Dogs Like Meat More Than Drugs

From The Vancouver Sun comes the unsurprising news that dogs like the smell of meat more than the smell of drugs:

Federal search dogs at international border entry points have a penchant for sniffing out one thing more than anything else: meat.

In fact, dogs trained to find animal products turn up meat around 20 times more frequently than drug-sniffing dogs find narcotics, according to government documents obtained by Postmedia News under access-to-information legislation.

The release of the data comes as federal officials question the necessity and effectiveness of the dogs, with the Canada Border Services Agency dismantling some of its search-dog teams over the past year – a move the federal union believes will erode the ability to quickly search incoming cargo and seize drugs and firearms.

Of the more than 31,000 seizures of meat, animal products and animals in the 201112 fiscal year, 7,179 of the seizures (more than eight tonnes of illegally imported meat) were attributable to detector dogs. That’s compared to 360 seizures of drugs such as cannabis, cocaine and heroin (totalling 1,259 kilograms) coming from detector dogs out of 10,187 drug seizures in the same period, according to the documents.

As Ilya Somin notes, the article spends a lot of time trying to explain why meat-sniffing dogs are more successful than drug sniffing dogs without stumbling across the rather obvious answer that dogs like meat and that they’ve been trained through tens of thousands of years of biology and instinct to be able to find it. Any one who’s a dog owner can testify to how little time it takes before something meat-related is opened in the kitchen for a dog to come running looking for a a treat. Drug-sniffing dogs are animals who’ve been trained to use the sense of smell to detect something else and, as Somin notes, there have always been flaws in the process:

[A]s I discussed in this post, drug-sniffing dogs often err because their main objective is to please their human handlers rather than find the drugs as such; as a result they tend to signal “false positives” if they sense that that’s what the handler wants. By contrast, meat-sniffing dogs have reasons of their own for finding meat. The point is so glaringly obvious that this could be considered a dog-bites-man story – except that it is actually much more common for dogs to bite pieces of meat than humans.

Unfortunately, there is a more serious side to the story. Despite the fact that drug-sniffing dogs have a high error rate, government policy – and even Supreme Court decisions – are often based on the assumption that they are far more accurate than the evidence shows.

Perhaps stories like this will cause the powers that be to rethink those assumptions

FILED UNDER: Science & Technology, , , , ,
Doug Mataconis
About Doug Mataconis
Doug Mataconis held a B.A. in Political Science from Rutgers University and J.D. from George Mason University School of Law. He joined the staff of OTB in May 2010 and contributed a staggering 16,483 posts before his retirement in January 2020. He passed far too young in July 2021.


  1. The real headline should be “Journalists Don’t Understand Conditional Probabilities.

    You have to situations here:

    P(dog detects drugs) = P(dog detects drugs | drugs are present) * P(drugs are present)
    P(dog detects meat) = P(dog detects meat | meat is present) * P(meat is present)

    They notice they find meat a lot more than drugs (P(dog detects meat) > P(dog detects drugs)) and are reporting it as proof that meat dogs are more accurate (P(dog detects meat | meat is present) > P(dog detects drugs | drugs are present)), when it probably only means that a lot more people cross the border with meat in their vehicle than drugs (P(meat is present) > P(drugs are present))