How Statistics Abuse Make Us Lazy, Biased, Misinformed and Stupid: The Slate Dog Chart


A pet peeve (HAR!!!): computer geeks and statistics experts reducing complex issue into “simple” charts and graphs that have apparent credibility because of their form rather than their substance. I encounter this seductive form of fake erudition—“You can’t argue with statistics!”—in every field I explore: baseball, politics (Sorry, Nate Silver), social science, science (climate change models are a spectacular example), education. “Simple, straightforward” arrays of statistics that hide biases, dubious assumptions, projections, value judgments, undisclosed definitions, and who knows what else are presented to persuade on the false representation that they are “hard” representations of fact.  Very frequently, they are not, and when they are not, they incompetent, irresponsible and dishonest. Also arrogant to the core.

You could find no better example of this than this dog chart, by David McCandless, which purports to summarize “big data”—read: “data that can be manipulated to show whatever you want it to show” indicating which dog breeds are “over-rated,” as well as how they score on a “costs and benefits” scale. The fact that anyone could take such a garbage graphic seriously is unsettling, but of course, it will only impress people who know absolutely nothing about dogs and dog breeds. That’s what all such arrays of statistics are for: to convince and mislead those who are too lazy or uninformed to really understand the topic at hand and its complexities, but who want to lay claim to an “informed opinion.”

Just look at this monstrosity (you can read it better here):

Dog Chart

I think its fatal weaknesses should be evident to dog aficionados and the sadly non-dogged, but here are some hints. In the lower left hand corner, you will see how each breed’s “score” was figured:

Dog Chart Data Score

Please note:

  • Individual members of dog breed categories, like individual humans who fit a particular race, vary widely in all abilities and aptitudes. Designating one kind of dog as smart or dumb is breedism, and the logical equivalent of racism.
  • Scientists still don’t understand dog intelligence. However, they are finding that dogs, and many other animals, are much smarter than were once assumed.
  • Thus the first category is dubious at best.
  • Longevity is a big issue: if they lived more than 5 to 8 years, I would have a series English Mastiffs gracing our home. But the decision whether one can handle the grief schedule attached to living with the most loving, sweet-tempered creature imaginable is a matter of the owner’s individual temperaments and philosophy. I know giant breed owners who literally don’t care about longevity, because they love the breed.  Six years with a breed you love may be preferable to 12 years with a hardier breed that doesn’t melt your heart every time you look into her eyes. (I miss you, Patience.)
  • So much for the second category.
  • Ailments: some breeds have more characteristic health problems than others. On the other hand, getting dogs from reputable breeders and regular veterinary care can mitigate this issue. Again, the real variable is the owner. If one has the money to spend, the dog’s health problems are not a huge drawback or even a consideration. That is true of the cost factor generally.
  • Grooming is a matter of taste, and the maker of the graphic has a bias that those who like long-haired breeds don’t share. Long hair is a drawback, if you don’t like long-haired dogs. The real issues for many owners are shedding (which is not the same thing: short-haired dogs can shed copiously) and allergenic qualities, which didn’t make the “formula” at all.
  • “Appetite” simply adds a pro-little dog bias to the data. It is also a subset of “cost.”
  • Where is “exercise required”? Temperament? Trainability? Stubbornness? Sense of whimsy (yes, I own a Jack Russell)? Do they get up in the morning, or sleep until you get up (like Jack Russells)? Good for apartments, or best in farms? Are they hunters, workers, or lovers? Do they bark a lot? Do they make you feel like a king when you walk them? Are people irrationally terrified of them, tempting you to tell them that they are idiots? Many if not all of these factors would have to be part of any useful “dog formula,” if there could be such a thing.
  • Most obvious of all, the factors in the formula aren’t weighted, and for an obvious reason: the weighting would vary according to the priorities, values, needs, preferences and personality of the owner. Without that, this “big data” is worthless.

What can you learn from this chart? About dogs, nothing. You can learn, however, the way public opinion on matters that the average “stupid voter” doesn’t care a lot about or pay attention to can be warped and manipulated.

24 thoughts on “How Statistics Abuse Make Us Lazy, Biased, Misinformed and Stupid: The Slate Dog Chart

  1. Jack, this is one of those stories that you know people are going to comment with their own dog anecdotes. So here it goes…there was some concern when I adopted a dog many years ago because she was predominantly Chow. That dog was the calmest, smartest, most obedient and loyal dog I have ever had. She didn’t have an aggressive bone in her body. She was a great dog with a cool black tongue. The dog I have now…we still don’t know exactly what she is. She is definitely part miniature schnauzer but…seems to also have some squirrel and nutria in her. She hunts junebugs. She is very loyal but mischievous. Also, has no interest in playing with other dogs unless the dog is an obese male pug. She also enjoys playing with male cats. All dogs have their own personalities and charts like these are so detrimental because they lead people into believing that they are getting certain qualities in a dog they might not get.

    • Sharon: I had two dogs at once that were much like those you describe. The boy was obviously half chow and likely keeshond on the other side. The girl was half chow and probably a spaniel together. Both were rescues as puppies and both were exceptional dogs. I’ll never forget Charlie and Kelly!

  2. I’m a cat person anyways… but I agree wholeheartedly on the failure of statistics. I’d like to think it’s my math background that makes it easy to pick up, but half the time statistics are used it’s by someone with that same background.

    Statistics are sometimes useful, but frequently misinterpreted. My personal litmus test for understanding is Simpson’s paradox. Anyone who doesn’t know what it is well enough to construct an example is not qualified to interpret statistics and draw conclusions. That includes the vast majority of the media that reports on it.

