Behavior and Training · Science

Death Throes of the Guilty Look

I just talked to a potential client who is interested in bringing his 7-month-old Golden Doodle to train with us at AutumnGold. His dog, Penny, has the usual young dog issues – jumping up, a bit of nipping during play, still the occasional slip in house training, etc. Penny also raids the kitchen garbage bin, removing and shredding food wrappers, napkins, and any other paper goodies that she can find. The owner tells me that he is particularly upset about this last behavior because he is certain that Penny “knows she has done wrong“. He knows this because…….wait for it…….”Penny always looks guilty when he confronts her after the dreaded act”.

If I had a nickel………

Like many trainers, I repeatedly and often futilely it seems, explain to owners that what they are more likely witnessing in these circumstances is their dog communicating signs of appeasement, submission, or even fear.

Guilty Look Fear Beagle               Guilt or Fear Shelter Dog               GUILTY LOOK?                                                            OR FEAR?

And, also like many other trainers, I often feel as though I am beating my head against the proverbial wall.

But wait! Once again, science comes to our rescue! And this time, it is a darned good rescue indeed.

science to the rescue

The guilty look is a difficult issue to study because it requires that researchers identify and test all of the potential triggers that may elicit it, as well as the influence the owner’s behavior and his or her perceptions of their dog may have. Tricky stuff, but lucky for us, several teams of researchers have tackled this in recent years, using a series of cleverly designed experiments.

Is it scolding owners?  The first study, published in 2009, was designed to determine if dogs who show the “guilty look” (hereafter, the GL) are demonstrating contrition because they misbehaved or rather are reacting to their owner’s cues, having learned from previous experience that certain owner behaviors signal anger and predict impending punishment (1).  The study used a 2×2 factorial design, in which dogs were manipulated to either obey or disobey their owner’s command to not eat a desirable treat and owners, who were not present at the time, were informed either correctly or incorrectly of their dog’s behavior. The box below illustrates the four possible scenario combinations:

2 x 2 Factorial

The Study: Fourteen dogs were enrolled and all of the testing took place in the owner’s home. All of the owners had previously used “scolding” to punish their dogs in the past and an additional one in five also admitted that they used physical reprimands such as forced downs, spanking, or grabbing their dog’s scruff. In addition, all of the dogs were pre-tested to ensure that they had been trained to respond reliably to a “leave it” command and would refrain from eating a treat on the owner’s instruction. During each test scenario, the owner placed a treat on the ground, commanded the dog to “leave it”, and then left the room. While the owner was out of the room, the experimenter picked up the treat and either (1) gave the treat to the dog or (2) removed the treat. Upon returning to the room the owner was informed (correctly or incorrectly) about their dog’s behavior while they were away. Each dog was tested in all four possible combinations. (For a detailed explanation of these procedures and controls, see the complete paper). Test sessions were videotaped and dogs’ responses were analyzed for the presence/absence of behaviors that are associated with the GL in each of the four situations.

The Results: Two important results came from this study:

  1. Scolding by the owner was highly likely to cause a dog to exhibit a GL, regardless of whether or not the dog had eaten the treat in the owner’s absence.
  2. Dogs were not more likely to show a GL after having disobeyed their owner than when they had obeyed. In other words, having disobeyed their owner’s command was not the primary factor that predicted whether or not a dog showed a GL.

First nail in the coffin – The owner’s behavior can trigger the GL.

What about Dogs Who “Tell” On Themselves? But wait, Joe next door (who happens to know a lot about dogs) says – “How do you explain my dog Muffin who greets me at the door, groveling and showing a GL, before I even know that she has done something wrong?”

Not to worry – The scientists got this one too.

