Looks like, well more than 'looks like' treat training is at the end of its controversial & dangerous road.

Partialy as a consequence of high proportion of food reward positive trained dogs turning up with behaviour problems at Lincoln behaviour clinic Lincoln (Danny Mills)
uni did a preliminary investigation with a veiw to further research via a full blown comparitive study. Food lure reward trained dogs had a 56% error rate, in comparison to simple operant shaped dogs had only 24% error rate.

I doubt that will come as much of a surprise to the victims of treat training & the inherant welfare problems that you buy with it.

Lincoln Uni: New study addresses fundamental question in animal training
http://www.lincoln.ac.uk/dbs/NewsletterDBS-0606.htm

>Looks like, well more than 'looks like' treat training is at the end of its controversial & dangerous road.

Far from it; it's not controversial (well, not amongst those who have the welfare of animals to the forefront) and not dangerous.

Denis, I remember when you asserted that clicker training was dangerous because a person can throw a clicker at a dog and 'might injure its eye'!

Far from it; it's not controversial (well, not amongst those who have the welfare of animals to the forefront) and not dangerous.

Oh well, Professor Mills must be wrong then.

That's perfectly possible; being a professor and publishing a paper doesn't mean you're automatically correct. Dr Wakefield was wrong, and condemned many children to suffer illness.

Well he is just one person so it is more than likely that against his finding a hundred will come up with different results - talk about making statistics prove what ever you want. I think I actually saw this questionnaire and I am very doubtful it would bring forth any meaningful result.

Denis, you're just revealing that you don't understand what you're talking about.

The link you have posted is comparing lure and reward training with shaping.  Therefore BOTH methods in the study are positive and BOTH use food as the ultimate reward.  The only difference is that in one variable, the food is put on the dog's nose and the dog is lured towards the mat with it and in the other study, the dog is clicked when touching the mat and the reward follows afterwards.

The study found that the presence of food as a lure - before the correct behaviour - causes the dog to be distracted and not to associate the correct response with the food.  Training using a marker device or clicker solves this, since the food is not presented until afterwards.

It would simply be pointless to have a study researching whether training with food is effective - since it is a well known FACT that it is.

It would simply be pointless to have a study researching whether training with food is effective - since it is a well known FACT that it is.

The victims say different

Have you actually read the article, Denis?

"However, food may be used in many different ways. It can be used as a reward for behaviour, which just happens to be right (simple operant shaping) or as a lure to encourage "faultless" learning."

Food is being used, as rabid has pointed out, for BOTH methods - you are trying to convince us that treat training is awful by comparing it with the better success of what?  Treat training!

Yes, lure-based training can slow progress - but this is a very limited study, and may not take into account different learning styles of dogs.  Just within my pack I have some who can do free-shaping (or simple operant shaping, in this article) and others who do not grasp the idea and must be lured - both types learn just as well, just in different ways.

"However, food may be used in many different ways. It can be used as a reward for behaviour, which just happens to be right (simple operant shaping) or as a lure to encourage "faultless" learning."

Yes, it's been in use for thousands of years, probably how neanderthal ended up with some kind of domestic relationship with those dog types of the upper plaestecine. It teaches dogs 'what 'to do' to get a treat' if thats what its priority preference is in the environment at the given point in time. Donkeys years old stuff, well understood as are its limitations as a reinforcer.

That is a victim of a bad trainer, not reward based training.  I can't believe you actually think that shows the dangers of treat training!  Good grief.

All that is said in that video is that the owner was advised to give the dog frankfurters for doing something good.  That is positive reward training at its very basic and no, it won't get a recall from a dog like that by itself - that is the fault of the trainer.  They should have got the owner and dog on a recall training plan, starting with minimal distractions and building them as the dog is recalling reliably at each stage.  Frankfurters may well work perfectly for this dog in this situation - but only with a planned and correctly applied training plan, which any decent trainer should know.  That's also not a reflection on the APDT - it's a reflection on the individual trainer.

Now, what exactly was the trainer filming the dog doing?  I saw no guidance for the owner on how to control him; all she's doing is randomly pulling him about on a harness designed to cause discomfort when the dog pulls, which isn't actually making any difference anyway.  The dog looks stressed when he's being pulled about.  The owner is stressed.  And, the dog is getting plenty of opportunity to practice his pulling and ignoring behaviour while the "trainer" is just standing there filming it and saying 'poor dog'.  And the comment about him needing to be punished?  Rubbish!  What he needs is a set training program, including other training besides recall, to build a relationship with his owner and teach him to pay attention to her - all of which is perfectly acheivable with reward based training, and not just food; toys, running, sniffing, whatever works for the dog.

My youngest was just like that when he arrived - and yet, now he will spin on the spot to recall, and is very attentive to me out and about where before he behaved like that lab.  I have not punished him once for not listening or pulling me, I have simply used reward training in a planned way to teach him to pay attention.

Jeangenie - I work in vaccination and I can tell you that even now children are still not being vaccinated because of his 'research'.

> Far from it; it's not controversial (well, not amongst those who have the welfare of animals to the forefront) and not dangerous.
>
> Oh well, Professor Mills must be wrong then.

