Volunteer studies of experimental, low-velocity rear-end collisions have shown a percentage of subjects to report short-lived symptoms, but the cause of these symptoms remains unknown. It is unclear whether the symptoms arise from biomechanical stress causing injury or from psychological stress causing symptom expectation and anxiety. Similarly, the cause of symptoms remains obscure in virtually all "whiplash" patients because it is impossible to identify acute pathology in many cases. In this study subjects were exposed to placebo collisions that almost completely lacked biomechanical stress. It was highly probable that if the symptoms reported following low-velocity collisions were not due to injury but to other factors (including misattribution of symptoms from other sources), then the proportion of subjects reporting symptoms would be similiar to that reported for volunteers in true (experimental) low-velocity, rear-end collisions. A total of 51 volunteers (33 males and 18 females, mean age 32.4 years) were recruited through local newspaper advertisements. An experimental set-up for a placebo collision was constructed using two standard European cars. At time T0, prior to the placebo collision, a history and physical examination was performed, including a psychological analysis (Freiburger Personality Inventory). A symptom history and physical examination were also performed at time T1, immediately after the placebo collision, and the subjects completed symptom questionnaires 3 days (time T2) and 4 weeks (time T3) after the placebo collision. Data analysis included a determination of the predictive value of psychological data for the presence of symptoms following exposure to a placebo collision. At time T1, 9 out 51 participants (17.6%) indicated symptoms. Within 3 days (time T2) after the placebo collision, 10 (19.6%) of the subjects had symptoms, and within 4 weeks (time T3) 5 subjects (9.8%) had symptoms. Of the last group, two of the five did not relate these symptoms to the "collision". Subjects who endorsed symptoms at time T1 had significantly higher scores on the psychological scale of psychosomatic disorders (measured at time T0). Subjects endorsing symptoms at time T2 had significantly higher scores on emotional instability. There was also a tendency to higher scores on this sub-scale for subjects with whiplash-associated disorders (WAD) at time T3. A discriminant analysis using all four psychological scales from time T0 had a power of 87%, 83% and 92% for correct classification of subjects as asymptomatic times T1, T2 and T3, respectively. Approximately 20% of subjects exposed to placebo, low-velocity rear-end collisions will thus indicate WAD, even though no biochemical potential for injury exists. Certain psychological profiles place an individual at higher risk for phenomenon.