The validity of a spatial visualization test for selecting candidate engineering students was assessed. The test was administered to 300 out of the 318 students enrolled during the winter quarter of 1989 in the Engineering Graphics 110 course at a major state university. Developmental steps undertaken to enhance the theoretical basis and validity of a preliminary version of the test are reported. The 48-item multiple-choice test was developed based upon an analysis of mental processes used by engineering graphics experts. Two major skills, visual synthesis and decomposition, were identified. Within these major skills, four other skills were identified: pattern recognition, two-dimensional versus three-dimensional visualization, ability to move images mentally, and ability to perceive an object from another direction by an orientation or rotation method. Scoring of tests and report generation were handled on a microcomputer. Results of the test administration were correlated with Scholastic Aptitude Test (SAT) and American College Test (ACT) scores as well as with engineering graphics course examination scores. Results indicate that the new test is a better student selection device than are the currently used SAT and ACT scores. Two sample test items and one data table are included. (TJH)