Dogs are being trained to sniff out COVID-19, and they're better than the tests

I expect the protocol used in this instance is sufficient to remove “clever Hans” responses (the article states the dogs don’t see passengers, they just have swabs presented to them). My personal feelings on the use of dogs as “production line sniffers” is that it would be difficult to maintain the dog’s interest. If you have a dog tracking game, this is something they have a natural motivation to do…a hound dog will follow a trail for days. Differentiating test samples for treats? Might work for a while, but to expect dogs to have a Protestant work ethic ( just like I don’t ) , that’s the reach for me. The expectation the a dog will perform to please the handler, this is the behaviour that would perhaps lead to false negatives; the dog has no commitment to truth, for the dog the goal is generating a positive response in his human.

Which shouldn’t matter. Do they see the handlers? Do those handlers, or any other staffers in the room see the passengers?

Cause the dogs take their cues from the handlers. Not the subjects of the search. Unless which swab is from who is blinded to the people in contact with the dogs biases of all sorts will appear.

This is exactly how the whole clever hans thing works. Hans took cues from his trainer.

The other end of it with detector dogs is the vaugeries of what’s considered a “hit”. There’s a significant amount of human interpretation in telling when the dog is indicating it’s found what it’s been trained to find. Even when there’s a specific behavior that’s trained for as the indication. False positives are often dismissed as the handler misreading, or the dog misbehaving. Additional signs are often added on the fly, or per dog. And there’s a ton of interpreting the general behavior of the dog, and attempts to extrapolate direction from it.

It’s very much like facilitated communication, or (again) a Ouija Board. That doesn’t go away when you seperate the passengers from the dogs. Cause it’s all rooted in the handler.

This is a known problem. As dogs are often used this way, particularly in airports. The dogs will be run up and down baggage, loading areas, border checkpoints, groups of people repeatedly. Over different sets, over long working days. The dogs apparently tend to get tired or bored and act out or act in ways that will be interpreted as a hit or the dog narrowing down. Dogs also have difficulty dealing with the extra stimulus in these circumstances, and will again act out or behave in ways that get read as “knowing” something. Even when nothing is found the dog’s agitation will be used to justify further search, questioning what have.

Interesting thing is we tend not to use scent hounds for this sort of thing. But shepherds. Which largely operate on sight, and we bred to control crowds of animals.

The other thing the hunting dog has going for it is a distinct, strong odor from a recently present animal in a defined area. Not traces of something in an endless group of strong smelling people and things.

1 Like

Scent hounds are not very biddable. I’ve had beagles and bassets; bassets are considered almost at the level of ability as bloodhounds. I think it’s because they have to value their own conclusions ahead of their handler, who after all doesn’t have any of the ability they are being employed for. The human’s role is to hang on to the rope and try to keep up. Bassets also won’t pile on if another dog thinks he hits a trail…beagles will. I can’t imagine trying to get a basset to check luggage or scan people for drugs. Sausage, maybe, but if you don’t let him have some he will lose interest pretty soon :slight_smile: Shepherds, OTOH, have been bred to work interactively with people, and they still have pretty good noses

That second test would need to be part of the testing strategy. I read nothing about that in the articles, and I assume it will not be.

Over here in Germany, a second test to confirm a positive result _is not mandated or even recommended _ in our testing strategy, even though a positive result requires people to self-quarantine.

So I take it that people making the rules often have little clue about this, and that goes for journalists writing about it, too.

In fairness, ol’ T. aquaticus may be more accurate; but there’s no way it’s more adorable than the dog option.
image
image

What, nobody’s brought up the old “Lab test and CAT scan” joke yet?

1 Like

This topic was automatically closed after 5 days. New replies are no longer allowed.