My Photo

Categories

follow us in feedly

Enter your email address:

Delivered by FeedBurner

Best Of KC Dog Blog

Become a Fan

« Weekly Roundup - Week Ending 3-10-13 | Main | Very good sentences...What I'm reading today... »

March 13, 2013

Comments

Randy

The good news about social media is it allows good information to be shared quickly. The bad news is it allows bad information to be shared quickly too! Morale of many of these stories is u probably don't want an animal (dog, big cat, snake, monkey etc.) that is big enough to kill you or your children. Thanks as always for the information!

Olivia

This is VERY well written. It's been sad reading all the hatred about Pits in general. I do not own, but have fostered many and to hear and read what people "would do to them if they ever came in contact with a Pit" is atrocious. That sweet, innocent little boy lost his life. The reason is irrelevant. We need to be there to support his family, no matter what type/breed pet you own.

Olivia from Milwaukee, WI

Dubv

"I've been in communication with several folks up in Wisconsin and there seem to be a fair amount of speculation as to the cause of the attack, but nothing tangible to date to report. "

I wonder if your experts that speculating have this option on the table: no cause can be determined, the mere fact that the baby was near the dogs was enough, and breed genetics makes a fatal outcome more likely.

This is what you folks don't get. A non-fighting breed dog is less likely to be aroused to the point of carrying out a full bore attack like this and non-fighting breed dogs are more likely to bite and release or if they do grip, an adult human delivering blows to the dog can make it stop more easily.

The reason pit bulls are more likely to kill things is taped to the end of your nose, I hope you read that memo.

Nancy Tranzow

DubV- I can only hope you continue to follow this blog to become more educated as your comments seem parallel to a certain organization who is full of misinformation.

Any dog can bite, and a well placed bite can result in a fatality. This is clear when dog fatalities are reviewed, with many different breeds being represented including, huskies, malamutes, chows, golden retriever mix, pomeranian, rotti, G.Shep, dobie etc.... "Arousal" is an interesting term and I'm not sure how that actually plays into dog behavior as it's not really a factor in bites. But I see several interesting descriptors in your comments like "full-bore". Also, what exactly is a "fighting breed"?

I think if you spent some time reviewing bite style on film you would see that no one breed has a "style". Dogs bite. They bite and hold or bite and release but neither is specific to a breed at all. A good example is tug-o-war where both dogs bite and hold on.

Pit bulls are not more likely to kill "things" and in fact if you stay off of biased sites and look to experts who are willing to put their name on research rather than hide behind inflammatory marketing titles, you will see that bites and fatalities fall across many breeds, under many circumstances BUT with themes of neglect, starvation, abuse, chaining, intact dogs and criminal parents being very common.

I wish for continued education for you Dub, and not on the sites we both know you frequent.

Diana Crawford

Were either of the dogs tied up? As we know, that can produce strong territorialism & aggression for the dogs to protect their little bit of space on a chain.

Take care.
~ Rosebud

Brent

Diana -- at this point not a lot of information is known about the circumstances surrounding the attack and likely won't be released until after the police report.

Nancy, Dub has no interest in intelligent dialogue on the topic.

Dub, dogs don't attack for no reason -- and any implication of that is why not listening to behavior experts is so dangerous. Also, the idea that you use the whole "fighting breeds" thing shows me you have no knowledge of genetics or the reality that a) most of the dogs classified as "pit bulls" today have not been used for, or bred for fighting in decades. Many were specifically removed from fighting lines 90 years ago. In canine genetics, entire breeds can be created in 4 generations over the course of a decade. Most 'pit bulls' now have as little in common with "fighting breeds" as standard poodles have to their hunting roots. But hey, don't let science get in the way of a good, scary catch-phrase.

Joel

I would also add that from what I've seen, posts of a dissenting opinion are not likely to be removed here as long as the discussion remains respecful. This is in direct contrast to the modus operandi of certain other organizations.

