A Rational Study of Ride-Related Human Error

Dr. Kathryn Woodcock and Janet Tsao, researchers from the School of Occupational and Public Health at Ryerson University in Canada, have published results of a two-part study on amusement ride-related human error. "Rider Responsibility and Amusement Ride Accidents: An Observational and Consensus Study of Rider Behaviours", can be read in its entirety in the Proceedings of the Human Factors and Ergonomics Society 49th Annual Meeting (2005). Highlights from the paper are presented here.

In 2002, the Natural Sciences and Engineering Research Council of Canada (NSERC) awarded Dr. Kathryn Woodcock a 4-year grant to establish a research program on the improvement of human performance in amusement ride safety. The THRILL project (Tools for Holistic Ride Inspection Learning and Leadership) was created to look at inspection-aiding by adapting lessons learned in aviation inspection to the area of amusement ride safety. Dr. Woodcock has worked closely with ride inspectors, including the Technical Standards and Safety Authority (TSSA), which regulates amusement ride safety in Ontario, Canada, and the National Association of Amusement Ride Safety Officials (NAARSO).

Woodcock and Tsao's paper compares field observation of patron and operator behavior at three Canadian carnivals with survey data from ride inspectors who were asked to compile consensus lists of the most troublesome patron and operator behaviors. The direct observations were conducted by non-industry observers trained in human error / human factors theory. The inspector survey was facilitated by NAARSO.

Human factors is an under-researched area of study in this industry. Investigation into amusement ride use/misuse patterns offers, perhaps, the greatest potential to improve safety -- particularly on older rides. Industry representatives attribute the majority of injuries sustained on amusement rides to rider behavior. The percentage commonly cited is somewhere around 80%, similar to the human error proportion cited in other highly mechanized industries, such as aviation or the automobile industry.

Human error is always a concern when devising machinery to be used by humans. Accepted theory suggests that systems be designed to provide error prevention, error capturing, and error tolerance. Woodcock and Tsao found that each of these is present in amusement rides, and each may also fail:

• Ride designers and owner/operators furnish certain error prevention mechanisms (restraints, warnings signs to deter certain behaviors and promote safe behaviors)
• Safety margins in ride design, installation, and operation help ensure a high degree of error tolerance .
• Ride operators perform error capturing functions such as enforcing height limits, guiding guests during load/unload, checking restraints, monitoring for unusual conditions, and performing an emergency stop if necessary. Woodcock and Tsao make the point that "operator's vigilance is not a perfect 'antidote' for rider error, however, because there may be no action available to the operator that would be safer than inaction.In addition, operators themselves may make errors."

Part 1 of Woodcock and Tsao's study involved field observation at three carnivals of small, medium, and large size in Ontario, Canada. Observers took note of the safety rules posted on or around the ride, then stationed themselves unobtrusively near the rides and observed for several ride cycles, noting safety-related errors including violations of posted rules. Errors were classified according to who initiated the error (operator, rider, or parent/observer), and by type of error.

The three carnivals observed in the study had 103 separate installations of 88 different rides and attractions. In total, over 5000 rider exposures were observed. 81 errors were noted. 60 of the observed errors would customarily be considered rider error; 19 were operator errors; and two were errors of others (adults outside the fence contacting a child on a moving ride). 43% of the rider errors were considered violations because existing rules contraindicated them, although "qualitatively, none of the errors defied comprehension to the observer." Only 6 minor injuries were observed, none requiring any treatment at the time. The observed rider error rate was approximately 1.4% of the ride exposures observed in comparison to the observed injury rate of 0.1%.

In almost all observed cases, the error tolerance of the ride was adequate to prevent injury (i.e., most of the patrons who made errors escaped injury, due to some combination of safety margins and luck). The high degree of error tolerance was fortunate, according to the authors, since error capturing and error prevention failed in almost every case: "Few of the rider behaviors were intercepted and some were not even observable from the operator's position. Some operators were doing two incompatible tasks at once (e.g., checking tickets, warnings riders not to run to the exit). Others were not watching the ride, or apparently decided to permit the violation or mistake. In a few cases, the operator encouraged the rider error...In almost every case, some type of error prevention mechanism existed but failed. For example, some restraints as designed allowed some riders too much discretionary movement and allowed them to adopt prohibited postures."

In Part 2 of Woodcock and Tsao's study, NAARSO organizers grouped participating industry members so as to distribute different levels of ride inspection experience evenly among 12 groups of 15 people on average. A seminar leader asked the groups to list "the top five bad things riders do and the top five bad things operators do", grouped by ride phase (in line & while loading, during ride, while unloading, and non-riders outside ride). The groups brainstormed, reached consensus, and reported back to the whole forum.

The list of behaviors created by the NAARSO inspectors was similar to the list of errors observed at the Canadian carnivals, with a few exceptions. The most common type of rule violation observed on the midway -- leaning/hanging/reaching out or not hanging on -- was not mentioned by any of the inspector groups in their lists of dangerous behaviors.

