Visual Expert Human Factors: Is The Moth Effect Real?

Marc Green

"'I had an almost overwhelming feeling that something was pulling me toward the lead plane…Sometimes I had the feeling that I could do nothing to prevent this.'" Clark, Nicholson, & Graybiel (1953), p 433.

Clark et al. studied pilot errors due to "fascination," concentration on some object or task that caused loss of voluntary control over response. They divided fascination into subcategories with the quote above exemplifying Type B-2, where pilots suffering felt drawn to a target and could not avoid the attraction. If the target is a light viewed at night, then the phenomenon is now called the "moth to the flame effect" or more often the "moth-effect."²

Some (e. g., Younger, 1997) believe that the moth-effect causes road accidents. In one common scenario, a driver inexplicably steers off the road and collides with an emergency vehicle parked on the shoulder. Presumably, the fascination/moth-effect caused the driver to steer toward the brightly flashing waning lights. Moth-effect advocates have usually supported this belief with anecdotal accident reports supplemented by evidence (Taylor & Sucov, 1974; Murdoch & Caughey, 2004) showing that humans innately orient toward light

Despite these arguments and the reports of airplane pilots, many (e. g., Summala, Leino, & Vierimaa, 1981; Agent & Pigman, 1990), are skeptical and argue that no empirical support exists. Wells (2004) summarized the view of many by saying, "There are no known studies that have not been disproven that substantiate the actual existence of this effect in real world driving." However, several recent experimental studies have convincingly demonstrated the functional equivalent of a moth-effect.

Literature Review

The first attempts to validate the moth-effect compared accidents rates for emergency vehicles equipped with and without warning lights. The moth-effect would appear as an increase in the accident rate for equipped over unequipped vehicles. Charles, Crank, & Falcone (1990) reviewed several accident analyses and concluded that the data showed no evidence for increased accident rates in marked cars and hence no clear evidence for the moth-effect.

In reality, however, these studies say very little. First, marked vehicles would be more visible. Any increase in accident rate due to the moth-effect would likely be offset by a decrease in accident rate due to higher visibility. At most, the studies could reveal the cost/benefit tradeoff of the two effects. Second, marked and unmarked vehicles may have differed in other ways that confound the studies. For example, marked cars might be newer and used more often or used in different ways. Further, the respondents in the Clark et al. (1953) report believed that type B-2 fascination caused about 5% of pilot errors while Sipes, Lessard, & Heideman (1998) agreed that the moth-effect causes accidents but concluded it was a relatively rare source of pilot disorientation. If these conclusions transfer to road accidents, the moth effect may simply be too small effect to rise above noise in these poorly controlled data. Lastly, the moth effect might depend on an enabling variable, such as driver fatigue or distraction that was not captured in the results. Lumping all accidents together would obscure key variables, so a simple examination of global accident rates is not likely to uncover evidence for the moth-effect.

Early experimental tests of the moth effect also failed to produce support. (Helander, 1978) sought to provide direct evidence by testing the steering of drivers as they met an on coming car. He found that drivers began steering away from the oncoming vehicle but that 2 seconds before the meeting drivers deflected the steering wheel in the direction toward the approaching vehicle. The deflection maximized just as the oncoming vehicle passed, a result that was consistent with a moth-effect. Helander (1978) further argued that the steering could not be a compensation for the initial steering away from the oncoming vehicle. The study, however, measured only steering wheel deflection and not actual lane position.

However, a later study of lane position succeeded in finding evidence for a moth effect. Kitamura & Matsunaga (1994) measured lane position for drivers passing stopped roadside emergency vehicles. The study tested three conditions: 1) emergency lights extinguished, 2) emergency lights operating with no specific instructions to the drivers and 3) emergency lights operating with drivers instructed to attend the lights. Drivers instructed to fixate the lights passed closer to the parked vehicle than drivers with no specific instructions and drivers who passed the vehicle when the hazard lights were extinguished. These results are the first clear, experimental demonstration of a moth-effect. Moreover, they implicate driver attention as a possible enabling variable.

Two recent studies provide even more compelling evidence. Readinger, Chatziastros, Cunningham, Bülthoff. H., & Cutting (2002) had drivers steer a simulator car down a virtual road while performing a Landolt C acuity task for targets fixated a various eccentricities. Drivers began steering their cars in the direction of fixation only a few seconds after the beginning of each trial. The magnitude of the lane deviation generally increased over time, although there was a tendency to make corrections giving the lane position a periodic component. The lane deviations averaged about 1.2 meters but reached distances as great as 2.5-3 meters (Chatziastros, 2006). These effects are certainly big enough to cause a driver to inadvertently leave the road. A subsequent study (Chatziastros, Readinger, & Bülthoff, 2003) replicated the original and also discovered that the effect was bigger when the scene contained few trees to provide visual information.

What Causes The Moth-Effect?

The last two studies showed that drivers might steer off the road in the direction of their fixation. Neither study, however, employed bright lights, so it is unlikely that the "moth-effect" results from an innate phototaxis. Moreover, people frequently look at lights by the roadside without steering off the road, so any simple appeal to phototaxis explains little. So what causes the "moth-effect?"

