Space Fortress as an Iq Test? Predictions of Learning and of Practised Performance in a Complex Interactive Video-game *

Claims that scores on pencil and paper IQ tests predict performance in easy laboratory perceptual motor rasks are weakened by methodological inadequacies. With an experimental design avoiding these weaknesses the AH 4 IQ test predicted rate of learning and performance after 5 days practice on ‘Space Fortress’ better than did age, between 18 and 36 years, or amount of previous experience at video-games. However, AH 4 scores only weakly predicted initial performance. Correlations between AH 4 raw scores and practised ‘Space Fortress’ game scores were at r = 0.69, that is, as high as are normally attained between different IQ tests. These findings remained robust even when re-examined within a sub-group of 43 subjects scoring within the top 10th percentile of the population, both on the AH 4 and on a much more difficult IQ test, the AH 5, designed for selection within University populations. When combined with AH 4 raw scores initial performance scores on training sub-tests for ‘Space Fortress’ were very efficient predictors of practised performance (69% or variance between individuals) and rate of learning (38% of variance between individuals). Clearly interactive video games can be tailored to provide very efficient selection and training instruments for complex industrial and military information handling tasks.