An in-vehicle information system (IVIS) was used to videotape drivers (N = 61) without their knowledge while driving 22 miles in normal traffic. The drivers were told that they were participating in a study of direction following and map reading. Two data-coding procedures were used to analyze videotapes. Safety-related behaviors were counted during consecutive 15-s intervals of a driving trial, and the occurence of certain safety-related behaviors was assessed under critical conditions. These two methods of data coding were assessed for practicality, reliability, and sensitivity. Interobserver agreement for the five different driving behaviors ranged from 85% to 95%. Within-subject variability in safe driving was more pronounced amoung younger drivers and decreased as a function of age. Contrary to previous research that has relied on self-reports, driver risk taking did not vary significantly as a function of gender. These results are used to illustrate the capabilities of the technology introduced here to design and evaluate behavior-analytic interventions to increase safe driving. (A)
Abstract