On Meaningful Measurement: Issues of Reliability and Validity from a Humanistic Constructivist Information-Processing Perspective.

In the past decade, there have been ample interests in the assessment of cognitive and affective processes and products for the purposes of meaningful learning. Meaningful measurement (MM) has been proposed which is in accordance with a humanistic constructivist information-processing perspective. Students' responses to the assessment tasks are evaluated according to an item response measurement model, together with a hypothesized model detailing the progressive forms of knowing/competence under examination. There is a possibility of incorporating student errors and alternative frameworks into these evaluation procedures. Meaningful measurement drives us to examine the composite concepts of "ability" and "difficulty." Under the rubric of meaningful measurement, validity assessment (i.e. internal and external validities) is essentially the same as an inquiry into the meanings afforded by the measurements. Reliability, measured in terms of standard errors of measurement, is guaranteed within acceptable limits if testing validity is secured. Further evidences of validity may be provided by in-depth analyses of how "epistemic subjects" of different levels of competence and proficiency engage in different types of assessment tasks, where affective and metacognitive behaviors may be examined as well. These ways of understanding MM can be codified by proposing a three-level conceptualization of MM, where reliability and validity are central issues for an explication of this conceptualization. (Author)