Tag Archives: training

Reading research – 8 classic red flags

Posted on by
11

Ten years ago, few trainers had access to research studies.  These days with Google University, we have moved into the era of research wars.  It is a battle of quantity, novelty and link littering.  Unfortunately, few seem to be reading past the abstract soundbites to see if the study in question is any good.  Even more problematic are lure of pop psychology magazines with sexy titles, articles that probably misinform more than educate.

Every professor and textbook on the subject of research sends a consistent message.  Read research with a critical mind.  Not all studies are well executed.  Peer reviewed academic papers are no exception.  Sometimes journals will publish research with poor design to inspire further research.  Looking at study design is scientific, not sacrilegious.

As readers of studies, we can take steps to improve our research reading abilities.  We can face our own biases directly.    Do we militantly tear apart research that goes against our point of view, while offering leniency to findings that feel warm and fuzzy?  More importantly, do we know how to read an analyze research?

I won’t pretend to be a research expert.  Rather, over the next series of blogs I will be highlighting what I have learned from Experimental Psychology by Myers and Hansen.  It is a worthwhile investment for anyone wanting to flex his or her mental muscles.

One core concept of research is internal validity.  As the name implies, we need to assess if a study is valid on the inside.  External validity, by contrast, would look at whether results apply to “the real world.”  Lesson number one is that internal validity should not be sacrificed for external validity.  If a study is not valid on the inside, there is nothing of substance to apply to the “real world.”

Campbell identified eight “Classic Threats to Internal Validity.”  They apply to research involving experimentation.  This includes true experiments and quasi-experiments.  True experiments have strict parameters or rules.  Quasi-experiments are not bad, just different.  In both researchers manipulate a variable and then measure the effect it has on another variable.

IV isolation

For example, we might want to know if training method A is faster than training method B.  We divide a number of dogs into two groups and compare results of those two methods.  The type of training is the variable being manipulated.  We call this the independent variable (IV).  The goal of experimentation is to isolate the independent variable, to ensure that no other factor is interfering or confounding the results.

Revisiting our dog-training example, let’s say that group A tested on a Monday and group B tested on Tuesday.  If Monday was sunny and Tuesday was stormy, any claim that treatment A was better is highly suspect.  Stormy weather could have agitated the dogs in group B.  The independent variable was not adequately isolated.  The study would not have internal validity.

The following itemizes Campbell’s Classic Threats to Internal Validity and provides examples.  One step we can take toward understanding research is to understand how these threaten validity.

Our thunderstorm example above is a history threat.  Dogs in group B had a shared history during the experiment that differed from dogs in group A.  Training methods varied.  However, so did weather.  No one can say for sure which training method was faster because the weather interfered.  History threats can be subtle.  Another example would be if one group receives an orientation while the other does not.  Orientation can prime one group, giving them a head start.  It would also be a history threat.

Maturation threat reflects internal changes.  An obvious example might be age.  Behaviour can change as puppies mature.  Maturation can also mean the maturation of knowledge.  Students handling dogs during experiments will have gained knowledge throughout the term. It would not be wise to test group A with new students and group B at the end of term.  Increased knowledge by the end of term can mean that students guess the hypothesis or influence results.

Subjects rarely get the same test results when re-tested.  Practice leads to improvement, even without treatment of any kind.  Suppose we take a group of anxious dogs and test their heart rate.  Heart rates can drop simply because the dog habituates and becomes more comfortable.  A second round of testing should show habituation.  It is not enough to ask if a dog improved, we need to know if the dog improved more so than dogs that did not receive any treatment.  Otherwise, we have a testing threat.

Measuring results is not without potential pitfalls.  Instrumentation threats involve data collection problems.  Equipment can fail or be unreliable.  Scales used to score results need to be set correctly.  Assume we want to know if dogs become anxious at the dog park.  Imagine if the measurement options are:  highly anxious; moderately anxious; mildly anxious and fully relaxed.  Answers obviously will be weighted toward the “anxious” side of the scale.  Unless a dog is fully relaxed, it is by default labelled as anxious. Had moderately relaxed and slightly relaxed been offered as choices, an entirely different picture may have emerged.

Random selection between groups is important.  This process helps balance the characteristics between groups.  When groups are not random by design or chance, this is a selection threat.  Assume that wrandom balancinge want to know which training technique obtains a faster recall.  Group A dogs are mostly short hounds and toy breeds.  Group B has mostly large dogs with a smattering of Border Collies and Whippets.  Under those conditions, we could not claim that group B training produced faster recalls.  To avoid accidental selection threats, random selection and balancing offers an even comparison between groups.  Researcher choice is not random.

