In studies of college students, normal children at three grades, and learning-disabled children, physical and nominal matching procedures were highly correlated, (range .72-.96). In each instance, expected time differences favoring physical matches were replicated. Comparable results were found with both tachistoscopic and paper-and-pencil methods. In two studies where relevant data were available, correlations were not significantly attenuated when the effects of motor speed were partialled out. The results challenge the premise that nominal and physical matching measures are empirically independent, invalidate their application as measures of independent types of information processing, and help explain inconsistent findings in hemispheric specialization research.