    For those unfamiliar with the name: It’s possible for Group A to do better on average than group B on some measure, but for every single subgroup in B to do better on average than the same subgroups in A on the same measure.

  3. I find myself irritated both by folks who unconditionally trust statistics in the form of infographics and by folks who categorize all statistics as lies. Regrettably, I’ve dealt individual associates who unashamedly fell into both categories, loving pretty pictures and loathing math, thus choosing what to believe mostly on presentation. (This may be a coincidence, but most were also loyal fans of Slate.)

    If the raw data can be trusted (itself a huge assumption), I prefer being given the tools to generate my own conclusions. Despite what I’ve said about infographics, many would be well served by a tool that let them visualize breed comparisons based on factors important to would-be dog owners. (Some will abuse this knowledge, but what else is new?) No breed is inherently better than any other breed, but some are better suited to certain owners than others.

    • I think that most of us who’ve been around the block a few times are aware of how misleading statistics (even accurate ones) can be depending on how they are presented. It’s the same as with pictures. Both CAN lie if you don;t know the background or circumstances involved.

  4. AARGH!!! My brother-in-law (whom Jack and I both agree is an idiot) and I have had this same argument many times, and it almost always starts with his statement that “You can prove anything you want with statistics”, thereby demonstrating an almost total lack of understanding of the scientific method in general and statistics, specifically. The reason: you can’t actually prove or disprove anything with a legitimate use of statistics. Unfortunately, statistics lend themselves to illegitimate (and unethical) use by people who have an agenda, like the publishers of this chart. There are so many things wrong with it, above and beyond what Jack has described, I can’t begin to list them. Which, incidentally, is also what is wrong with most poll’s.

    • It depends if you can also see the methodology.I’m skeptical the moment methodology isn’t included, because I assume they’re attempting to snow me. But if the methodology is included, and it makes sense to me, I’ll trust the numbers. It also matters what they’re trying to prove. It’s AMAZING how many cause-effect errors and subjective problems the average study contains.

  5. I saw the bias for small dogs on the chart, but they also made stupid math errors. Ailments should NOT be a plus, that would mean a dog who develops a spontaneous case of rabies every six months would get a better score. Also higher costs should be a negative too. If the maker of this graphic really thinks that, I’d like to sell them a million dollar puppy as it would be beyond perfect for them. They tried to quantify why their favorite breeds are the best and went backwards to make it so. If there was any objectivity there should be a comparable number of dog breeds in every quadrant and not have clusters like the working dogs dumb and ‘rightly ignored.’

  6. I take issue with this being represented as “science”, rather than dumb entertainment. It would be funny to say the Saint Bernard is “overrated”; it is intelligence dulling to say these numbers mean anything.

  7. Breeding is eugenics gone mad. In less than 1,000 years we’ve taken one of nature’s most intelligent, agile, and resilient animals and have bred them into household accessories. Specific breeds aren’t more prone to ailments because of some evolutionary mistake, but rather because we bred in the mistakes over generations.

    The genes that give the dalmatian it’s iconic coat also makes it more prone to blindness. The genes that give the terrier it’s characteristic tenacity also makes them prone to obsessive compulsive disorder. Doberman’s, despite being extremely loyal, were bred for ferocity and, as a result, face a number of behavioral issues. And then there’s the chihuahua — humanity’s answer to creating a helpless child that will never grow up. “You know, I’ve always thought Down’s syndrome made for cute kids, I think I’ll find a few and play match-maker. Maybe I can open a farm!”

    Frankly, it’s horrific. Early on, dogs were bred for naturally desirable traits: strength, hunting ability, temperate attitudes, etc. Today, it’s little more than an abomination of natural selection. But, I guess none of that matters if it “makes your heart melt.” Repugnant.

    • I don’t see what’s repugnant about it. Is this some sort of creationism absolutist creed you’re espousing? Should we just have wolves, then? Dogs made their deal with mankind to support various needs in exchange for care, love and feeding. It isn’t unreasonable for mankind to customize an animal dedicated to serving humanity (at a price) for specific needs. Dogs promote mental health, sreve the disabled, assist the military, rescue kids (when their earlier incarnations would have eaten them)…and breeding dogs to serve sepecific needs makes perfect sense, and helps the dogs as well.

      I agree that some breeds, including one of my faves, the English Bulldog, meets your criteria for irresponsible eugenics—the dog is just physiologically unsound—as do the extreme giant breeds, which is why I won’t own them any more; it is not unusual for a Mastiff to drop dead of a heart attack at the age of one, because they get too big, too fast—but your generalization is overblown and unfair. Terriers suffer from OCD: be serious. That’s a human disorder, not a terrier disorder. That’s what makes terriers terriers. They are happy, functional, loving dogs who function very well (OCD is an impediment for humans.) They help treat depression in human. If a human can’t handle a dog that will bark for a year until you rescue his chew toy from under the couch, that human shouldn’t own a terrier.

      It’s a provocative comment, Neil, and I’m grateful for it, but except for extreme circumstances involving unethical close interbreeding and inexcusably toxic breeds like the Chinese Crested, it seems like a bizarre complaint. If you want to object to the AKC’s insane breeding for appearance, you’ll get no argument from me. The proper Jack Russell, however, still not an AKC breed, is given its papers according to behavior, not appearance: intelligence, perseverance, physical mobility, hunting instinct, temperament. No, they aren’t wolves any more. I don’t see that as a problem. and certainly not as “repugnant.”

Leave a Reply

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

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

Facebook photo

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

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.