Dubv

"Dub, dogs don't attack for no reason -- and any implication of that is why not listening to behavior experts is so dangerous."

Of course, I believe in cause and effect as much as anyone. I believe in determinism. But the reason for a dog attack can be so small that it hardly counts or the reason can be internal to the dog. That it is always traceable to a human misstep is YOUR cognitive bias and one shared by so called "experts".

"Also, the idea that you use the whole "fighting breeds" thing shows me you have no knowledge of genetics or the reality that a) most of the dogs classified as "pit bulls" today have not been used for, or bred for fighting in decades."

I know a great deal about genetics. Once the genes have been selected for, unless they have been selected against, then they remain in place and will only change via a slow process called genetic drift. This is why working dog lines that have been bred for dog shows and not the field still carry out genetic behavior such as herding, but not to the same degree as a line that is still worked with all breeding based around it. Further, whether or not a particular pit has been fought does not change its DNA. It's funny. The folks at more responsible pit bull orgs recognize that the fighting lineage does make pit bulls unique and that they carry those traits still. I guess you didn't get the memo.

Dubv

" In canine genetics, entire breeds can be created in 4 generations over the course of a decade. Most 'pit bulls' now have as little in common with "fighting breeds" as standard poodles have to their hunting roots."

Brent, yes or no.

Dog fighting still exists. Some of these fighting dogs are bred. Some end up outside the fighting circles. Dog fight bust dogs are sometimes adopted out to the public (ala' vicktory dogs).

What you wrote about the speed at which a new breed can be created is a smokescreen.

Demonstrate that there has been any real effort to breed away from fighting dog traits and that the pit bull is now substantially different in the key areas that would relate to their danger levels. You can't do it. The breed stewardship for pits has been atrocious. They still carry the genes for things like dog-directed aggression, although there is a probability distribution around this trait as is the case for all things.

You are resorting to making things up. Again, the poodle thing is a red herring. Show me that the changes in the poodle are mirrored by pit bulls. If not, there is no reason to bring them up.

Dubv

It seems that Cesar admits there are fighting breeds and that genetics matters. I could post pit bull owners/advocates all day that agree with me on the basic breed traits. They just don't complete the syllogism that concludes with pit bulls being an above average risk.

A Cesar Millan quote taken from

http://fullcomment.nationalpost.com/2012/11/28/barbara-kay-cesar-the-dog-whisperer-should-whisper-the-truth/

What do we mean by that? Well Cesar, you yourself know what the traits of a fighting dog are. I quote you: “Yeah, but this is a different breed…the power that comes behind the bull dog, pit bull, presa canario, the fighting breed – They have an extra boost, they can go into a zone, they don’t feel the pain anymore. … So if you are trying to create submission in a fighting breed, it’s not going to happen. They would rather die than surrender. If you add pain, it only infuriates them…to them pain is that adrenaline rush, they are looking forward to that, they are addicted to it… That’s why they are such great fighters.” You go on to say: “Especially with fighting breeds, you’re going to have these explosions over and over because there’s no limits in their brain.” Wow, is that what you want in a pet? A dog that has “explosions over and over” in its brain?

Caveat

Man this Dub likes the 'sound' of his own nonsense, doesn't he? And quoting BK is just plain nuts. She knows less about dogs than Dub apparently does. She's just a crazy old Canuck lady whose son is editor of the page she bloviates at. As for Cesar, bless his heart but he is not an expert on dog breeding or genetics either. 'Even' Cesar indeed - he is an entertainer on television.

First of all you can't 'breed' dogs to 'fight' (or train them to do so) and nobody wants a purpose-bred dog that is unreliable regardless of the job at hand. In dog breeding for a purpose vs for appearance, you try to select for a certain temperament (which includes intensity, reactivity, etc) and overall physical ability. What is heritable is not related to complex behaviours such as species-specific aggression, or 'fighting' - both of which would be genetic dead ends. In the case of dog fighting, the desired temperament was one that was actually very stable. If you read any of the books by the old dogmen, such as Colby, you will learn that even though he was strongly selecting for fighting ability, he often got entire litters without one pup that he could use for that purpose. Sometimes he got one or maybe two. This is always true - otherwise every field bred retriever would be a champ, every Agility winner would produce dogsport champions. Unfortunately, genetics just don't work that way.