Back off man Scientist

The Study: Experimenters set up a series of scenarios involving 64 dog/owner pairs (2). The testing took place individually in a neutral room with just the dog, the owner and the researcher present. After acclimatizing to the room and meeting the experimenter, the dog was commanded by the owners to “leave” a piece of hot dog that was left sitting on a table. The owner then left the room. In this experimental design, the experimenters did not manipulate the dog’s response. Instead, they simply recorded whether the dog took the treat or not. Before calling the owner back into the room, the treat (if not eaten) was removed. The owners then returned to the room but were not informed about what their dog did (or did not) do in their absence. The owner then was asked to determine, by their dog’s behavior whether or not the dog had obeyed their command. In this way, the experimenters ingeniously tested for the “dog telling on himself” possibility.

Results: Just as the first study found, a dog’s behavior in the owner’s absence was not correlated with showing a GL upon the owner’s return. (Corroborating evidence from independent studies is always a good thing). The researchers also found that when they controlled for expectations, owners were unable to accurately determine whether or not their dog had disobeyed while they were out of the room, based only upon the dog’s greeting behavior. In other words, the claim that dogs tell on themselves and therefore must have an understanding that they had misbehaved was not supported.

Second nail in the coffin – Dogs don’t really tell on themselves (it’s an owner’s myth)

The most recent study, published in 2015, parsed out a final two factors that could be involved in the infamous GL – the presence of evidence as a trigger and guilt itself.

Guilty Look Rover

Guilt itself: If indeed, as many owners insist, a dog’s demonstration of the GL is based upon the dog having an understanding of the “wrongness” of an earlier action, then this would mean that the trigger for the GL would have to be directly linked to the dog’s actual commitment of the wrongful act, correct? Likewise, if the dog herself did not commit a misdeed, then she should not be feeling guilty and so should not demonstrate a GL to the owner.  It is also possible that the mere presence of evidence from a misdeed (for example, a dumped over garbage tin) could become a learned cue that predicts eventual punishment to the dog. In this case, a dog would be expected to show a GL in the presence of the evidence, regardless of whether or not he or she was personally responsible for it. This last study tested both of these factors (3). Using a similar procedure to those previously described, the researchers created scenarios in which dogs either did or did not eat a forbidden treat in their owner’s absence. They then either kept the evidence present or removed it prior to the owner’s return to the room. Owners were instructed to greet their dog in a friendly manner and to determine whether or not their dog had misbehaved based only upon their dog’s behavior.

Results: Owners were unable to accurately determine whether or not their dog had misbehaved based upon their dog’s greeting behavior and the dog’s actions did not increase or decrease the inclination to greet the owner showing a GL.  A dog’s inclination to demonstrate a GL was also not influenced one way or the other by the presence of evidence. The second finding suggests that the presence of evidence is not an important (learned) trigger for the GL in dogs. Rather the strongest factor that influences whether or not a dog exhibits a GL upon greeting appears to be the owner’s behavior.

Third and final nail – Neither engaging in a misdeed nor seeing evidence of a misdeed accurately predict whether or not a dog will show a GL.

The witch is dead

Take Away for Dog Folks:  This series of studies tells us that at least some dogs who show signs of appeasement, submission or fear (aka the GL) upon greeting their owners will do so regardless of whether or not they misbehaved in their owner’s absence. We also know that an owner’s behavior and use of scolding and reprimands are the most significant predictor of this type of greeting behavior in dogs. These results should really be the final death throes of belief in the GL. Good Riddance to it, I say.

Now, all that needs to be done is that trainers, behaviorists and dog professionals everywhere work  to educate and encourage all dog owners to PLEASE, PLEASE STOP DOING THIS:

Fear Guilty Look Husky             Fear not guilt

Cited Studies:

  1. Horowitz A. Disambiguating the “guilty look”: Salient prompts to a familiar dog behavior. Behavioural Processes 2009; 81:447-452.
  2. Hecht J, Miklosi A, Gacsi M. Behavioral assessment and owner perceptions of behaviors associated with guilty in dogs. Applied Animal Behavior Science 2012; 139:134-142.
  3. Ostojic L, Tkalcic M, Clayton N. ARe owners’ reports of their dogs’ “guilty look” influenced by the dogs’ action and evidence of the misdeed? Behavioural Processes 2015; 111:97-100.

