Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Study of pet dogs shows breed does not predict behaviour

nature.com

Post reply on HN
Print view View original

41–50 of 191 posts

Re: Study of pet dogs shows breed does not predict behaviour

#41

Post by shard972 »

echelon wrote: post #34
ShockTohp wrote: post #7

From the source: "The team found that some traits were more common in certain breeds. For example, compared with a random dog, German shepherds were more easily directed; beagles, not so much. And the authors’ genetic studies revealed that mixed-breed dogs with a particular ancestry were more likely to act in specific ways. Mutts with St Bernard ancestry, for instance, were more affectionate, whereas mutts descended…

This study/paper is ridiculous. Anecdotal, but I've had lots of dogs growing up. We have five right now. This does not reflect the reality I know. Shiba inu act like cats and prefer to spend time away from their person. They frequently get kicked out of dog parks for not being friendly to other dogs. It's documented and well-studied. They're not at all like other dogs. Corgis are protective, commandeering, and can be…

No post body was provided.

Re: Study of pet dogs shows breed does not predict behaviour

#42

Post by cato_the_elder »

echelon wrote: post #34
ShockTohp wrote: post #7

From the source: "The team found that some traits were more common in certain breeds. For example, compared with a random dog, German shepherds were more easily directed; beagles, not so much. And the authors’ genetic studies revealed that mixed-breed dogs with a particular ancestry were more likely to act in specific ways. Mutts with St Bernard ancestry, for instance, were more affectionate, whereas mutts descended…

This study/paper is ridiculous. Anecdotal, but I've had lots of dogs growing up. We have five right now. This does not reflect the reality I know. Shiba inu act like cats and prefer to spend time away from their person. They frequently get kicked out of dog parks for not being friendly to other dogs. It's documented and well-studied. They're not at all like other dogs. Corgis are protective, commandeering, and can be…

It seems that every day we are getting one step closer to Lysenkoism [1].

[1]: https://en.m.wikipedia.org/wiki/Lysenkoism

Re: Study of pet dogs shows breed does not predict behaviour

#43

Post by TameAntelope »

dTal wrote: post #39
TameAntelope wrote: post #21

Earlier quoted context omitted.

Trusting your gut over a paper published in Nature. Bold move! Some people would use this as a moment to question what they know about dog breeds, but you choose to double down on your beliefs. Very human of you! [0][1][2][3] [0] https://www.psychologytoday.com/us/blog/the-pathways-experie... [1] https://en.wikipedia.org/wiki/Belief_perseverance [2] http://aphilosopher.drmcl.com/2015/12/11/doubling-down/ [3] https://…

https://en.wikipedia.org/wiki/Argument_from_authority If Nature put out a paper saying the sky is red, it would be entirely rational of me to question what on Earth is going on at Nature, rather than my years of experience of blue skies.

https://en.wikipedia.org/wiki/False_equivalence

They didn't put out a paper saying the sky is red, they put out a paper saying the magnitude of impact from breed on dog behavior is substantially less pronounced than previously thought.

It is not rational for you to double down on this belief.

Re: Study of pet dogs shows breed does not predict behaviour

#44

Post by bpodgursky »

dTal wrote: post #38
User23 wrote: post #20

Earlier quoted context omitted.

I'm not going to spell it out because it's one of those observations that's taboo even here, but it's pretty obvious why in the prevailing political climate this piece would get published in prestigious journals and a piece finding otherwise would not.

You should spell it out, or stay silent. This middle ground is cowardly and contributes nothing.

"It is currently politically unacceptable to make any claims that human behavior is primarily genetically influenced, and strong evidence in support of genetically-determined behavior in dogs will beg the question of why (or if) we don't see the same patterns in human populations."

Re: Study of pet dogs shows breed does not predict behaviour

#45

Post by InitialLastName »

daenz wrote: post #35
InitialLastName wrote: post #26

Earlier quoted context omitted.

They aren't saying that there's no physical differences in dogs' capabilities. Pits can bite and shake because they have ridiculous jaws and a powerful build. If a miniature poodle tried, it wouldn't have much effect. The reporting on this is also sketchier than the science; the honest headline "Study of pet dogs shows breed is imperfectly predictive of behavior", which could be confirmed by anyone with multiple dogs…

>Pits can bite and shake because they have ridiculous jaws and a powerful build. If a miniature poodle tried, it wouldn't have much effect. They got that way because humans designed them that way intentionally with selective breeding. It's not so much "they do X because Y" than it is "they do X because they were designed to do X, and Y is the result of that design."

For sure, but again, that's a characteristic that is enabled by a physical difference.

Literally nobody here is saying there's no physical difference between breeds of dog.

