction may remove pressures to teach
isolated facts and skills while offering teachers incentive to provide extended thinking activities.