Interested in Learning More about Canine Nutrition? Take a Look at the new Science Dog Courses!

Interested in more dog training and behavior myth-busting? See Linda Case’s book – “Dog Smart: Evidence-based Training with The Science Dog“!


42 thoughts on “Death Throes of the Guilty Look

  1. I agree with almost all of this and I’m glad you are educating people on this subject with scientific evidence. I don’t think leaving a treat on or off the table was a big enough trigger to use in the test regarding dogs telling on themselves I have a dog (a lab) that gets into the trash if I forget to put it behind a closed door before I leave. He does not even greet me at the door when he has done this. He is sitting in his “time out” spot when I come home. This happens to be in front of a window where I can see him as I walk up the sidewalk. I have no idea if he has misbehaved when I am walking up that walk because he is always looking out the window when I come home. When I get to the door and open it and he is not there wagging his tail that is my first clue that something is amiss. I believe the trash mess is his trigger for this behavior because my husband will send him there to lie down while he cleans up the mess. There was one instance where he had not been able to hold his bladder and relieved himself in the house. When I came home he did not greet me at the door and was in his time out spot. I immediately thought I forgot to put the trash away but no, that wasn’t it. I coaxed him out of his spot and tried to make him feel better. It was only about an hour later that I found out what he had done. . Granted, this dog could be an exception. He is a very smart dog! 🙂
    PS – I don’t use time out with the dog, my husband does. He thoroughly believes in the guilty look and often our sensitive Golden Retriever gets blamed for things because he thinks she looks guilty but I know that she looks like that whenever she thinks either of us is unhappy. This blog is getting forwarded to my husband right now!!


  2. Might be worthwhile to add to the 64 dog study by always having a treat on the plate when the owners returned.


  3. Great column; I’m looking forward to anything you find on dogs and self-awareness. The only study I know if is Marc Bekoff’s “yellow snow” studies. Would love to see MORE evidence, since it is clear that dogs absolutely are self-aware. Just need that science …


  4. Just a funny story on this topic: my 2 year old female, Maybe, had finally demonstrated she could be trusted in the house alone, but being that she has destroyed things (pillows, comforters, countless plush toys) in the past, still leery, i kept it to half days. I’d go home at lunch, let all three dogs out for ~30 minutes, then crate the puppies and go back to work. It went well for about 3 weeks, my female greeted me at the door each evening with what humans call “dog happiness.” The last time i left her alone in the house i was greeted by a sad looking girl. Ears back, tail tucked, avoiding eye contact. I walked into the family room and was so completely shocked!

    Maybe had shredded – absolutely destroyed a ~12’x14′ throw rug. which was a new thing – she’d never done it before, never seen me angry/stressed over a destroyed rug, and she greeted me with the “GL” before i’d even seen what she’d done. The mess she made killing the rug was huge – but the true shock was the brain behind the scene; Science will say she didn’t know right from wrong – based on the history of her behavior, etc, and i don’t disagree, but she did know well enough to tamper with the evidence.

    My 42 pound Maybe got a hold of the crate mat under 40 pound dog #3’s crate, dragged her across the laminate floor and positioned that crate right in the middle of where the rug once was. She literally framed the foster dog for her mess.

    I laughed so hard. it was really too funny to do anything other than laugh. thankfully she didn’t view my laughter as approval or permission to kill any other trespassing rugs. 🙂


    1. Hi “Fumfels” – My apologies for the delay in responding. I was out of town for a while and we recently brought a new pup home, so my computer time is very limited right now! 🙂 You are welcome to translate the article, but please provide complete attribution (my name as author) plus a link to the Science Dog site. Thanks and happy training!


  5. If I recall correctly, Temple Grandin wrote that in a two dog household where just one dog makes a mess, the dog that DIDN’T do it is usually the one to get blamed since he KNOWS the rules, anticipates the owner being upset and this is mistaken for a “guilty look”.


