Affiliation:
1. State University of New York at Albany
Abstract
Various models of assessing construct validity are considered and found to be lacking. In addition to the two dimensions of (a) trait to be assessed; and, (b) method of assessment, a third dimension of time or age is offered as a logical extension when involved in reading research. A computer simulation for learning performance on five hierarchically arranged skills is offered as an illustration. The results of the simulation suggest that assessment specifically based on hierarchical models requiring prior skill development should take into account the point in time at which the assessment is made and that, more generally, all settings requiring the learning of data are time related.