Words of One Syllable department

Short answer: No. No, it doesn't. That's "it," as in no, your dog probably doesn't feel guilty, and no, that's not what the study says (which is "no," or, more broadly, "no such result").

What we have here isn't really a study of "animal behavior" but a study of human behavior -- specifically, anthropomorphism, or attributing human-like traits to something nonhuman. You know, like looking at the "front page" of a "major metropolitan daily" and assuming that what you see reflects "news judgment." And by all appearances, it's a pretty good study. But it's not on the front page because Austin has a regular process for vetting the week's peer-reviewed research (cognitive psychology, linguistics, epidemiology or whatever your favorite discipline), then giving the best play to the most relevant and rigorously conducted studies. It's there -- three weeks after it was promoted in a press release from Elsevier, publisher of the journal (Behavioural Processes) in whose special dog issue the study appeared, one might point out -- for, well, behavioral reasons.

The story isn't new by the time it reaches us today. It was reported that very week by the New York Times, Fox, the Beeb and UPI, in Mexico and Canada ("Pas de culpabilite pour Fido"), and in the Netherlands, Norway and Brazil.* But it did get a new boost by appearing on the front of the Washington Post this week, and those 1A budgets from the big-town papers do have an agenda-setting effect out in the sticks.

It needs to be noted that we're not talking about a bad study here (and since it isn't about using gummint money to study foreigners' sex habits, even Fox seems to like it). It's a neat design, and the researcher is appropriately cautious in talking about what it does and doesn't do.** (You'll note that she doesn't claim to prove dogs don't feel guilty, just to have shown that what's seen as "guilt" isn't associated with with the dogs themselves did.) But it gets a little boost from the press release, which calls the research design ingenuious, and another from the Post, which personalizes it with a dose of Elongated Yellow Fruit syndrome:

Many dog owners have had this experience: Arriving home, they discover their pooch looking sheepish, with its head down, ears pulled back, tail tucked between the legs, maybe slinking behind the sofa. Puzzled, they soon discover the reason: a favorite pair of shoes chewed to pieces, or perhaps the kitchen garbage can upended.

But is their canine companion really acting guilty? Or is this an example of people projecting a human emotion onto their four-legged friend?

And when it gets to Austin, it gets that Local Anchorperson touch: "Up next, the weather. And does your dog really feel guilty? We'll find out!" It's what you're told to do -- almost literally -- in journalism textbooks, which remind you that data don't mean anything unless you relate them to people and that scientists can't be trusted to describe their findings on their own.

Too bad, because the Elsevier folks do a reasonably good job of explaining it:

This study sheds new light on the natural human tendency to interpret animal behavior in human terms. Anthropomorphisms compare animal behavior to human behavior, and if there is some superficial similarity, then the animal behavior will be interpreted in the same terms as superficially similar human actions. This can include the attribution of higher-order emotions such as guilt or remorse to the animal.

The editor of the special issue, Clive D.L. Wynne of the Department of Psychology, University of Florida, explained, “this is a remarkably powerful demonstration of the need for careful experimental designs if we are to understand the human-dog relationship and not just reify our natural prejudices about animal behavior.”

Interesting, huh? But you can be forgiven for assuming from the headline that those aren't the reasons it was interesting at the 1A budget meeting.

Once again, not a case of Bad Science, and not an especially bad instance of journalism either. Just a reminder, perhaps, that dogs aren't the only ones who draw strange inferences about the interests of human audiences.

* In case you're practicing, the Portuguese for "study says" appears to be "diz estudo." One wonders how often balloon photos say "Up, up and away."
** And, of course, anyone who titles a canine cognition book "Inside of a Dog" is automatically invited to the next Marxist Media Studies Conference at Stately HEADSUP-L Manor.***
*** I shot a critical theorist in my pajamas last night.