Error rates in a multiple-choice spatial manipulation test.

: Error rates in a multiple-choice spatial manipulation test were analyzed post hoc on a number of different test item variables with the intention of building a computational cognitive model that simulates error rates on a spatial manipula tion test. 12 men (Army soldiers), ranging in age from 19 to 38 years, with an average age of 29 yr., participated. The variables included degree of rotation, connectivity, number of blocks in each polygon, types of distractors used, and compactness. Significant effects of connectivity, number of blocks in each polygon. type of distractor, and compactness on error rates were found, with mirror-image distractors being selected more often than nonmirror-image distractors. The amount of rotation of each image was not an influential factor in overall error rate; however, there were indications from verbal protocols that participants were using nonrotational strategies to solve the problems. Our findings suggest that complexity, i.e., connectivity, number of blocks in each polygon, and compactness, may affect error rates for nonrotational manipulation of visual images. Findings also suggest that participants may have difficulty evaluating "handedness" of images while using nonrotational manipulations.