Those pit dogs were not aggressive in the sense that most people mean it, since sudden attacks are usually the sign of a dog that is afraid/unstable. There are pictures of an old-time grand champion playing with his neighbour, a poodle.

Now add in the fact that the majority of alleged "pit bulls" are randomly bred mutts of unrelated and unknown ancestry, and that the AmStaff has been a conformation dog since the 1930s and you really have to wonder about people's overall knowledge of what dogs are and aren't. They aren't manufactured items. You can breed for a pathology, which is what the hysterical people are describing when they talk about "pit bulls" and other types, but it is very difficult because recessives always carry other recessives so you never know what you will get. It is also unnecessary because that's not the way things work.

As for the silly challenge re: whether APBTs have been 'bred away' from fighting ability, nothing could be easier: the APBT is always in the top 5 of purebred dogs in the US. Millions of them are kept as family pets by people from all walks of life. Very few of them exhibit this bizarre pathological behaviour so beloved of the bogeyman crowd.

How about the Vick dogs? Look at how they are living normal lives, getting along with other dogs and generally behaving like good family pets.

Wrong place for nonsense, Dub.

Emily

I could talk about breed, genetics, behaviorally-mediated aggression, the history of "pit bulls," etc., but there's so no point. Rule out neglect, rule out abuse, rule out lack of socialization and training, rule out going nuts stuck on the end of a chain, rule out poor health, poor husbandry, lack of management, etc.--rule those out first, and then maybe breed is worth talking about. The trouble is, once those factors are ruled out, there's rarely an incident to talk about...

Dubv

"And quoting BK is just plain nuts."

It was BK quoting Cesar.

Dubv

"First of all you can't 'breed' dogs to 'fight' (or train them to do so) and nobody wants a purpose-bred dog that is unreliable regardless of the job at hand."

I can't believe this. It is obvious that there are fighting breeds purposefully crafted for fighting. Stop denying reality.

"What is heritable is not related to complex behaviours such as species-specific aggression, or 'fighting' - both of which would be genetic dead ends."

Please explain this to the owners of border collies that notice puppies herding things. I suppose herding things is not as complex as fighting. Also, where could you possibly get that fighting ability is a genetic dead end? Humans selected for it and propagated. Are you trying to compare this to an animal in the wild?

"If you read any of the books by the old dogmen, such as Colby, you will learn that even though he was strongly selecting for fighting ability, he often got entire litters without one pup that he could use for that purpose."

So now they can be bred for fighting, got it. Oh, and Colby, a dog in his yard killed his nephew and he kept the dog, bred it, and sold the puppies.

"Sometimes he got one or maybe two. This is always true - otherwise every field bred retriever would be a champ, every Agility winner would produce dogsport champions. "

Yes, some greyhounds are not fast enough to win a race, but on average, greyhounds are faster than labs. Do you think my general statements only hold true if they hold for every particular pit bull? Please.

" There are pictures of an old-time grand champion playing with his neighbour, a poodle."

I can find pictures of a gangster being nice to his grandkids, so what?

"They aren't manufactured items."

Domestication is not exactly like manufacturer, but it is similar.

" but it is very difficult because recessives always carry other recessives so you never know what you will get."

Recessive genes do not carry other recessive genes. Complex traits are rarely coded by a single gene and the closer the genes are to each other on the chromosome the more likely they are to be found together in offspring. So what? You are typing to someone that has forgotten more about genetics than you currently know.

And APBTs are not in the top 5. That is wrong on its face.

Dubv

"Rule out neglect, rule out abuse, rule out lack of socialization and training, rule out going nuts stuck on the end of a chain, rule out poor health, poor husbandry, lack of management, etc.--rule those out first, and then maybe breed is worth talking about. The trouble is, once those factors are ruled out, there's rarely an incident to talk about..."

