Tests of spatial ability were analyzed for their analog (visuospatial) and nonanalog (verbal reasoning) components, using factor analyses of items and test scores. The self-selected sample consisted of over 2000 clients (average age about 26 or 27) employing the Johnson O'Connor Research Foundation's aptitude evaluation services in 12 metropolitan laboratories across the nation. Spatial tests administered were: (1) Incomplete Open Cubes; (2) Guilford-Zimmerman Spatial Visualization; (3) Wiggly Block; and (4) Paper Folding. Raven's Advanced Progressive Matrices was included as a measure of Spearman's general intelligence factor (G), and two tests of verbal reasoning ability (analytical and inductive) were given as a reference for nonanalog abilities. TESTFACT results distinguished item features that resisted nonanalog solution strategies (high loadings on first factor) from item features that promoted nonanalog strategies (high loadings on second factor). The sex difference favoring males depended on the analog component, and was largest for the Guilford-Zimmerman test, which requires the most "spatial" processing. The factor structure for test scores was less clear, partly because of ceiling effects. Two nonanalog factors were found, and it was noted that many tests currently used to evaluate spatial abilities depend largely on verbal reasoning components. (LPG)