A continuing and vexing problem associated with survey instrument development is the creation of items, initially, that correlate favorably a posteriori with constructs being measured. This study tests the use of symbolic-logic matrices developed by D. G. Coleman (1979) in creating factorially "pure" statistically discrete constructs in survey instrument design. The study used symbolic-logic (verbal logic) "truth tables" to increase statistical factor loadings in confirmatory factor analysis studies and to develop discrete constructs. Initial experimental work with truth tables increased, hypothetically, construct purity in construct validation. Construct validation helps to assure that items created for measuring specific theoretical or real constructs do, indeed, represent the distinct qualities measured. In November 1996, the National Association of Elementary School Principals (NAESP) approved the revision of the Administrative Diagnostic Inventory (ADI), the title given to the diagnostic procedure conducted within their assessment center. To assist the NAESP in the use of the ADI for the assessment of current and prospective school principals, this study evaluated the construct validity of the instrument with 203 teachers enrolled in educational administration programs. Results show the improved content validity and construct validity of the revised ADI. Factor analysis shows that the descriptors within the skill dimensions are highly correlated and are significant measures of the constructs for which they were written. (Contains 2 tables and 20 references.) (SLD)