|
Abstract
|
All of the considerations which prompt an investigator to use a laboratory rather than a reallife situation in research on behavior apply with full force to the study of accidents. But laboratory research on risk-taking and accident-inducing behavior is beset by several limitations. First, the ethics of laboratory experimentation preclude the introduction of hazards that can produce the more serious injuries all too common in real life. Secondly, it is doubtful whether a laboratory can replicate (or even effectively simulate) the numerous and complexly interacting behavioral and environmental variables involved in even the simplest accident-whether it be a vehicle-pedestrian collision or the ingestion of a household chemical.
As a result, every laboratory study of risk-taking must be subjected to two questions. First, is the subject behaving in the laboratory in much the same way as he behaves in the real world. This question, which is by no means peculiar to accident research, is asked in essentially quantitative terms: Is the subject as spontaneous? Is he more inhibited?
The second question is essentially qualitative: Does the laboratory situation elicit the same kind of behavior that a real-life situation would elicit? Is the subject driving a simulator responding to the stimuli presented by a real automobile on a real highway? Is the laboratory task eliciting all the responses involved in risk-taking or does it require of the subject only an affectless calculation of subjective or objective probability? A review of the laboratory experimentation on risk-taking uncovers few studies indeed that provide defensible answers to these questions.
It is for these reasons that the study of risk-taking behavior in a natural setting is crucially important. Yet the natural site, too, presents problems. Because accidents are infrequently occurring events, studying them in a natural environment is time-consuming, expensive, and often frustrating. To eliminate the possibility of observer effect, the investigator must remain concealed. And the question of ethics does not disappear. If the investigator observes an accident about to occur, does he welcome it because it provides data or does he break out of his role and try to prevent it?
The paper that follows offers interesting solutions to several of the problems posed above. By using cinematography with a concealed camera, the investigators not only eliminated the problem of observer effect but also produced a permanent set of objective data that lent themselves to repeated and detailed analysis. Their choice of time and place maximized economy in data gathering. And their decision to study the "near accident" as well as the injury-producing accident substantially increased the proportion of relevant data.
With respect to the study itself, however, some critical questions remain. Aside from economic considerations, was the site chosen for data-gathering the best possible one? Was the behavior of the subjects influenced by their proximity to the school, or by specific school instruction in street-crossing behavior? Would children in a playground or other non-school setting exhibit more spontaneous behavior?
Moreover, broadening a study to include the "near accident" must be done only after very careful consideration. If this is done on the assumption that the "near accident" is sufficiently similar to the injury-producing accident to be treated in the same way, the investigator may neglect a careful examination of the differences between the two events; in these differences may lie clues as to effective countermeasures.
Language: en |