Identification and measurement of select factors which bias causation analyses of accident phenomena

Abstract

The allocation of accident countermeasures to prevent injury, damage, and other losses within a man/machine/environment system is partially based upon the findings of accident cause analyses. The effectiveness of countermeasures is related to the accuracy of these analyses. To the extent that bias is introduced into cause analyses it will, in turn, introduce error into the rational allocation of accident countermeasures and result in higher overall losses than what might otherwise occur. The primary objectives of this research were to identify and characterize two possible sources of bias in accident cause analyses: (1) the bias introduced by the investigator's academic background and (2) the bias introduced by knowledge concerning the severity of the accident outcome. Relative differences in cause allocation were examined for select individuals with (1) education, (2) management, and (3) engineering academic backgrounds. Relative differences in cause allocation were also examined for accidents associated with (1) relatively "minor" outcomes, (2) relatively "severe" outcomes, and (3) outcomes designated as "unknown." A bi-polar scale using the terms "unsafe acts" (human errors) and "unsafe conditions" (physical/environmental errors) as extremes was utilized. Differences in allocation of cause along this scale were examined, while holding sample accident descriptions constant, and varying academic background and outcome severity..