Convergence of self-report and archival crash involvement data: a two-year longitudinal follow-up.

This study constructively extends Arthur et al. (2001) by assessing the convergence of self-report and archival motor vehicle crash involvement and moving violations data in a 2-year longitudinal follow-up. The relationships among these criteria, conscientiousness, and driving speed were also assessed using both predictive and postdictive criterion-related validation designs. Data were collected from a 2-year follow-up sample of 334 participants. Results suggested a lack of convergence between self-report and archival data at both Time 1 and Time 2. In addition, the predictor/criterion relationships varied across research design and data source. An actual application of our findings is that the interpretation of relationships between specified predictors and crash involvement and moving violations must be made within the context of the criterion-related validation design and criterion data source. Specifically, predictive designs may produce results different from those of postdictive designs (which are more commonly used). Furthermore, self-report data appear to include a broader range of incidents (more crashes and tickets), and thus researchers should consider using self-report data when they are interested in including lower threshold crashes and tickets that may not be reported on state records (e.g., because of the completion of a defensive driving course).

Three dimensional spatial memory and learning in real and virtual environments.

Human orientation and spatial cognition partly depends on our ability to remember sets of visual landmarks and imagine their relationship to us from a different viewpoint. We normally make large body rotations only about a single axis which is aligned with gravity. However, astronauts who try to recognize environments rotated in 3 dimensions report that their terrestrial ability to imagine the relative orientation of remembered landmarks does not easily generalize. The ability of human subjects to learn to mentally rotate a simple array of six objects around them was studied in 1-G laboratory experiments. Subjects were tested in a cubic chamber (n = 73) and a equivalent virtual environment (n = 24), analogous to the interior of a space station node module. A picture of an object was presented at the center of each wall. Subjects had to memorize the spatial relationships among the six objects and learn to predict the direction to a specific object if their body were in a specified 3D orientation. Percent correct learning curves and response times were measured. Most subjects achieved high accuracy from a given viewpoint within 20 trials, regardless of roll orientation, and learned a second view direction with equal or greater ease. Performance of the subject group that used a head mounted display/head tracker was qualitatively similar to that of the second group tested in a physical node simulator. Body position with respect to gravity had a significant but minor effect on performance of each group, suggesting that results may also apply to weightless situations. A correlation was found between task performance measures and conventional paper-and-pencil tests of field independence and 2&3 dimensional figure rotation ability.