Re: Study of pet dogs shows breed does not predict behaviour

#46

Post by srveale »

synaesthesisx wrote: post #29

To be frank, this "study" is absurd & invalid, and there is 100% an association between breed and behavior (we literally breed dogs for certain behavioral traits!). On that topic, I realize it's controversial but think pitbulls should be banned outright - they are responsible for the majority of fatal dog-related attacks and there is a wealth of evidence indicating that they are not a desirable breed. https://www.red…

From the study: "Owner survey responses are susceptible to rater bias, including the influence of breed stereotypes." That's all I could find acknowledging that this bias exists. I'm inclined to believe they know this type of bias completely invalidates the study, which is why they didn't want to explore it further.

Imagine a dog owner saying "Strongly agree" to something like "Dog has a tendency to attack (or attempt to attack) other dogs", or "Dog behaves aggressively towards children." (I believe these were questions in the surveys given to the dog owners).

> No no, not my precious Bulldozer, he's a sweety, just gets excited sometimes. I'll only put "neither agree nor disagree", because he only tries to attack dogs who are smaller than him.

I mean, I get it, nobody wants to think ill of their beloved pet, but that's the point. A study like this needs objective measures.

Re: Study of pet dogs shows breed does not predict behaviour

#47

Post by defen »

tkahnoski wrote: post #23
ShockTohp wrote: post #7

From the source: "The team found that some traits were more common in certain breeds. For example, compared with a random dog, German shepherds were more easily directed; beagles, not so much. And the authors’ genetic studies revealed that mixed-breed dogs with a particular ancestry were more likely to act in specific ways. Mutts with St Bernard ancestry, for instance, were more affectionate, whereas mutts descended…

Lots more details from the authors on twitter. https://twitter.com/eenork/status/1520450288095797251

The author says "We found no support for policies targeting particular breeds" [1] which is completely absurd even if you postulate that their study's results were even stronger than they are. Imagine that their study found zero correlation between breeds and behavior. Even if that were the case, the outcomes of that behavior are wildly different. Some breeds will bite you and not even break skin. Some breeds will bite you, and you'll need stitches and antibiotics. Some breeds will literally rip your arms off. How could any reasonable person conclude that "no difference in behavior" leads to "no support for policies targeting particular breeds"?

Is there anyone here who would let their 6 year old child play, unsupervised, with a random pit bull? How about a random chihuahua?

[1] https://twitter.com/eenork/status/1520459786575396869

Re: Study of pet dogs shows breed does not predict behaviour

#48

Post by dTal »

TameAntelope wrote: post #43
dTal wrote: post #39

Earlier quoted context omitted.

https://en.wikipedia.org/wiki/Argument_from_authority If Nature put out a paper saying the sky is red, it would be entirely rational of me to question what on Earth is going on at Nature, rather than my years of experience of blue skies.

https://en.wikipedia.org/wiki/False_equivalence They didn't put out a paper saying the sky is red, they put out a paper saying the magnitude of impact from breed on dog behavior is substantially less pronounced than previously thought. It is not rational for you to double down on this belief.

https://en.wikipedia.org/wiki/Motte-and-bailey_fallacy

The headline does not say "substantially less pronounced than previously thought". The headline says "breed does not predict behaviour (nature.com)". Any fool who works with dogs will tell you that yes, it ruddy well does.

Re: Study of pet dogs shows breed does not predict behaviour

#49

Post by daenz »

_aavaa_ wrote: post #24
onphonenow wrote: post #17

This is such a weird turn in new science. For hundreds of years folks breed dogs for various traits. Labs are friendly (and not great guard dogs). Greyhounds can run fast and like to chase. Pitbulls can use their strong bite and a shake behavior to devastating effect. And the list goes on. Now we have nature coming out with a study that shows breed does NOT predict behavior? What's interesting is I'm not sure some br…

I suspect part of it is a fear of the following slippery slope: admitting of genetic differences between breads -> genetic differences between races -> providing scientific proof that some races are better than others -> therefor some races should be treated better/worse than others. I personally find this line of argument flawed. Putting aside any discussion of the idea of race, it simply does not follow that a) one…

I can think of 4 things that would make a science-based discussion around those sensitive topics more palatable: 1) All data is in distributions, and so no predictions can be made about an individual based on distributions. 2) "Better" is a meaningless term in these discussions. 3) Any traits are the result of optimization for a particular environment. 4) Everyone should be treated with respect.

Re: Study of pet dogs shows breed does not predict behaviour

#50

Post by jamesash »

The Siberian silver fox study makes it very difficult to avoid the conclusion that behavior (e.g. tameness) in canines is a heritable trait that can be selected for.

Beginning in 1952 wild silver foxes have been bred in Siberia selecting solely for friendliness. Notable changes were found after 6 generations. By the 30th generation, 70-80% of the selected population were "domesticated elite", which "are eager to establish human contact, whimpering to attract attention and sniffing and licking experimenters like dogs." Similar experiments were performed in the opposite direction, selecting for aggression. https://en.wikipedia.org/wiki/Domesticated_silver_fox

Post reply on HN