I think it is you that have it wrong. in that article you posted Prof Mills certainly never said using food was ineffective or dangerous. Infact the article indicates using food works. it is investigating which  method of using  food is most effective.

May be tunnel vision, we all suffer from it over something or the other, you read and only take in that which fits your way of thinking not what is actually said.

> Partialy as a consequence of high proportion of food reward positive trained dogs turning up with behaviour problems at Lincoln behaviour clinic Lincoln (Danny Mills) uni did a preliminary investigation with a veiw to further research via a full blown comparitive study.

Where you did you come up with that?? That wasn't the reason for the study atall.

The article you linked to gives the reason behind the study:

"Amongst trainers there has been considerable debate over the impact of food lures on learning ability, with proponents arguing that it facilitates rapid response acquisition by eliminating irrelevant or incorrect behaviour during learning and opponents countering that the presentation of food, biases attention towards feeding so the animal does not focus on the task being taught.

The aim of the Lincoln study was to conduct an initial investigation of this issue, using a group of students with no bias towards one method over the other."

> Food lure reward trained dogs had a 56% error rate, in comparison to simple operant shaped dogs had only 24% error rate.

Yes, and as you'll see from the write up, explaining the study, that what they refer to as 'simple operant shaping' is using food to reward a behiaviour. -
"However, food may be used in many different ways. It can be used as a reward for behaviour, which just happens to be right (simple operant shaping)"

The article is about food LURE vs food REWARD.


Thank you for bringing it to every bodies attention that this study found FOOD REWARDs produces great results :)

Where you did you come up with that??

Professor Mills, he did not specify food, what he said was he was allways getting owners turning up with "positive" or 'reward' trained dogs which were playing up in some way and owners telling him they could do nothing with them.

It can be used as a reward for behaviour

What is your understanding of the word 'reward' in animal training (inc humans)?

> Professor Mills, he did not specify food, what he said was he was allways getting owners turning up with "positive" or 'reward' trained dogs which were playing up in some way and owners telling him they could do nothing with them.

But the article about the study (both reasons for carrying it out & the subsequent findings) you linked to has no mention of this.

> It can be used as a reward for behaviour
>
> What is your understanding of the word 'reward' in animal training (inc humans)?

Rather vauge question considering the text of mine you linked to.
My vague answer is: a reward is something the subject experiences as a GOOD thing.

In context to what I actually said,
A food reward is the use of food as a reward. To further explain, to reward a dog with food would be to give the dog some food (liver, cheese, kibble or anything that is food) as a rewward in order for the dog to LIKE the 'reward' and subsequently make the connection between the liking of the food and the doing of the behaviour that resulted in the dog getting the food (reward/good thing).

My vague answer is: a reward is something the subject experiences as a GOOD thing.

No thats not the meaning at all. When your talking about a >any< stimulus (food in this case) in animal learning the correct term is 'reinforcer', a reinforcing stimulus strengthens a behaviour, 'reward' means anything the individual wants it to mean but nothing specific, if the behaviour does not strengthen and tend to repeat it is not a reinforcing stimulus, operant does not prejudge if something will weaken or strengthen a behaviour untill the behaviour tends to repeat.

But the article about the study (both reasons for carrying it out & the subsequent findings) you linked to has no mention of this.

Why should it? unis don't make announcments on everything they do or think every hour of the day, if you go to Oxford uni you'll find the courses not the thousands of thoughts and ideas each tutor & students have every hour of the day!

> No thats not the meaning at all. When your talking about a >any< stimulus (food in this case) in animal learning the correct term is 'reinforcer'

Nope, when I am talking about a REWARD, I mean exactly what I said previoulsy.
You asked what MY idea of a REWARD was, not the 'correct' use of the word 'reinforcer'.

By the way, as  I never actually mentioned the word reinforcer, nor did the article, how have you linked my use of the word 'reward' with the use of the word 'reinforcer', unless the 2 are commonly interchangable? (which for some strange reason you seem be trying to argue is NOT the case, even though it is completley irrelevant)

>(ME) But the article about the study (both reasons for carrying it out & the subsequent findings) you linked to has no mention of this.

>(YOU) Why should it?

Because you stated it as the reason for the study.

Professor Mills, he did not specify food, what he said was he was allways getting owners turning up with "positive" or 'reward' trained dogs which were playing up in some way and owners telling him they could do nothing with them.

Why is it assumed it's the method at fault and not the owners application of it, or indeed other factors in the dogs home environment ?
I'm sure organisations such as Canine Partners who use reward and positive training, would be most surprised to learn their methods, that have been turning out highly trained dogs for some years now, are actually ineffective.

> Why is it assumed it's the method at fault and not the owners application of it, or indeed other factors in the dogs home environment ?

Indeed.  Negative associations and stress can affect learning too - I worked with a dog for a year who came from quite a stressful household (we worked on that, of course) and whose owner was a bit hyper; as a result he was a very sensitive dog, easily stressed, and took a while to learn anything from his owner or respond to her because he was on edge a lot.

Yet with me, he would do absolutely anything I asked immediately (the owner often said it was like magic, the difference in him between working for me and her) - I have no stressful association for him.  The methods were totally positive all the way - aversives would have likely caused a nervous breakdown in that dog - and they worked VERY quickly for him with me, but the environment and stress associations with his owners and normal life hindered things.