Mortality should be listed in an experiment.  It is the dropout rate.  When large numbers drop out of an experiment, it’s a red flag.  According to the text, “Often it means that the treatment is frightening, painful, or distressing.  If the dropout rate is very high, it can mean that the treatment is sufficiently obnoxious that typical subjects would choose to leave, and the ones who remain could be unusual in some respect.”  Assume we are testing a protocol to help reactive dogs.  Many drop out.  Those who remain seem to improve.  The obvious question is whether those who left were distressed or deteriorated so much so they did not return.  That is critical information.

The seventh threat comes with a big word:  Statistical Regression.  Extreme test results are unreliable.  Think back to grade school I.Q. tests.  Scoring low could mean you had the flu.  If an experiment uses subjects with extreme characteristics, we can expect some of that to level out on its own.  Testing a new anxiety treatment on highly anxious dogs can appear to work.  That result looks similar to statistical regression.  As with a testing threat, it is not enough to ask if an animal improved.  We need to ensure that improvement happened because of the treatment.

Finally, we come to selection interaction threats.  It’s the one-two punch of threats.  It happens when a selection threat combines with another threat.  Returning to our experiment that asks which dog training method is faster, suppose we non-randomly select dogs from two training schools.  Immediately, that creates a selection threat.  Now suppose school A has a trick team.  Students at this school are motivated to join the team.  The second training school does not offer tricks sessions.  That creates a history threat.  Trick dogs would have a wide array of skills to draw on – to guess the right answer instead of learning it via the training method tested.  Selection threat combines in this case with a history threat to create one hot mess.

classic threats to validity copy

Campbell’s Classic Threats are the tip of the iceberg in terms of red flags.  It can make it seem no research can hold up to its standard.  Following a defined process for evaluating research is a far sight better than pointing to number of subjects and chanting “the n value is too low.”  It may not be possible to control for every bump in the road.  Experimental Psychology states, “Control as many variables as possible.  If a variable can be held constant in an experiment, it makes sense to do so even when any impact of that variable on the results may be doubtful.”

Knowing the threats to internal validity are only useful if you start using them to read studies more carefully.  It might be tempting to annihilate an experiment you dislike.  Perhaps a more interesting exercise would be review an experiment you love and have shared.  Challenge your bias.  Look at the design and the various threats to internal validity.  Did you find any?

(Part 2, 3 and more….about those n values, non-experimental research and more.)

 

 

Shock Collars are so NOT like T.E.N.S.

Posted on by
3

I had a more practical blog lined up.  Humor me, as we look at one more expression.  This one keeps coming across my news feed and it’s interesting.  Some trainers claim that, “Shock collars are like T.E.N.S. machines.”

I have used a T.E.N.S. machine many times under the guidance of various therapists.  For those unfamiliar with this treatment, mild electric current flows to pads secured to the patient’s skin.  It causes muscles to contract.  Dr. Ho’s Pain Therapy System seen on infomercials is a T.E.N.S. machine.

Here is one of the machines that I use on a regular basis.  At the bottom are two dials for “intensity.”  There are two intensity dials because the machine regulates the left and right pads separately.

tens

Comfort levels change between individuals.  They also change from moment to moment for the same person.  Intensity levels vary from one area of the body to another, hence the two dials.  There is no “consistent” setting.

If you asked a patient using a T.E.N.S. “Are you okay?” the answer should be “Yes.”  Discomfort results in a reduction in intensity.  If asked, “Do you want me to turn it off?” the answer should be, “No.”  Stimulation should be neutral or pleasant.

With shock collars, it is this underlying premise that is fundamentally different.  Bark collars shock dogs to suppress barking.  Barking reduces because the dog does not like the shock.  If the shock were neutral, nothing would change.  If it were pleasurable like a T.E.N.S. machine, barking would increase.  Barking would increase because the dog would be learning that barking leads to massages.

The same goes for electronic training collars used as negative reinforcement.  Negative reinforcement is like the game “Uncle.”  Children pinch one another.  They keep pinching until the other child says, “Uncle.”  The behaviour of saying uncle leads to cessation of pinching or relief.  Similarly, dogs turn the shock off by obeying a command.  Reinforcement is the cessation of the electronic stimulation – relief.  Dogs that respond quickly can “beat the buzzer”, avoiding shock altogether.

