Abstract
The task of visual search is ubiquitous in everyday life. I present the case for studying one real-world search task - lifeguarding - in the laboratory. The lifeguard task encompasses a scanning component, a fixation component, and an evaluation component where the lifeguard determines whether a swimmer might need assistance. Research seems to indicate that lifeguards' scanning is continuous, so missed drownings may be due to faulty evaluation. Lifeguards are taught behaviors associated with drowning - bobbing vertical body with flailing arms and tilted head - but, more importantly, they must anticipate a drowning by closely examining weaker swimmers, who are most likely to drown. In my study, three groups of subjects saw 60 short video clips of varied swimming scenes while an eyetracker monitored their gaze position. The Naïve group was told to look at things that interested them. The Taught group was taught the behaviors above and told to act as a lifeguard and look for those behaviors. The Lifeguard group contained trained and certified lifeguards who monitored as if on the job. Two analyses were done - one on categorizing the target of fixations longer than one-third second and one on “critical events” (i.e., events in my clips identified as “important to monitor” by two persons who teach and research lifeguarding). So far in the analysis, Lifeguards spend more time scanning and less in fixation than Naïve subjects, and spend more time looking at weak swimmers (indicated by, e.g., slow speed). Lifeguards also fixate more critical events than Naïve subjects, although they are not perfect and even Naïve subjects fixate some of the critical events. The Taught subjects perform in between the other groups, and even though they look for the behaviors, they do less well at anticipating future problems and frequently saccade away from a critical event prematurely.
Supported by NIH Grant MMH067793 to C.M. Moore.