Three-dimensional spatial skill training in a simulated space station: Random vs. blocked designs

Background: Astronauts floating inside a spacecraft must be able to recall the direction to surrounding visual landmarks, regardless of their viewing perspective. If 3D orientation skills are taught preflight, should perspective sequences be blocked or randomized? Can standard spatial skill tests predict performance? Methods: Undergraduates (40 men and 40 women; ages 19-24) learned 3D spatial relationships among landmark pictures in a cubic chamber simulating a space station node. Subjects learned to predict picture directions when told one picture's direction (the one behind them) and the subject's simulated roll orientation, which was changed between trials by rotating pictures. The dependent variable was the proportion of correct predictions. A between group (n = 40 per group) independent variable was training type (random vs. blocked sequencing of perspectives). Experiment phase (familiarization, training, transfer, and 2 retention phases) was a within group variable. Subjects also took three standard spatial skill tests: Card Rotation, Cube Comparison, and Group Imbedded Figures. Results: As hypothesized, during training, performance for the random group (0.56) was worse than the blocked group (0.83); during transfer, the random group (0.75) was better than the blocked group (0.56); during retention-1, the random group (0.70) was better than the blocked group (0.55); and during retention-2, the random group (0.76) was better than the blocked group (0.65). Spatial skill tests correlated differently across the two groups, indicating that random sequencing elicits different skills. Conclusion: Random presentation enhances 3D spatial skill transfer and retention. Standard spatial tests can predict performance and have the potential to customize training.