In a data analysis situation, you do not RULE out variables. You estimate the role each plays while attempting (either by experimental design or mathematical reasoning) to hold all other variables steady. This is the root of the famous Latin phrase ceteris paribus (which I'll leave you to look up) and encapsulates a large part of the scientific method.

I love being talked down to by you folks. It is funny to me.

Erich

Dubv, you will never be able to perform the precise sort of data analysis you describe. These stories are media articles written by journalists. Even the breed of dog is a guess, and certainly all the other factors you want to "estimate" are either guessed at or not even addressed.

If society were more interested in finding out why these attacks happen and blaming the parties responsible the facts might be gathered. Instead, if anyone cares, it is a pet owner, a family member, and insurance company debating a settlement.

Two things are clear to me, be careful with dogs and small children and dogs in groups take cues from each other. I have fostered dogs and if there is potential for a problem I use crates and electric collars.

Emily

Nice dodge, Dubv :) Okay, don't rule out the variables mentioned; factor them in or control for them. Until that's done, there's really nothing to talk about.

Dubv

"Nice dodge, Dubv :) Okay, don't rule out the variables mentioned; factor them in or control for them. Until that's done, there's really nothing to talk about."

Not a dodge at all. You either do not know how a proper analysis would be carried out OR you are imprecise with language.

By the way, with the data set as is, you could not perform a formal analysis that would be telling. You would need to collect as much information on potential covariates from each animal, owner, and situation of interest (probably a serious attack or killing) and then collect the same on a random sample of animals that did not attack/kill.

You could then perform what is called a retrospective study which would have some limitations, but is probably the best you could get given time/money/data constraints.

So, what Brent, Crosby, and others are doing by ONLY looking at attack cases is faulty.

At least I am mentally comparing the sample of killing/attack dogs to what I know of a random subset of non-killing/non-attacking dogs.

Looking at it this way tends to put Brent's claims into perspective. I don't think he, Crosby, Delise, or other pit advocates even know that is the way the problem should be addressed.

It is faulty to merely look at each positive (attack) case and then find some exculpatory factor. Their type of analysis would not even pass for an undergraduate thesis.

But pit bull owners tend to both be narcissist and suffer from the downside of Dunning-Kruger effect, and so they will think I am just a hater making things up and rambling.

Joel

Oh great, yet another one who has somehow performed a psychiatric anaysis of pit bull owners.

Please tell your BSL friends that they should include the 99.something percent of pit bulls who don't attack/kill in their analyis as well.

Build a dataset where you collect information on the dog's age, breed, health, gender, status, socialization level, quality of care, aggression level, history of attacking, circumstances of any attacks, etc. Perform some sort of chaid or cart analysis and tell us what you find. In short sentences, since we're on the slow side. Seems like you've got this all figured out. We'll wait for your findings. Until then, not too many knowledgeable people are buying the idea that breed is the determining factor.

Emily

Dubv, sorry, the world is filled with extremely intelligent people with passionate opinions and no data. I breathlessly await real data; the rest is... hmm. BSL folks make assertions absent any evidence to the scientific standards you yourself express. They've made the claims; let them cite the studies and present the data in peer review. Until then... nothing to talk about but passionate opinions, yawn.

Dubv

" BSL folks make assertions absent any evidence to the scientific standards you yourself express. They've made the claims; let them cite the studies and present the data in peer review. Until then... nothing to talk about but passionate opinions, yawn."

You are making claims to my beliefs. I am not as interested in BSL as I am the following: stopping the white wash of likely breed influence that endangers humans and pets.

Pit advocates make claims about pit bulls being fine family pets. To make that truth claim, you would need the type of evidence you are asking of me. You do not have it. The record is clear enough that if someone were exercising the precautionary principle, then they would be better served with a common breed of dog that is not embroiled in this controversy.