    1. I would like to believe that it was because the vast majority of owners already know that the look is anxiety rather than guilt, but suspect it is more down to the complexity and expense of setting up a study like these with larger numbers of participants…


  6. This was really brought home to me when a bottle of shampoo tipped over in my bathroom – by the time I found it there was a large, sticky mess all over the floor. The dogs peered in at the scene – grumpy owner on hands and knees, muttering while scrubbing a mess off the floor, and quickly and quietly both removed themselves from the vicinity and went to wait in the bedroom… It was absolutely nothing to do with them, but they knew that the same scenario usually involved mild dog-directed complaints (and stronger reminders to myself to take them out more often!).


  7. I figured this out years ago while I was baking cookies. The dogs were not allowed in the kitchen back then and I had everything laid out to start mixing. Had to head to the bathroom for a moment and when I returned there was a wrapper from a stick of margarine on the living room floor licked clean! I said,”Who did this”!!!’ Both looked guilty as heck. Next morning, Shannon, the older Border Collie, was laying on the back porch, stiff as a board. Rushed her to the Vet’s , she had a bad case of Pancreatitis, several days of IV’s ,she was ok. Learned several lessons that time! You yell, they all look guilty and no fats for dogs!


  8. There is a TV programm in Australia called Insight and it is basically a Discussion type programme that gets dfirreing demographics and experts to discuss a particular topic on that night.
    One night was wther or not dogs felt emotions etc. They had a human neuoscientist as one of the main guest who was doing studies with CT scans of dog’s brains and the way it reacted in a similar fashion to humans under certain stimulai.
    This “scientist” was convinced his dog felt guilt and he provided a video of an incident to prove. The dog had apparently stolen a couple of cooked chickens of the kitchen bench and the owner was scolding the dog (no physical punishment) and yelling for the dog to come and own up to what he had done. The dog slowly crept into the kicthen with all the usual appeasement signals etc – the vidoe went on for a long time and it really was heartbreaking to watch, although many in the audienced laughed and accepted it as the dog showing its guilt
    There were several vets in the audience, including Behavioural specialist, politly said they felt the dog was just repsonding to the owners tone adn tand not showing the GL
    “No way” said the “scientist” (notice I keep putting him in ” ” – you will see why shortly) “You are wrong and I can prove it with another story” and proceeded to tell us how one night when he was living in South Africa he heard a noise down stairs in the middle of the night that sounded like glass breaking. He fearfully crept down the stairs and was met by his dog…..(this is the clincher). Because they were both scared the “scientist” started to push his dog forward declraing what a good dog it was and it was big and brave and he said he was being really positive towards it. But the dog refused to come out from behind his legs and according to the owner had the GL. Once they arrived in the lounge room they discovered a lamp had crashed to the floor and smashed…..obviously (apparently) the dog had tipped the lamp over and broken it because the dog had behaved in a guilty fashion on the stairs – this was his proof, the dog knew it had done something wrong and was behaving guiltily …….he would not be swayed that maybe, just maybe the dog was just as frighted as it’s owner by the unexpected noise that happened in the middle of the night (possibly right next to it) ……..Yep true genius “scientist”


    1. Hi Gillian – Thanks for sharing this! It appears that no one (even a scientist!) is necessarily immune to the draw of the GL. What astounds me is that people think watching a dog in such distress is humorous – it breaks my heart to see that (as well as those awful “shaming” photos that are so popular today – I literally cannot stomach looking at those). Thanks for reading – Linda

      Liked by 1 person

  9. Your ability to take hard core science and share that information with us in an easy to understand and humorous manner is outstanding. I promise to never say out loud that Sally has a GL! Thanks for the insight.


  10. I had a feeling this wasn’t a thing. Will need to show my folks how erroneous the guilt-training style is.

    Next to debunk: Dogs do not have self-awareness like gorillas, dolphins, elephants, or people. Anyone care to comment?


Have a comment? Feel welcome to participate!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s