The purpose of this study was to design a prototype written test of cognitive development based on the model of Piaget. Although the results were not consistent with Piaget's theory, test development was only partially successful. Specifically, validation of the scales was excellent or poor depending on the perspective. Analysis of the item difficulties within scales was excellent for the exclusion and combination scales and poor for the proportion scale. In fact, the prediction for the exclusion and combination scales was "perfect". Such results lend support to Piaget's contention of the invariant sequence of logical development for those logical structures tested. An analysis of those subjects' responses and reasoning on the formal items revealed that most had rotely learned how to solve proportion problems without understanding the underlying structural process. Such evidence would invalidate the contention that the "low" difficulties on the concrete items was due to the "high" readability of the items. Thus, it would seem that from a within scale analysis, the written test was generally successful, the criterion of cognitive level existence is still subject to question, and a different type of formal proportion item must be devised to alleviate the effect of specific past learning without understanding. (Author/DEP)