Toward Intelligent Systems for Testing
This chapter explores one way the technologies of testing might combine with certain cognitive science techniques to provide an adaptive form of instruction. The form of diagnostic testing, used often, in small amounts, provides continuous assessment and instruction as opposed to relatively standard pretest–treatment–posttest designs for individualizing training and instruction. Domain expertise is always embodied in instructional decision making, either explicitly or implicitly. Curriculum knowledge is the specification of the goal structure that guides the teaching of a body of expertise. The planning knowledge includes two sets of planning concerns: the arithmetic difficulty of problems that are presents to the student and the circuit complexity. The basic approach is to begin each cycle of diagnosis by sweeping through the curricular goal structure, noting which subskills are ripe for testing. The processing capacity demands of one subskill may be so great as to make impossible the execution of another newly formed subskill.