No, the dog doesn’t discriminate by religion, odd clothing choices, or nervous facial tics, but the dog team might mixed-metaphorically bite you on the rear.
The value of canines in explosives detection has been hotly debated in recent years, and I, for one, [full disclosure: being related to a highly experienced police K-9 trainer], have always been a believer in their capabilities, even with their real-world operational limitations. Some new recent research in animal behavior suggests another potential weakness may reside in the relationship of dog vs. handler.
Scientists at the University of California, Davis, wondered whether small unintentional signals from a handler might affect the dogs’ response. Seems like they do.
Trained dog teams (13 drug dogs, 3 bomb dogs, 2 dual-trained) each performed two sets of multiple searches; their hit signals varied between bark and sit. The physical environment was consistently controlled, but prior to the test searches, each handler was told some search areas might have three target scents, and in two cases those scents would be marked by pieces of red paper. The handlers were not told that two of the targets contained only decoy scents (sausages) to get the dogs’ interest, and none had drugs or explosives. Thus, a hit anywhere must be a false alert.
The data: 144 area searches with up to 3 targets each, yielded only 21 with no alerts; the rest had at least one alert; with a total of 225 alerts, all false. Why? The researchers claim when handlers saw the red markers, they were much more likely to say their dog signaled an alert. In rooms with red paper and no sausages, 32 out of a possible 36 alerts were raised. In rooms with both red paper and sausages, the figure was 30–about the same. Only 17 hits occurred in areas with a sausage but no red paper for handlers to react to.
The conclusion: the dogs were distracted roughly half the time by the sausage stimulus; human handlers were distracted virtually every time by the red paper, and apparently unintentionally transmitted that distraction to the dog, which responded accordingly.
It’s unclear how that might work in a complex and fast-moving operational airport environment with travelers from every walk of life and every corner of the world. But a dog handler, and nearly everyone else – yes, you too – has built-in biases and prejudices to varying degrees based on our collective life experiences, education, training and media exposure. For example, a handler might simply “tense up” when approaching a group of rowdy skinheads, sign-carrying protesters, or middle-eastern thobe-clad businessmen, any of whom, like many people, raise the ante by tensing up themselves when approached by a uniformed K-9 handler, who then, unconsciously or not, signals the dog for higher attention. The UC-Davis testing suggests a potential for as much as a 45-50% false alert rate. Hopefully, the minimal presence of this complaint in the anti-TSA flaming on internet blogs is a good sign. Alternatively, perhaps people, like K-9 dogs, are becoming trained to agree that this is an acceptable level of suspicion.
