Sunday, November 11, 2007

Weird

Can anyone get their head around this experiment conducted by a Canterbury University student?

The conclusion is dogs can read people's minds. But I cannot follow the methodology at all. Surely a smart dog would show a preference for the choice of the person they saw hiding something. How does it follow that the dog can read someone's mind? I am totally bamboozled.

6 comments:

Anonymous said...

Right between 62% and 70% of the time, huh?

Perhaps the dogs should play Blackjack, where the odds of winning are about the same.

Kevin said...

Its bollox. Its slightly amusing that a student can do a research degree by sitting around playing with a few dogs and self indulging in their hobby but then you realise that the taxpayer is paying for it and you begin to cry for our knowledge economy that will never be. Its the tip of the iceberg though - think bogans and sunscreen questionaires at the students favourite beach and research institutes in tin garages at Whanganui....

Anonymous said...

There is lots missing, such as what kind of food and how was it handled. It says one of two people put the food in a can but the dog couldn't see who. The dog went to the person who put the food in the can.

So why the surprise? If one person handles the food and the other doesn't the first will give off a distinct odor that dogs can pick up readily. Their hands will smell like food. The other one wouldn't.

Just shere random luck would give a 50/50 chance of being right. Toss in the smell factor and I would have thought the percentage would have been much higher.

But without proper controls and the methodology being explained there is nothing here but gullibile student and a media looking for a story whether there is anything of a worthwhile nature there or not.

PM of NZ said...

Methodology definitely in the same league of academe as giving $96K to study boganolgy...

Eric Crampton said...

The headlines on this story have been horribly inaccurate, and draw from an off-the-cuff remark by the researcher.

I've not read the full paper, nor am I likely to. The report I read on it is from the Chronicle, found here. The headline is rather bad, but the discussion of the method is a bit better than you'll find in the Press. Go to page 5.

From that story:

The study has nothing to do with mind-reading. Rather, it's checking whether dogs have a theory of mind. Rather a different thing, that. From the method described, dogs could follow cues from two people leading it to food; the dogs would more likely choose the person that they saw had been able to see where the food was put. Previous studies have found that primates have such a theory of mind but hadn't looked before at dogs.

I actually find the result pretty interesting. And, I'd be astonished if the methodology were as bad as anonymous suggests -- we've actually got a rather good psych department at Canterbury.

The Kiwi knee-jerk anti-intellectualism really depresses me. Go dig up the study and give a substantive critique if you've a mind to critique it. Don't guess what the methodology might be and then complain about your imagined version of it.

Eric Crampton said...

For clarity: I wasn't addressing the last comment to Lindsey, who properly was puzzled by the rather shoddy reporting on the topic, but rather at the other comments...