“Pleasant” shock, under the level of aversiveness wouldn’t be effective here either.  Ask a dog, “Do you want the electrical current to stop?”  The answer needs to be, “yes” or the collar will not work as designed.

Both may involve electrical current but the intensity is on a gradient.  Many things are on gradients.  Music is nice, unless it is too loud.  Massage feels good, unless Olga the Horrible is hurting you.  (My apologies to anyone named Olga.  Olga was the name of the massage therapist who hurt me.)  Light helps us see unless it is glaring.  Flowers smell pretty unless it is an overdose of noxious perfume.  Cool water makes a refreshing swim.  Cold water is icy and painful.  Food is usually fabulous, unless you have eaten excessively and feel nauseated.

No one thing fits neatly into a naughty or nice category.  Dogs decide what they find pleasant, neutral and aversive.  We infer how the dog feels by observing their change in behaviour.

I would have gladly paid Olga the Horrible to stop the massage.  Massage therapist Mike was different.  He had magic hands.  I would pay money for longer massages.  Olga’s massages were aversive.  Brad’s were appetitive.  I was not screaming and thrashing in pain with Olga.  I just wanted it to stop and it changed my behaviour.  Contrary to popular belief, it’s not the dog’s attitude that will tell you if the training is aversive.  It is the dog’s responses.

There is a second key difference between these two devices.  Shock collars are training.  With T.E.N.S., the goal is long-term if not permanent pain elimination.  It is not a teaching tool.

Olga may have been brutal, but her intent was to offer long-term relief.  Magic Hands Mike offered the same goal under threshold.  Suzy who does scalp massages at the hair salon….she just gives nice feeling massages.  Training cannot be compared to these scenarios.

In training, the trainer creates the plan with the goal of creating a change in behaviour.  We determine which behaviour we want to teach.  Therapies for pain relief have no skills development.  Olga never taught me to dance a jig in return for pain relief.  She did not use my discomfort as a tool.

I never went back to Olga.  I had the choice to leave.  More importantly, I had a choice to never return.  I could refuse to participate.  Our dogs do not have that choice during training because repetition is part of the gig.  Once the aversive ceases, the aversive is back in play for more training.  We can’t see all the differences unless we look at the specifics of how each of these work.

Olga vs Mike

If you are ready to jump in and say, “You don’t need batteries to use aversives,” then you won’t get any argument from me.  There are many forms of aversives in dog training.  Some are easier to swallow than others.  You do not need batteries to use aversives.

But, let me broaden the statement.  Using aversives in dog training is not the same as using a T.E.N.S. machine.  It’s not even close.  It does not matter if the aversive is added to the dog, the dog is added to an aversive scenario or whether we use a dog’s discomfort for our training purposes.  Using the aversive is trainer’s choice.  If the dog is put into a position where they want to leave, they have been put between a rock and a hard place.  It’s not free will.  The trainer has free will.

Training paradigms are not the same as T.E.N.S. machines.  In dog training, the choices to use an aversive are ours and ours alone.

Stop Nagging – Start Training – Using the environment as a command

Posted on by
2

I cannot stand micromanaging my dog’s behaviour.  There is a certain point when I expect my dogs to behave.  I hate repeating myself, and I really hate sounding like a nag.

It’s exhausting.

When we think of obedience, the command/obey sequence is what typically comes to mind.  It is what most dog owners first learn.  Say sit, and when the dog sits, reward the dog.  Continue practicing until the dog sits when told to do so.  I have no objection to this.  Being able to communicate with your dog is very useful.  Stylized commands are extremely useful in competitive environments.

Constantly reminding the dog of expectations can get old fast.  Some expectations never change which means it would be nice if the dog did them without having to give a command.  For example, I never want the dog eating the cat’s food.  My dinner is always off limits.  Jumping on visitors is never okay.  Darting out the car door without permission is downright dangerous.  Really, I am never going to want any of these things.

I don’t want to give a command.  I don’t want to give reminders.  Dear dog:  Just do what you are supposed to do.

It’s a lot like little children.  We expect that young children need reminders to “flush the toilet.”  However, there comes a point in time when seriously – you just should not need to be reminded to flush.