The moth effect seems to have two enabling conditions. The first is minimal optic flow information, as when a driver is on a dark road at night or perhaps traveling in bad weather. The second is an intense attentional fixation on a roadside target. The lack of optic flow likely removes the normal source of heading information, forcing the driver to rely on a sense of egocentric direction relative to a landmark - the fixated object. When people look in the direction of travel, the egocentric straight-ahead direction and eye direction are the same. When people fixate away from the direction of travel, then they must then use knowledge of eye position to calculate maintain a proper sense of egocentric direction. If the calculation is correct, then the person has maintained directional constancy. Studies (e. g., Hill, 1972; Morgan, 1978) show that people are unable to maintain their sense of egocentric direction when fixating eccentrically. Instead, the sense of straight-ahead moves in the direction of fixation. In other words, the driver looking right while attempting to travel down the road straight will steer to the right of roadway in an attempt to steer straight.

The intense attentional focus can play several roles. Attention is a zero-sum game, so the more attention focused on one task, the less available to others. Concentrating attention on a target might reduce attention available to maintaining directional constancy. Even the most automatic behavior, such as maintaining lane position, still requires some degree of attention. It might also prevent the driver from noticing cues for steering correction. The perceptual narrowing might prevent the driver from monitoring road delineations in peripheral vision. The driver would not be aware they he last lost lane positioning. Similarly, the driver may fail to notice to tactile cues that occur when a driver leaves the paved roadway. Research (Summala, 1998) shows that steering becomes generally erratic when drivers fixate eccentrically.

Lastly, drivers who start with less attentional resource should suffer a greater chance of suffering the moth-effect. Drivers who are fatigued, bored, affected by drugs or alcohol, or older should be more prone to steer off the road. However, it is not these conditions that directly cause the result. It is the way they affect distribution of attention.

Conclusion

The "moth-effect" is a myth in one sense and reality in another. The idea that drivers may steer off the road when they fixate flashing lights is likely correct, but they are not drawn to the lights like moths to a flame. Rather, they inadvertently steer rightward, which may or may not take them into collision with the roadside vehicle. Even normal, alert drivers in daylight conditions may steer in the direction of eye position during periods of intense fixation. The cause is likely error in judging heading under eccentric fixation when optic flow cues are minimal and when attentional focus prevents sensing of the need to correct the error. Although bright lights and fascination are not required, of course, it is impossible to rule out these factors in some accidents.

References

Agent, K., & Pigman, J. (1990). Accidents involving vehicles parked on shoulders of limited access highways. Transportation Research Record, 1270, 3-11.

Charles, M., Crank, J., & Falcone, D. (1990). A Search for Evidence of the Fascination Phenomenon in Roadside Accidents. Washington D.C.: AAA Foundation for Traffic Safety.

Chatziastros, A., Readinger, W., & Bülthoff, H. (2003). Environmental variables in the "moth effect". Vision in Vehicles X.

Chatziastros, A. (2006). Personal communication.

Clark, B., Nicholson, M., & Graybiel, A. (1953). Fascination: A Cause of Pilot Error. Aviation Medicine, October, 429-440.

Helander, M. (1978). Drivers' steering behavior during traffic events: A case of perceptual tropism? Human Factors, 20, 681-690.

Hill, A. (1972). Directional constancy. Perception & Psychophysics, 11, 175-178.

Kitamura, F., & Matsunaga, K. (1994). Influence of Looking at hazard lights on car-driving performance. Perceptual & Motor Skills, 78, 1059-1065.

Morgan, C. (1978). Constancy of egocentric visual direction. Perception & Psychophysics, 23, 61-68.

Murdoch, J., & Caughey, C. (2004). John Flynn and the Psychological Effects of Lighting. Lighting Design & Application, 34, 69-73.

Readinger, W., Chatziastros, A., Cunningham, W., Bülthoff. H., & Cutting, J. (2002). Gaze-eccentricity effects on road position and steering. Journal of Experimental Psychology: Applied, 8, 247-258.

Sipes, W., Lessard, C., & Heideman, D. (1998). Spatial disorientation: what kinds and how often? Proceedings of the 1998 36th SAFE Annual Symposium Sep 14-16 1998 (pp. 164-172). Phoenix, AZ, USA : Survival Flight Equipment Assoc.

Summala, H. (1998). Forced peripheral vision driving paradigm: evidence for the hypothesis that car drivers learn to keep in lane with peripheral vision. In Gale, A. E., Vision in Vehicles VI, 51-60. Amsterdam: Elsevier.

Taylor, L., & Sucov, E. (1974). The movement of people toward lights. Journal of the Illuminating Engineering Society, 3, 237-241.

Wells, J. (2004) Florida Highway Patrol Emergency Lighting Research & Prototype Evaluation [Web Page]. URL http://www.theiacp.org/div_sec_com/Committees/Less/FHPevaluation.pdf.

Younger, J. (1997) The Moth Effect and How to Beat It [Web Page]. URL http://www-afsc.saia.af.mil/magazine/htdocs/win98/mothefect.htm.

Footnotes

¹A version of this article appeared in Accident Reconstruction, May/June, 2006.

²The technical term for orienting toward light would be either "phototaxis" or "phototropism."