Equivalence of Conventional and Computer Presentation of Speed Tests

This study examined the effects of computer pre sentation on speeded clerical tests. Two ratio scores— average number of correct responses per minute and its inverse, average number of seconds per correct re sponse—were examined as variants of the conven tional score, number of correct responses in a fixed in terval of time. Ratio scores were more reliable than number-correct scores and were less sensitive to test ing time. Tests administered on the computer were found to be at least as reliable as conventionally ad ministered tests, but examinees were much faster in the computer mode. Correlations between paper-and- pencil and computer modes were high, except when task differences were introduced by computer imple mentation.