If that sounds demanding, it is.  I demand it of myself to train my dog to understand that certain behaviours are expected at all times.  In order to achieve this, we need to expand the idea of what constitutes a command.

stop nagging

Traditionally, commands are words or hand signals that tell the dog what to do.  Who says that commands have to be limited to words or stylized gestures?  Actions and situations can act as commands too.

Dogs readily learn that our actions mean something.  We pick up a leash and they run to the door in anticipation of a walk.  You don’t have to say a word to speak volumes to your dog.

You can use this ability to intentionally to create “commands” for your dog.  If the context or situation reminds the dog of what it should be doing, you no longer have to.

Environmental commands or contextual cues have a wide array of uses.  Teach a dog to sit for visitors, without being told to do so.  Have them wait until released from the car.  For this blog, let’s work through “leave the cat’s food alone”.

Start by ensuring your dog has a reasonable grasp at leave it.  I teach it by reinforcing movement away from treats as shown in this video that demonstrates with American Sign Language.  A verbal cue is taught in the same manner.

Next, bring out the cat’s food.  Continue working leave it, giving the command and reinforcing the dog when it is correct.

At some point, the dog will jump the gun, offering the leave it BEFORE you ask for it.  I call this a genius moment.  Your smart dog has decided to leave cat food alone without being asked

Celebrate this.  Reinforce it – generously.  Give treats.  The dog has noticed that the cat dish being placed on the ground predicts the leave it command.  With repetition, placing the food on the ground will become a command of its own.

Cat food in the bowl = leave it.

The finished product makes for a very peaceful co-existence with our dogs.  The following video shows Kip, Karma and Icarus demonstrating a few variations.  You will see both dogs leaving the cat food alone.  In addition, you will see the animals leaving treats thrown to the others.  I do not like my guys charging and battling over food.  I never want my dogs charging at dropped or thrown food unless given permission to do so.  It’s an excellent candidate for a context based command.

As you watch the video, notice the following key points.

  • There are no verbal commands or reminders given.
  • The dogs are free to do any appropriate behaviour they like.  This is not a stay.
  • Notice the calm and disinterest.  It comes from consistency and generous positive reinforcement.

Most importantly, listen to the silence.  Our dogs are capable of learning that situations have meaning.  Our training has to come up to their abilities.  Watch for those moments when dogs offer genius.  They might be fleeting at first.  With reinforcement, they can blossom into so much more.

(I have used the word command throughout this piece for readability.  Many pet owners recognize it.  Trainers often use the word cue instead because it implies communication while the word command can feel like an order.  By command I really only mean a word, gesture or situation that communicates information to the dog – an antecedent.)

Nine Steps to a Calm, Relaxed, Quiet Canine. Have a Go at DRO.

Posted on by
11

I like quiet, calm dogs.  Most owners I know want quiet, calm dogs.  Visions of a dog cuddling during a relaxing evening shatter with the reality of a pushy, loud, obnoxious pest.  Where classes are concerned, pestering pups embarrass owners.  I like my classes quiet so I don’t have to yell over the ear piercing, migraine inducing screaming of an out of control, demanding dog.

Typically, owners are offered a long list of tips ranging from increased exercise, busy toys, supplements, gadgets and massages.  Sufficiently exercised, some dogs fall asleep.  A sleeping dog isn’t well mannered.  It’s just sleeping.  When they wake, we’re back to pestering.

Owners can feel like they have become the dog’s personal entertainment center. Perhaps we should use training instead? Many owners try rewarding an incompatible or alternate behaviour.

For example, the dog that is lying on a mat cannot be clawing at your legs.  However, I like to use something a little different because DRIs and DRAs can create behaviour chains.  The dog swings back and forth between good and obnoxious behaviour like a pendulum on a clock.  Pester – treat – pester – treat – pester – treat.

When I get noisy, hyperactive, unfocused, pushy, pestering dogs in classes, I start them on a Differential Reinforcement of Other Behaviour (DRO) plan.  It uses positive reinforcement to reduce unwanted behaviours.  The absence of a specific problem behaviour over time is reinforced. DRO’s work on a wide array of problems from barking, pawing, pushing, bouncing at the end of a leash and more.  I do not however use it with dogs with fear and anxiety problems that would be better served through desensitization and counterconditioning.

What many fail to realize is that a DRO rewards TIME, not RESPONSES. Simply reinforcing the absence of a problem doesn’t address that we want the dog to behave for a prolonged period of time.

