Collecting Non-Experimental Driving Behaviors: A Method Leading to External Validation of Experimental Outcomes

摘要

Human factors research in transportation safety has not clearly demonstrated howexperimental behaviors are related to public driving behaviors. The appropriateness ofdetermining external validity measures has been debated, and it is known that externalvalidity is not a necessary component in research if those findings do not need to transferoutside of that context; however, results in the transportation safety community are frequentlygeneralized, if improperly, to a stated risk in public driving situations. Few studies havevalidated driving behavior between experimental and non-experimental (observed behavior inpublic) settings. Proposed is an empirical method to establish a baseline of non-experimentaldriving behaviors to compare with experimental driving behaviors. This method involvesvideo recording public driving behaviors during the same time interval that data collectionoccurs in the real world (instrumented vehicle) and virtual world (driving simulator). Thesevideo recorded non-experimental driving behaviors at locations along an experimental routecan then be compared to experimental driving behaviors in instrumented vehicles and drivingsimulator scenarios (if developed) to establish empirical differences and potentiallycontribute to measures of external validity. This method introduces collecting nonexperimentaldriving behavior in order to establish methods that will generalize experimentalresults as they relate to public driving situations.