Although the experienced inspectors identified similar behaviors to the errors tabulated by researchers, the inspectors used what the authors called a "discourse of violation" when talking about human error. In other words, the neutral observers saw a range of possible causes for usage errors (e.g., mistiming, losing place in a sequence of steps, lack of information, misinterpreting the situation), while the inspectors' comments mirrored the prevalent industry view of patron errors as intentional, irresponsible acts. "This type of description tends not to acknowledge the potential for errors to occur in legitimately motivated, non-reckless behavior through sensory illusion, excess mental workload, or lack of information"

Woodcock and Tsao note that using thrill rides involves repeated short periods of high mental workload, conditions that can hamper situational awareness. Enjoyment of the ride, cited by the authors as the rider's primary job, requires immersion in the ride narrative or illusion. "In the short duration of the ride, the rider must process rapid, extreme sensations and illusions as well as her responses, such as surprise, fear, and physical discomfort. This may leave little or no capacity for the additional mental workload of recognizing emerging situations in the real physical world, recalling numerous safety instructions and abstract principles, and applying the principles promptly during the ride. The task would require the rider to maintain situation awareness of two parallel situations: the physical reality and the ride illusions."

The authors point to effectively designed affordances as an ideal design strategy for error reduction in the amusement ride industry. For example, individual lap bars that automatically open when the ride reaches a safe unloading point provide the dual benefits of 1) effective physical protection for riders of all sizes and 2) cues to inform the rider that it's safe to exit. "Design features that match affordances to the actual hazards of each ride would be the most effective strategy, eliciting safe behavior naturally through cues in the world."

At the operations level, clear communication and consistent enforcement of all rules was suggested. When operators allow, or actively encourage, riders to break some rules (e.g., riding with hands in the air), patrons are likely to take all safety rules less seriously.

Woodcock and Tsao have given us new tools for understanding human error related to use of amusement rides. As I see it, here is the takeaway value for consumers:

• Both parts of the study show that consumers need to pay closer attention to safety, both for themselves and their children. When visiting amusement parks and carnivals, make sure you understand the instructions, warnings, and rules -- and consistently follow them. The authors warn that posted instructions are sometimes too vague ("do not take dangerous articles onto ride") or cryptic ("participate responsibly") for children and adolescents to apply reliably to their own circumstance in real time. Parents should take extra time to explain the meaning of the rules, and the reasons behind them, to children.
• Parents and guardians should be aware that the combination of high error tolerance and low error capture can create situations where unsafe behaviors by children and adolescents become learned. When both natural consequences (i.e., injury) and logical consequences (i.e., scolding by operator or revoking of riding privileges) are rare, children may not take safety rules seriously. Parents should set and enforce consequences, regardless of whether the park or carnival enforces its own rules, to ensure that children know unsafe behavior around thrill rides is not tolerated.
• Most importantly, parents should take careful note of Woodcock's and Tsao's caution regarding the high mental workload of riding thrill rides. Amusement rides that appear quite simple and benign to an adult can be confusing, frightening, and/or dangerous for some children. Be cautious when deciding whether a ride is right for your child. Take time to look at the machine and the load/unload cycle from a child's point of view. Watch a cycle with your child before he or she rides. Point out how the ride and its restraint system work, where/how riders load and unload, and read the rules together. If you or your child have any worries, find another ride or attraction to enjoy. Amusement parks and carnivals offer something for everyone, but it may take more than a simple height measurement to determine whether a particular ride is a good, safe choice for your own child.

From Saferparks' perspective, Woodcock and Tsao's paper offers an additional benefit by creating a space in which the two different interpretations of human error can coexist, where both have their place and both are respected. Given what ride inspectors see and do, they cannot help but be frustrated by the errors we humans make around thrill rides, and the disastrous consequences that sometimes result. Yet it's also important for inspectors and those in the industry to imagine what it might feel like for a 5-year-old child who's being herded onto a huge, powerful, unfamiliar machine that looks cool and scary at the same time, or his worn-out mother who has to figure out how to secure her children inside yet another huge, powerful, unfamiliar, cool, and scary machine in the few seconds before it launches -- and then navigate the family through three more "must see" attractions before closing time. Mistakes will be made.

Serious accidents do result from human error, despite the admirable level of error tolerance found in the sample of 5000 exposures observed by Woodcock, et al. Just about every type of error cited in this paper has caused serious amusement ride accidents, most often involving children.

All amusement rides are not equally error tolerant, and all riders are not equally protected by error tolerance. A child riding with hands in the air is at higher risk of ejection than an adult who does the same. Rides that use single lap bars on bench seats are less error tolerant than rides with individually-adjustable restraints, especially for children. Parents need a reliable way of determining whether a particular ride is designed to tolerate the types of errors children like theirs make when using amusement rides. I anticipate that Dr. Woodcock's work will help move us toward that goal.

Dr. Woodcock has been invited to address a task group of the ASTM F24 committee next month. She is currently working on a human error investigation tool for the amusement ride industry and human factors analysis of simple amusement ride control panels.

Her research interests cover a range of topics in health and safety and injury prevention. They include hazard perception, safety inspection and accident investigation practices and tools, professional development of safety specialists and safety program decision-making and management, achieving safety through design and health, safety and ergonomic implications of deafness and assistive technology. Dr. Woodcock can be reached by email at kwoodcoc@ryerson.ca.

Saferparks wishes to thank everyone involved in this project, including Dr. Woodcock, Ms. Tsao, Dymytro Sheyrnin, the Natural Sciences and Engineering Research Council of Canada, Ryerson University , the Technical Standards and Safety Authority (TSSA), Steve Elliott, the National Association of Amusement Ride Safety Officials (NAARSO), and the Human Factors and Ergonomics Society.