Let’s use barking as an example.  We do not want to create a dog that barks and hushes, looking for a treat each time it quiets.  We want a dog that stays quiet.  Our criteria is not the absence of problem behaviour.  Our criteria is the absence of problem behaviour over time.  Dogs can absolutely learn time based or temporal criteria.

Back to our barking dog, we might begin by rewarding short periods of quiet.  It does not matter if the dog is sitting, standing, spinning, chewing a bone or doing a headstand.  If the dog remains quiet for a pre-determined length of time, they earn their reward. Gradually, we would ask the dog to stay quiet for longer and longer until barking rarely, if ever, happens.

Close attention to passing time prevents the dog from learning that they can bark and hush to get a cookie.  That’s because we are being very clear that our criteria is not the act of becoming quiet, but the act of staying quiet.  The dog is free to do any safe, appropriate behaviour it likes – so long as it is quiet.  The same goes for any other nuisance behaviour we seek to eliminate. dro A well executed DRO follows a process and some rules. There are variations based on the type and timing of reinforcements.  This is the one I typically use in classes and at home with my dogs.

The Process

Step one:  Identify the problem behaviour with clarity.  (My dog barks when near other dogs.)

Step two:  Measure the frequency of the problem to create a baseline.  How often is the dog barking?  (My dog barks when near other dogs on average every 2 seconds.)

Step three: Set the length of time your dog needs to “behave” in order to earn a reward.  This should be slightly less than your baseline.  (I will reward my dog every time he is quiet for 1 second.)

Step four:  Use a reward that motivates the dog.  (My dog likes meat.)

Step five:  Allow the dog to engage in normal activities that are safe and appropriate for the context of the situation.  You are not asking the dog to do anything.  Count quietly in your head and reward the dog each time they meet your criteria.  (One Mississippii – treat.  One Mississippi – treat.  One Mississippi – treat.)

Step six:  Aim for A level student grades.  The dog should be right at least 80% of the time.  We want the dog practicing appropriate behaviour, not rehearsing the problem.

Step seven:  Continue to re-evaluate progress and measuring the dog’s responses.  Ask for longer duration of appropriate behaviour as the dog demonstrates that they are ready.  Stay at each level until the dog is consistent.

Step eight:  Switch to a random schedule when the dog has developed sufficient duration.

Step Nine::  Generalize the behaviour in various locations as needed.

Now for the rules:

  • Count!  Do not eyeball this exercise.
  • As your dog improves, you can increase criteria more rapidly.
  • If the dog misbehaves before time is up, get their attention and re-start the time.  (Do not reward the attention back to you.)
  • If the dog is not hitting A student level, decrease your expectations.  Make it easier.
  • Be careful not to reinforce other nuisance behaviours.  For example, if you want to eliminate barking, be careful that you are not rewarding pawing or jumping.  If the dog engages in too many alternate problem behaivours, reduce your time criteria.
  • Problem behaviours may initially increase before dropping off dramatically.
  • During initial stages,ignore other training goals.  Once problem behaviour disappears, you’ll have plenty of time to teach new things.
  • If the problem fails to improve, communicate with a qualified pet professional.  It is possible that the dog is misbehaving due to medical problems or anxiety issues.

Perhaps one of my favourite parts of executing a DRO in group class is watching the dogs.   Young, powerful, adolescent dogs barking en masse and too distracted to learn anything start to breathe.  Their muscles relax and the room goes still.

The owners are stunned that for the first time ever , their hyper dog has gone from a maniac to sane – a cool dude lying by their feet.  It happens in a class of other dogs.

A DRO is how we taught Kipper the ex-crotch ripper to settle quietly at the end of a busy day. Sometimes we get so focused on telling the dog what to do, we become micro-managers instead of teachers.  Sometimes we forget that we often don’t actually care what a dog is doing, so long as it’s appropriate.

Often we forget about counting – the time factor.  It’s a shame we don’t focus on time more often because frankly – it’s really effective.  Results are often dramatic and can come quickly.  If you want to eliminate any number of nuisance behaviours, remember to keep one eye on a clock.

Science says … a lot of things.

Posted on by
20

There seems to be an explosion of science circulating through dog training groups, and that is rather exciting.  I started collecting studies over a decade ago.  I am a huge fan of science and seem to have gotten a reputation as a go to person for links.  Often I receive messages that say:

“Do you have anything that proves that…..<insert topic here>.”

The truth is that you could insert almost any topic and I probably have something.  Heck, I could send you study links to “prove” that aliens exist.

Where dog studies are concerned, I have studies that show negative reinforcement is linked to stress.  However, I also have studies that show no increase in cortisol – a stress hormone – in dogs trained with negative reinforcement.  Pick any topic and there will likely be studies that draw very different results.

I can “prove” both sides.

Science is a lot like Lego.  Each block is important.  You can’t see the whole of the structure by looking at just one block.  Research studies are the pieces.  Together they give you a complete structure.  As you stand back, you might see that some don’t fit quite right.  Some pieces for whatever reason don’t work in a given spot.  Sometimes you might even get one of those cheap knock off bricks that doesn’t fit anywhere at all, except perhaps the trash bin.
IMG_8216 copy

There are knock off studies – pay to publish.  Money talks, sometimes a little too much.  Not all research is free from the dilemma of who pays and why.  We live in an era where corporations can hire researchers to “test” their products.  How biased those studies are depends on the construction of the study.

Other times statistics pose problems.  Small or pre-screened samples create a huge margin of error.  Who cares if one or two pre-screened dogs act a certain way?  One proverbial guinea pig is not a large scale study with blind controls and random assignment to groups.  That is rather important if you want to know how the average dog behaves.  That is not to say that small studies are bad.  It is what it is.

Scientists question studies, trying to replicate interesting findings.  If only one research team is getting a particular set of results, we should probably ask why.  It’s not personal, nor is it an insult.  Questions are good.  Researchers do it all the time.

Let’s not forget that mistakes can happen.  Media outlets reported that neutrino particles moved faster than the speed of light (apparently an amazing physics discovery).  Testing and re-testing confirmed the results.  Yet, other scientists kept digging into the controversial finding.  Eventually it was determined that a loose cable caused faulty results.  In the age of the internet, you can still Google the obsolete (2011) results.  Quote it all you like, it’s wrong.

Questioning research doesn’t make one a jealous Debbie downer.  The scientific process is all about throwing stones.

The question is whether we allow our own opinions and bias to determine which studies we blindly accept, and which we evaluate with a critical eye.  Searching for studies is not the same as searching for truth.  Validating our own choices, or heaven forbid our own business product is biased and self serving.

Don’t get me wrong, Google Scholar has a place.  But it’s not really a place where we should hunt down support for our own opinions.  “I knew I was right, I found an obscure abstract, skipped over the flaws and quoted one paragraph that proves my point.”  Of course, no one ever phrases their findings using those words – making the practice difficult to spot.

Instead, we should be looking at all studies with the goal of ascertaining truth.  If we have made an error in our thinking, we can seek to correct it, or wear blinders, plugging our fingers in our ears chanting “na na na na I can’t hear you.”

Our dogs deserve better from us.  They deserve us to care enough that we look for truth.  By doing so, we can see that conflicting studies just give different representations of information that might oddly fit together.

Returning to the negative reinforcement example from the beginning, there really is no controversy.  We know that successful avoidance of aversives can provide temporary stress reduction.  Both outcomes are possibly true under different scenarios.  Conflicting results can support one another.

It’s like saying that I don’t fear spiders in my home because I bug bomb regularly.  Does that make bug bombing is a good strategy in treating spider phobias?  No.  It can reduce my stress levels inside my home until it is time to spray again.  Conflicting results have surprisingly logical explanations.

The goal should be to keep asking questions, discarding pseudo science and disproven theories.  We should aim for the very elusive goal of seeking truth – ever mindful that we all carry a bias.  The antidote to that bias is to kick the studies that appeal to us with just as much ferocity as those that offend us.

That goes double for studies quoted by other people.  Read the studies.  Read the opposing points of view.  Simple truth: If you believe everything that comes with a link, you’re letting other people do your reading and thinking for you.  That is an idea that I just find, unthinkable.

Feb 21st:  Great blog done by another great writer – Eileen – on the subject.  Has some excellent links on assessing the qualify of journals.

Feed Problem Dogs Last

Posted on by
2

Kip, Karma and Icarus sleep on my bed.  Although, Icarus, the cat, prefers to use Karma’s tail as his own personal flirt pole rather than sleeping.

When Kip first came to us, he preferred to sleep under the bed.  It took some time using coaxing and treats to convince him that it was safe to sleep in the open.  When puppy Karma earned the right to come onto the bed, I was sensitive to the fact that Kip likes his space.  If others are too close, he retreats to the floor.

I brought treats with me to bed, ensuring that I quickly set some ground rules.  “You lie down here and you lie down there.”  Both were close to one another, but both had their assigned areas.

I had anticipated that Kip would be uneasy.  Surprisingly, Karma snapped at Kip.  “Back off – my space- my treats – my mom.”  She is a spunky little gal.  While mild resource guarding might be natural, I do not like it.  I do not want it in my house.  I do not want it when I’m around other dogs either.  Mild resource guarding does not scare me, but I am on it immediately.  Therefore, I made the decision to follow the following rule:

“Karma, you get your treat last.”

Many trainers claim that problem dogs are trying to assert themselves, trying to become the alpha.  By feeding them last, you are driving home the point that the dog is at the bottom of the pack, the omega.  “I am alpha, you are not.  Knock it off.  Alphas eat first and you need to be the alpha.  Feed the dominant ones first, and you cannot let it be the problem dog.”

Trainers and owners swear that this strategy works.  I have no doubt that it does.  After all, I just said that I implemented it.

A treat for Kip, then a treat for Karma.
A treat for Icarus, then a treat for Karma.
A treat for Icarus, a treat for Kip and then a treat for Karma.

Not only did we do this as we prepared to go to sleep.  Karma had snapped at a couple potential dog friends.

A treat for the strange dog, then a treat for Karma.

Initially I would have one dog to the left of me.  Karma was further off to the right.  The distance between my outstretched arms increased the distance between the animals.  With time, I allowed the other dogs closer and then initiated a pause.

A treat for Kip, followed by a pause, then a treat for Karma.

By pause, I mean the briefest of moments.  It feels like you have taken a short breath and are still waiting to exhale.  With time, those pauses grew longer.  “Please learn to patiently wait for your treat.”

Finally, we added the element of motion.  If I tossed a treat to another animal and it went astray, I wanted Karma to back away instead of fighting over it.

Throw a treat to Icarus, watch the cat bat it around like prey, then a treat for Karma.

Execution matters.  We worked in careful measured steps.  This is an overview, not a how to.  Had these been adult dogs, I would have put safety precautions in place.  I hope that the gist is clear.

Karma eats last.  Proponents of dominance theory use the same exercise.  The problem dog eats last.  Yes, it can work.  I do not think it has anything to do with dominance.

There is this principle in science called Occam’s razor.  It states that if there are two hypotheses, the one with the least assumptions is likely true.

We can assume that dogs act like wolves.  We can assume that the desire to be the alpha dog motivates them to act in dangerous and destructive ways, even if it seems counterproductive and unstable.  Our assumptions can extend to the idea that dogs keep a tally based on when they eat in relation to others.  We can leap to the conclusion that dogs are too dense to realize we belong to another species.  We can believe that we can integrate ourselves into this battle of social rank and that we can influence the dogs.  We can ignore the caveat that dominance is about relationships between conspecifics – members of the same species.

Would it not make more sense to say that dogs make associations?  Pavlov rang a bell and the dogs salivated.  If Kip gets a treat, then Karma gets a treat.  I am creating an association.  See a dog get a cookie and salivate.

With repetition, the dog’s internal emotional state changes so it no longer wants to drive away other animals.  It wants those dogs closer.  “Please, come closer.  If you get a cookie, I get one too.  I’m drooling in anticipation of your presence.” 

Karma and Ic napping

Feeding the problem dog last can be part of an effective strategy for many dogs.  However, the fact that it works doesn’t prove that dominance played any role in the problem.   Similarly, just because some trainers justify the strategy with dominance based explanations, it does not make the technique flawed.

Occam’s razor is probably right.  It usually is.  Classical conditioning as an explanation has an additional benefit.  Labeling our dogs as dominant, painting them as creatures set on usurping our authority is combative.  These magnificent creatures share our homes and our hearts and deserve better than a negative bias based on assumptions.  Ulterior motives for misbehaviour justify anger, frustration and punitive measures.

If I have two competing hypothesis, I choose to go with the one that paints the dog in the more flattering light.  If I feed, pet, play or give attention to a dog last, I’m very likely creating an association.  We can break down all protocols into basics confines of learning theory.  We really don’t need an explanation that is more convoluted than that.

Treat Training Trinity – Why positive reinforcement did not work for my dog.

Posted on by
105

About fifteen years ago, when I started apprenticing as a trainer, I used leash corrections and other forms of “discipline”.  I no longer leash correct, and have not for more than a decade.  This is not because I had a moral agenda.  I simply needed an effective training solution.

Kiki, my learning dog pulled like a tugboat.  We tried so many techniques we could have been the poster child for:

“But I tried positive reinforcement and it did not work.”

I chuckled and snickered with other trainers, “Ignore bad behaviour?  So you just LET the dog knock grandma to the ground?”  Teaching with food and then proofing with corrections seemed to make more sense.

We ran the gamut on protocols:

Food luring Collar corrections – flat collar
Collar corrections – nylon slip Head halter use
Head halter to reposition dog Head halter corrections
Chain choke collar correction Special choke collar correction (sits behind the dog’s ears)
Clicker instead of verbal marker Penalty yards
Be a tree Change of direction
Reward dog for releasing leash pressure Reward with approach to distractions
Hiding food rewards/surprise Finger poking the dog

 

Failure was not from a lack of effort or poor timing.  That special choke collar came from a very large, well-known training facility.  Ironically, they market themselves as being positive.

I have heard people say that positive reinforcement did not work for their pets.  Upon further digging, their chart looks oddly similar to mine. This is not positive reinforcement.  This is a mixed bag of reinforcements and punishments, more aversive than not.  We don’t even know if any of the elements were executed correctly.  In hindsight, I know that in Kiki’s case, they were not.

Kiki was lucky because I am stubborn.  Despite multiple trainers telling me to give up on her, I kept searching for a solution.  Our journey took us to a workshop and some private sessions where some exercises had a profound impact on my way of thinking.  I realized that intending to use positive reinforcement was not the same as using it effectively.

In each of these exercises, we initially fed so quickly it was obscene.  The clicker took on a machine gun staccato.   Feed!  Faster!  Click – TREAT!  Click – TREAT!  Again!  Again!  Faster!  You’re clicking too slowly!

I can’t emphasize how fast we were feeding that dog.  She was right and right and right.

Twitchy toes Kiki, who couldn’t stand still on account of her tail wagging so hard, learned a stand for exam.  On the formal recall, she learned to target a bulldog clip attached to the front of my shirt.  I stepped back 5 cm, asked for the touch – click and treat.  Five centimeters seems absurdly easy!  (That’s two inches for my American friends.)

Wouldn’t you know: it worked.  Not only did it work, we finally moved AWAY from food rewards.  The gasp I heard at her first competition as she charged toward me and sat in a perfect front , I can still hear it.  It’s one of my fondest memories of Kiki.  Within months of measured, planned and well-executed positive reinforcement, Kiki stopped pulling on leash.  Did I mention, she did this on a body harness?

Properly executed, positive reinforcement works.

It all comes down to the Bob Bailey trinity.  Timing – criteria – rate of reinforcement.  While providing feedback at the right moment in time is important, it is equally important to raise expectations in small, measured increments.  Too big of a leap and the dog goes from right to wrong.  We also need to provide fast feedback in the initial stages of training.  Pause too long between reinforcements and the dog checks out, gets bored or would rather be elsewhere.

Trinity
The greatest irony being that strong technique allows owners to wean away from food treats faster.  A few weeks or a month of good technique may include bursts of obscene machine gun clicking.  The alternative is months or years of slow feeding and searching for the miracle cure that may or may not come.

It doesn’t help when fear of food mongering encourages people to reduce treat use.  Stingy feeding and large leaps in criteria create a weak technical foundation.  That weakness is what creates, “I tried positive reinforcement, but it did not work.”

Technique matters…period.  The tougher the dog, the more important technique becomes.  I see those dogs as the ones that point a paw at me and say, “You have more learning to do….I shall point out your technical weaknesses!”  I cherish those dogs for the lessons.  I truly hope that I never fall into the mindset to think that I know it all.

Re-visit the dog training trinity first when a dog is not responding in the way you want.  We seem to spend so much time, effort and money adding new tools to the toolbox.  It pays to sharpen and hone the essential tools we already have.

It’s like buying a food processor because your dull knife won’t cut a tomato.  Of course the knife isn’t working.  You have not maintained it.  Sharpen it.  Don’t neglect the essential tools of timing, criteria and rate of reinforcement.  They are the ones you need to count on.  Really, it’s not a moral issue.  At the end of the day we want dog training that works.