Let's say a full study has not been conducted yet that indicates that large trucks are more dangerous to other motorists than say a smaller car. All preliminary evidence, from design to a general perusal of traffic accidents, indicates that large trucks are more dangerous to others. Now, imagine if a large truck club were stating that they could explain away the apparent danger of large trucks and that large trucks are fine, even if you care what will happen if you smash into a family. Would you feel this was perverse on their part?

When the pit bulls apparent record is explained away, and the "it's all how you raise'em" claptrap is espoused in the face of what we know about what the breed was created to do, then you are acting in a similar manner.

Emily

What "record" are you referring to, Dubv? Please present the citations and links to the data.

Joel

CHAID and CART can handle non-parametric data, but that has nothing to do with whether or not they can test hypotheses vs. predict outcomes. They can be used to establish decision trees which could be used to study the most likely determining factors in the dependent variable. The fact that these analyses are non-parametric is the point, because the data can be ordinal, nominal, or categorical. It's not perfect but it can be used to investigate a dataset of mixed variable types. The problem is accurately coding the behaviors and conditions, which you mentioned.

I'm not trying to be clever. It's clear that nobody here can measure up to you. You've made that abundantly clear.

Nobody is saying pit bulls are for everybody, or that larger dogs aren't likely to cause more damage than small dogs if they bite someone. But the demonization of the entire breed using a bunch of phony data and scare tactics does nobody any favors.

Dubv

"CHAID and CART can handle non-parametric data"

Data are not nonparametric or parametric. Models are either parametric, nonparametric, or of other classes. That you would make this basic error should give you pause. Please google "nonparametric data" and then "nonparametric models" and see which one gives more intelligible information.

They do not determine most likely determining factor in any precise sense. You are typing to someone that has published statistical packages in the most popular open source language that handles these (R Cran).

You are wrong in your thinking. I can state it definitely.

That they are nonparametric is useful in certain scenarios but precludes them from testing hypotheses in any rigorous manner. It is rather important to realize that nonparametric models cannot test hypotheses. It has everything to do with this, and you are simply unaware of it. A hypothesis can only be tested when the general relationship (perhaps linear or inverse square or some other) between dependent and an independent variable is specified in advance and model fitting only estimates the parameters governing the general relationship (and nonparametric techniques do not, by design, do this). That is very primitive knowledge. However, nonparametric techniques can yield useful aggregate output, which is the predictions. They cannot parse the strength of contributing factors in any rigorous sense, although some techniques have been designed to get at this somewhat indirectly, but they are usually classed as exploratory techniques.

And many parametric techniques can handle a mix of data type among the independent variables (categorical, ordinal, or continuous). I happen to be running a mixed model right now that contains both continuous and categorical variables. So, you stating that nonparametric techniques are especially useful in that context is incorrect. You are likely someone with a bit of research experience (perhaps an MS) that has a hammer in their toolbox and so thinks that everything is a nail, and is not familiar with the limitations of hammers or the physics behind them.

Accurately coding behavior and conditions is buzzword white noise. Stating that, at minimum, means you stayed awake for at least one hour in a basic stats course.

I would suggest you spend several years more of study into what you are attempting to discuss. As far as you not being able to compare to me in this field, well, I think that is rather obvious.

Dubv

I will add a caveat to my previous statement. Some nonparametric techniques (such as permutation tests) can be used to test hypotheses. That is because they work on a well-defined framework. For instance, I fit a regression model but the errors are not normal. The relationships in the regression are specified a priori and the distribution of the errors not being prespecified is the only limiting factor. In a case like that, something like resampling can get around this limitation somewhat.

In the case of your decision tree techniques, the relationships between independent and dependent variables are nearly infinitely flexible and driven by the data (and so post hoc). Therefore, they cannot be used to test hypotheses in any rigorous manner. I am currently writing a paper using ensembles of decision trees, and I would not dream of presenting a hypothesis test without a million caveats. I would be eaten alive if I did.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment