Call For Participants: Canine Temperament Testing

[WoofnHoof]

The test isn't whether a dog meets the standard for it's breed. The test is sbout whether a dog has a suitable temperament to rehome safely. Fearful dogs don't fit that criteria ( unless they're acted because they're in a shelter, in which case a good temp test will help determine whether this is the cause of the fear).

My girl would fail this test no doubt. She is a lovely, well mannered little sweetie. She is fearful and reverts to aggression if not managed properly. She is not a suitable dog for 99.9% of families and SHOULD fail a good temp test. This doesn't make her a bad dog, just a dog that isn't safe in the hands of Joe average.

Agree - and a test for each breed is wildly impractical anyway, especially since most of the animals that come through will be cross breeds. So you'd be pouring resources into developing a test for a dog that pops up, say, once a year? Once every 5 years?

But a standardised test seems impractical to me anyway given the vast differences between breeds and crosses, so what is the point other than covering asses in case someone gets bitten?

[WoofnHoof]

Didn't she just explain that this test is more about matching the right dog to the right family, not passing or failing dogs.

To me standardisation is about getting rid of individual differences to get statistically signficant results, since matching dogs to owners can really only work on a case by case basis I fail to see the relevance or practical application of a standardised test for dogs from a wide range of breeds and backgrounds.

[dancinbcs]

I would like to see research done on the owners who dump dogs in pounds and shelters, to find out why they think pets are disposable and why others should have to fix the problem they created. This is the root of the problem and if no one dumped dogs then there wouldn't be any need to worry about re-homing them.

[WoofnHoof]

I would like to see research done on the owners who dump dogs in pounds and shelters, to find out why they think pets are disposable and why others should have to fix the problem they created. This is the root of the problem and if no one dumped dogs then there wouldn't be any need to worry about re-homing them.

Absolutely, why spend time trying to standardize the dog population to suit people who are only going to stuff up the 'perfect' dog anyway? I know the ARG has done a lot of work on this 'perfect pet dog' concept and while it's useful to know the sort of dog the average Joe is wanting in terms of working out which breeds are best suited in a general sense and promoting those breeds rather than having them buy an unsuitable breed.

But at the end of the day though it's still a cop out for irresponsible owners, they can put in an order for their 'perfect' dog, non shedding, great with kids, medium size etc etc but they still get to dump it if it craps in the wrong spot on their lawn or they go on holiday and are too cheap to pay for boarding. There is simply no responsibility on the owner's part, no onus on them to learn about the dog, or the breed, or management, or training or anything really.

While I do see the relevance of testing and making life easier for shelters who have the hard job of choosing which dog is the most rehomable, the amount of variability in dogs just makes me wonder how applicable all these tests are in the grand scheme of things. There is no guarentee that the perfect dog that was the most rehomable according to the test still isn't going to bite some feral kid that sticks a pencil in it's ear. It's not going to guarentee the dog stays perfect after it spends a year or so in the backyard with little to no interaction. So is it really going to achieve the desired outcome of no bouncing back to the pound if there is no standardized test of the owner's suitability?

[corvus]

To me standardisation is about getting rid of individual differences to get statistically signficant results, since matching dogs to owners can really only work on a case by case basis I fail to see the relevance or practical application of a standardised test for dogs from a wide range of breeds and backgrounds.

Well, no, it's not. It's about using a single approach so that there are limited external variables that may influence the results. Like I said before, a standardised test needn't ignore breed. It could mean the individual tests within a behaviour assessment are carried out in a way that doesn't vary from dog to dog. Breed can be accounted for in the statistical analysis afterwards if there is enough information to do so. A standardised approach would mean the same things are measured in the same way using the same scale so comparisons can be made. This does not mean comparisons between breeds would be assumed meaningful. That is something you find out later in analysing the results. Ideally a standardised test would be well supported by science before it was adopted, but this isn't always the case.

Standardisation does not necessarily equal loss of information or smoothing over details. That's all in the design. It does mean the ability to make comparisons and test for validity. This is useful.

[Sheridan]

To me standardisation is about getting rid of individual differences to get statistically signficant results, since matching dogs to owners can really only work on a case by case basis I fail to see the relevance or practical application of a standardised test for dogs from a wide range of breeds and backgrounds.

Well, no, it's not. It's about using a single approach so that there are limited external variables that may influence the results. Like I said before, a standardised test needn't ignore breed. It could mean the individual tests within a behaviour assessment are carried out in a way that doesn't vary from dog to dog. Breed can be accounted for in the statistical analysis afterwards if there is enough information to do so. A standardised approach would mean the same things are measured in the same way using the same scale so comparisons can be made. This does not mean comparisons between breeds would be assumed meaningful. That is something you find out later in analysing the results. Ideally a standardised test would be well supported by science before it was adopted, but this isn't always the case.

Standardisation does not necessarily equal loss of information or smoothing over details. That's all in the design. It does mean the ability to make comparisons and test for validity. This is useful.

It's useful - to academics. If you think that such a test would be used wisely and or well, instead of the more likely scenario that Woof has posited, then you need to get out of the university. The real world, Corvus, is about what's easiest and cheapest.

[WoofnHoof]

To me standardisation is about getting rid of individual differences to get statistically signficant results, since matching dogs to owners can really only work on a case by case basis I fail to see the relevance or practical application of a standardised test for dogs from a wide range of breeds and backgrounds.

Well, no, it's not. It's about using a single approach so that there are limited external variables that may influence the results. Like I said before, a standardised test needn't ignore breed. It could mean the individual tests within a behaviour assessment are carried out in a way that doesn't vary from dog to dog. Breed can be accounted for in the statistical analysis afterwards if there is enough information to do so. A standardised approach would mean the same things are measured in the same way using the same scale so comparisons can be made. This does not mean comparisons between breeds would be assumed meaningful. That is something you find out later in analysing the results. Ideally a standardised test would be well supported by science before it was adopted, but this isn't always the case.

Standardisation does not necessarily equal loss of information or smoothing over details. That's all in the design. It does mean the ability to make comparisons and test for validity. This is useful.

You can limit external variabiles till the cows come home, the point I'm trying to make is the internal variables within each dog can't be standardized, not without making some fairly large leaps of faith such as timidity being a significant problem, it may be a problem in some dogs in others it may just be a natural reaction to a highly stressful situation, not necessarily indicative of a significant fault in temperament.

[Alkhe]

Isn't that the whole point of the study though? What we currently have is a pretty lacklustre, patchy "system" whereby many great dogs that would suit many homes never get there, because they behave a certain way in a pound environment.

In recognition of the fact that the temperament tests that are currently used by pounds and shelters aren't able to take into account the fact that timidity (or whatever) isn't necessarily an insurmountable problem, or a factor that makes a dog unrehomeable.. this study seeks to enable various other things to be taken into consideration.

I fail to see how this is a bad thing. This goes a much greater way toward taking into account different breed characteristics and tendencies than anything that exists for pounds to use so far. Of course it's not perfect. And of course it won't take into account the particular breed attributes and differences. But it's better than what we have now. Argue all you want about how other things would be better - but someone here is actually doing something. I for one think it's great, was really pleased to help, and can't wait to see the results.

And as for the key problem being people who dump their pets at pounds - well, yes. But recognising that that's something to tackle doesn't mean there shouldn't be better ways of dealing with the result of that. Which is dogs at pounds. Which are often assessed poorly, because assessment tools available and/or widely used are crap.

It's like saying what we really need to do is stop war, and anything less than stopping war is a waste of time. Go stop war if you want, but don't jump on the people who are plugging away at their own piece.

Edited February 4, 2013 by Alkhe

[corvus]

Standardisation does not necessarily equal loss of information or smoothing over details. That's all in the design. It does mean the ability to make comparisons and test for validity. This is useful.

It's useful - to academics. If you think that such a test would be used wisely and or well, instead of the more likely scenario that Woof has posited, then you need to get out of the university. The real world, Corvus, is about what's easiest and cheapest.

Validity and comparisons are useful to everyone. It's the cornerstone of being able to make informed decisions and reliable predictions. Sheridan, I have been dragged out of the university to speak to people in the industry who badly want standardised, objective testing. They don't want it because it would be easiest and cheapest, believe it or not. I think such a test could be used well if it were carefully formulated. It has been done before in other fields. Even in dogs to some extent. The Swedish DMA, for example, has been used to help collie breeders breed timidity out of the Swedish collie population. The DMA is not perfect, but there's enough science out there on it to support its use. It's geared towards working dogs, though, and may be too harsh for some dogs in the broader population.

WnH, no one CAN standardise internal variables in dogs? There is plenty that we can do, though. We can observe and work backwards, for example. There are plenty of people working on indicators of internal states and predictors of future behaviour. This is my life at the moment. I haven't the time or inclination to try to convince people on a dog forum. I can just say I've been neck deep in this literature for 3 years and have spoken to a lot of people about it, both academics and people in industry. I watch a lot of dogs behaving. The particular problem of behaviour assessments has been a bone I've been gnawing on for months. You can consider I maybe have some idea what I'm talking about or not. I really am not bothered. I am bowing out because I don't have the spare time for this discussion. I should never have got involved. I think that this study sounds great, though, and would encourage people to get involved.

[WoofnHoof]

Alkhe I see your point but I don't think the chances of standardising this type of thing is all that realistically feasible, I mean it takes hundreds of years of selective breeding to standardise a temperament and morphology. In the case of crosses it is unlikely that any test will be overly accurate anyway, we see so much variation in phenotype that it's unrealistic IMO to think that temperament will be any easier to characterise. In the case of purebreds then yes there should be a more predictable temperament but it's the breed standard that will guide this, generations of selective breeding can't be compared to a one-size-fits-all temperament test.

ETA corvus you posted while I was still typing, your points are fair enough but I am afraid I still disagree for the reasons I've outlined above.

Edited February 4, 2013 by WoofnHoof

[Sheridan]

Standardisation does not necessarily equal loss of information or smoothing over details. That's all in the design. It does mean the ability to make comparisons and test for validity. This is useful.

It's useful - to academics. If you think that such a test would be used wisely and or well, instead of the more likely scenario that Woof has posited, then you need to get out of the university. The real world, Corvus, is about what's easiest and cheapest.

Validity and comparisons are useful to everyone. It's the cornerstone of being able to make informed decisions and reliable predictions. Sheridan, I have been dragged out of the university to speak to people in the industry who badly want standardised, objective testing. They don't want it because it would be easiest and cheapest, believe it or not. I think such a test could be used well if it were carefully formulated. It has been done before in other fields. Even in dogs to some extent. The Swedish DMA, for example, has been used to help collie breeders breed timidity out of the Swedish collie population. The DMA is not perfect, but there's enough science out there on it to support its use. It's geared towards working dogs, though, and may be too harsh for some dogs in the broader population.

WnH, no one CAN standardise internal variables in dogs? There is plenty that we can do, though. We can observe and work backwards, for example. There are plenty of people working on indicators of internal states and predictors of future behaviour. This is my life at the moment. I haven't the time or inclination to try to convince people on a dog forum. I can just say I've been neck deep in this literature for 3 years and have spoken to a lot of people about it, both academics and people in industry. I watch a lot of dogs behaving. The particular problem of behaviour assessments has been a bone I've been gnawing on for months. You can consider I maybe have some idea what I'm talking about or not. I really am not bothered. I am bowing out because I don't have the spare time for this discussion. I should never have got involved. I think that this study sounds great, though, and would encourage people to get involved.

Ah, but you bow out before you explain why the people who 'badly want standardised, objective testing' actually want it and what they will use it for. And I am still at a complete loss as to how breed traits will be assessed in standardised testing. And since Diana R hasn't been back, before you disappear completely, do feel free to explain how gameness can be tested in this objective testing. Given gameness is erroneously labelled as dog aggression by people who don't know what they're talking about, I would really like to know.

Edited February 4, 2013 by Sheridan

[corvus]

Ah, but you bow out before you explain why the people who 'badly want standardised, objective testing' actually want it and what they will use it for. And I am still at a complete loss as to how breed traits will be assessed in standardised testing. And since Diana R hasn't been back, before you disappear completely, do feel free to explain how gameness can be tested in this objective testing. Given gameness is erroneously labelled as dog aggression by people who don't know what they're talking about, I would really like to know.

I can't speak for them. It wouldn't be fair given they are not here to correct me if I misinterpreted. I think a lot goes unsaid because it's delicate. And it's kinda complicated as well. They have to take into account their volunteers and employees as well as their public image and on top of that they do actually want what's best for dogs given their limited resources, but they also have to protect the community. They are doing the best they can and have a difficult job. My feeling is they would welcome anything that made it less arbitrary and subjective.

The answer to your other questions is a matter of statistics. The short answer is depends on whether you're a frequentist or a Bayesian, but either way, very complex models. And as for gameness, lots of baseline data. There has certainly been talk about this in academic circles. Breed specific behaviour is one of those things we know exists but have trouble capturing due to sample sizes. That's all I have time for, sorry.

[WreckitWhippet]

Let's put it another way then Corvus, if YOU were developing a temperament test, how would YOU assess "gameness" in a standardised test ????

[Diana R]

Hi Diana, it was me who mentioned to you that it was on DOL. I'm Feather and Kibahs (the scaredy blue whippets) mum. :)

Hi Kirislin, thank you :) I've had a number of people contact me after seeing it on DOL, so I must thank the OP as well.

I'm not sure if it was someone from here, but the link to the Adelaide story was sent to my main supervisor over the weekend too and he is currently following it up with them - he had the same initial thoughts that I (and by the looks of it, possibly Corvus) had, that they may have got their wires crossed and it's something to do with Pauleen Bennett's work (though if Tammy didn't know about it, I also doubt that...).

I'll go through the posts one by one now (sorry I haven't been back sooner - my OH was taken to emergency last night with burns to his face and hands, so I went straight there after we finished up yesterday and we've just gotten home again from the hospital).

[Diana R]

It's all well and good to say that a standardised method is useful in pounds in getting dogs that suit Joe Blow who can't be stuffed doing their own breed research. What if the research ends up being used by councils etc to apply to the general population? I think the whole concept of standardisation is problematic when you are dealing with a variety of breeds. How about creating a standardised test for Joe Blow to assess their suitability to cope with the dog they want to buy? Ultimately you can have the most adoptable dog in the world but it doesn't matter if Joe is a numpty who knows nothing about dogs and doesn't need to know in order to adopt one.

Personally I feel all this research into temperament testing is a waste of time and resources and does nothing to address the disposable mentality of the general public towards pets. I've come across people who have owned huskies in the past, it ran away, it escaped etc etc, the slightest bit of breed knowledge would have told these people not to buy a husky in the first place and if they still wanted one to have a bit of understanding of the breed and cater for it's needs. The breed is the most critical factor in the behaviour and temperament of the dog and yet it is largely overlooked in favour of standardizing all dogs. Makes no sense to me.

I think you may have commented again since this W&H, but to address these points - I'm still not clear on whether you are concerned about a standard test (i.e. a set of stimuli that's used in a particular way to elicit responses which give us some info about the dog) or a standard pass/fail assessment based on that info? I completely agree that a standardised pass/fail test isn't a suitable way to assess the suitability of a wide variety of dogs given that they all fulfil very different needs and roles. I can see lots of useful things about a standard set of stimuli and a method for reliably gathering info about a dog though.

Sure, there are lots of people that don't research the dogs they subsequently purchase and end up in trouble, and in the case of those people that purchase purebred dogs with known histories, they have no one but themselves to blame when predictable issues arise. However, there are also lots of people who come into shelters and have a fairly decent idea of what type of dog they would like (a chilled out socialite that would do well in an urban setting or an intelligent, energetic dog that they can compete in sports with) and at the moment, we don't have a reliable way (that's consistently been shown to work) to help them decide which dog would suit them best - this is what we are trying to address.

I honestly would have thought that the purebred fraternity would have the best understanding of why knowledge of a dog's temperament is so important for a new owner, as that's pretty much the whole basis behind purebreds.

[Diana R]

Didn't she just explain that this test is more about matching the right dog to the right family, not passing or failing dogs.

To me standardisation is about getting rid of individual differences to get statistically signficant results, since matching dogs to owners can really only work on a case by case basis I fail to see the relevance or practical application of a standardised test for dogs from a wide range of breeds and backgrounds.

I think we're still having the same crossed purposes here - in the VAST majority of dog bite cases, the only consistent factors that seem to have some type of 'causal' effect are those related to owner behaviours, so the idea that a temp test of any type, on any dog, in any situation, will somehow 'proof' a dog against biting in the future is flawed.

That said, there are factors associated with temperament that will make a dog more likely to be stressed/aroused and therefore react 'badly' in particular situations - for example, a dog with a very passive coping style, who is also very emotionally reactive and has a poor recovery (i.e. he becomes stressed easily, doesn't outwardly show overt signs of that stress and stays stressed for extended periods of time) would be a very bad choice to go into a home with novice owners with small, noisy children - even if he never bites and the owners hardly ever recognise how uncomfortable he is because he sits quietly and melts down (which means they are more likely to stress him even further inadvertantly, meaning he is more likely to bite at some stage...), it's a welfare issue for the dog. I'm failing to see us being able to recognise this and place the dog accordingly is a bad thing.

For those that are concerned that this test will be used by shelters simply to justify euthanasia, without another decade research following on from this, in which a wide variety of dogs that have been tested are followed and their 'aggression records' and temp results are analysed, they can't. Even if they did do that, I can save you the time and tell you that the results will most likely show the same results that every other investigation into dog bites shows - that owner behaviour is the most influential factor. That doesn't mean that we should rehome every dog, o even that we shouldn't pay attention to aggressive responses that dogs show in the shelter environment - we just need to work out a way to help shelter staff (who are often poorly trained and have hugely varied levels of experience) to gather more info about the dog before making any decisions.

[Greytmate]

I honestly would have thought that the purebred fraternity would have the best understanding of why knowledge of a dog's temperament is so important for a new owner, as that's pretty much the whole basis behind purebreds.

There is suspicion from a lot of people here about behaviour testing and its perceived use as a reason for euthanasia. Also there are people who believe that any type of dog behaviour can be handled with competent management and anyone who cannot do this has failed the dog.

But there are also people here who use behaviour tests and foster assessment in an effort to find out all they can about the dogs they sell, and I don't see them criticizing here. Good luck with your work. Anything useful you discover can only help rehome more dogs.

[Diana R]

I would like to see research done on the owners who dump dogs in pounds and shelters, to find out why they think pets are disposable and why others should have to fix the problem they created. This is the root of the problem and if no one dumped dogs then there wouldn't be any need to worry about re-homing them.

I have to get up again and I will come back tonight, but this work has been done (and continues to be looked at) - and a key part of addressing this is making sure that owners (who have realistic expectations of what dog ownership is about) end up with a dog that will meet those expectations - owners with dogs that meet their expectations have better bonds with their dogs and will invest more resources into their care...

I will be back :-)

[Diva]

For those that are concerned that this test will be used by shelters simply to justify euthanasia, without another decade research following on from this, in which a wide variety of dogs that have been tested are followed and their 'aggression records' and temp results are analysed, they can't. Even if they did do that, I can save you the time and tell you that the results will most likely show the same results that every other investigation into dog bites shows - that owner behaviour is the most influential factor. That doesn't mean that we should rehome every dog, o even that we shouldn't pay attention to aggressive responses that dogs show in the shelter environment - we just need to work out a way to help shelter staff (who are often poorly trained and have hugely varied levels of experience) to gather more info about the dog before making any decisions.

My main concern is it being used as part of a drift to a "McDogs" culture, where only a certain sort of bland and (to me) boring temperament is considered valid. Some of us are very attached to our somewhat less than mainstream breeds and their idiosyncracies. I'll admit it was the media article that raised my concerns, before that I looked at it as an interesting piece of research. But that article put quite another, somewhat worrying, slant on it. I dont think that is your fault at all, it is just reflective of a failure of trust between some of the the dog community and government legislators more generally.

Edited February 5, 2013 by Diva

[Diana R]

Sorry, Diana R, but you cannot classify by group. A bedlington is as different to a cesky to a staffy. Yes, they're terriers, but a dog that only goes to earth is going to have a different temperament to a dog that does multiple tasks. How, for example, do you measure gameness?

Define 'gameness' - if we are on the same wavelength when talking about gameness, I don't consider it a trait of it's own accord, but rather a certain 'collection' of underlying temperament traits that results in the behaviour that we see as 'gameness'. Basically, I would say a dog that has moderate to low emotional reactivity, a very proactive 'coping strategy', few social and non social fears (and a good rebound) and some marked degree of chase proneness (aka prey drive or tendency towards strong predatory behaviours) is what most people would refer to as gameness (i.e. the dog thinks on it's feet, is comfortable using force when pressed, is 'balanced' in how it responds to environmental stimuli and shows relatively strong predatory tendencies). Bearing in mind that we've done no analysis as yet and I'm just commenting on what I've seen during the tests we've done so far (so when we've done the analysis this could all mean nothing), the only thing listed above that we don't seem to getting a fairly clear picture of during the tests is chase prone-ness/predatory tendancies, as there seems to be issues with arousal/stress levels inhibiting the dogs responses to prey stimuli.

I'm not sure I'm fully understanding your comment further on about dog aggression and gameness, as I'm reading it as you implying that they are mutually exclusive (i.e. that a game dog which reacts inappropriately and aggressively towards other dogs is not dog aggressive or potentially lacking in social skills because it's game)?