Lies, damn lies and statistics

I would like to demonstrate to you how unreliable statistics can be and why many of the figures quoted in the recent programme Pedigree Dogs Exposed are unreliable

The programme flung statistics around like the cream pies in an early silent movie. I was told that the programme had been ‘meticulously researched’: if this was the case some indication of how unreliable statistics actually are should have been included. It is not just that they are subject to misinterpretation but as the collection of data which make up the final figures can be so very different the results are almost always incompatible with other sets of figures. You can listen to  this talk at

As an example of how inaccurate statistics can be, let us look at the numbers of dogs that are estimated to reside in Britain. The Pet Food Manufacturers Association (PFMA) website states that in their latest survey there are 7.3 million dogs in the UK in 2008. They say 75% of those dogs are pedigree dogs (although do not define whether that means those dogs are registered with the Kennel Club), 11% are ‘cross breeds’ and 14% are ‘mixed breeds. My immediate reaction is that 7.3 million is not a realistic figure and 75% being pedigree dogs is very unlikely. Why do I think this? Well, the PFMA researchers also asked where the dog owners obtained their pets. 16% said ‘from a private advert in a newspaper’, 7% said ‘from a pet shop’, 8% said ‘through the Internet’, 25% said ‘from a friend’ and 16% said they obtained their dogs from a breeder. The rest, 32% said ‘from a rescue centre’. Now this is where the high total may come from – the number of puppies available from rescue centres is actually very low and as most dogs are adult before coming up for rescue many would be counted twice and some might be ‘recycled’ several times. What is more, dogs from a rescue centre are allocated a ‘breed’ by the staff so many who look a bit like, say, a Border Collie or a Border Terrier would be counted as pedigree even if they had no papers.

There is more. In 2007, the Kennel Club registered 270,707 puppies. In relation to 7.3 million this means that the registered pedigree dog population is about 43%! This is long way short of 75%. There will, of course be many dogs which are classified as having a ‘pedigree’ by their owners even though they are not KC registered but is it likely that they comprise 32% of the canine population? A much less scientific method is to observe the dogs appearing at your local veterinary surgery. You will find that the numbers of pedigree compared to mongrel dogs is about 50/50, and this probably reflects the proportion of pedigree/crossbred dogs in the county as a whole.

If there are, in fact about 7 million dogs and the average life of a dog is, say, 12 years (the KC survey suggested that of the 30,000 dogs in their report into canine morbidity in 2004 the average age was 11.3 years) this means that approximately 600,000 puppies are born each year and, of course, 600,000 dogs die. Let us now look at the dogs which die each year. We have seen that this is likely to be around 600,000. It is likely that most die (or are put to sleep) because of infirmity or old age. We can have no way of knowing how many are put down because of illness when they are not old but we do have some figures put out by the RSPCA and the Dogs Trust about dogs put to sleep ‘unnecessarily’.

The RSPCA’s latest figures for 2007 say that they put about 6000 dogs to sleep for humane/medical reasons – i.e. they were ill or old. And just over a thousand were put down because they could not be re-homed i.e. they were likely to be a danger to the community. As a proportion of the total number of dogs handled by the RSPCA this is not unreasonable. Putting a dog to sleep is regrettable, of course, but it does not seem an unreasonable figure although it is seriously at odds with the headline grabbing press releases sometimes put out by the organisation which incidentally added £23m to its reserves in 2007.

Dogs Trust began its own Stray Dogs Report’ about five years ago. This report says that in 2007 nearly 100,000 strays were collected by local councils of which 6,700 were put to sleep for want of a home. That is an average of 18 each day but it also means that over 92,000 were, presumable, safely returned to their owners or were still in the care of the local authority at the end of the year.

Quite how these figures are collected is unclear as, in general, Dogs Trust does not take in strays collected by dog wardens although when they have room and the dog is re-homable (for the charity boasts it never puts a healthy dog down) they may take dogs in after the seven day period. On the other hand some RSPCA centres do so – as do Battersea and the city based dogs homes such as Birmingham, and Manchester – so again, some of the dogs might well be counted twice.

Of course, no one has yet begun to define what all these terms mean and each organisation has different methods of compiling the figures so it is virtually impossible to work out what the proportion of strays there are in any given year, what proportion of those strays are put to sleep because no home is available and precisely what proportion of dogs are killed compared to the number who die naturally or are humanely put to sleep to save them pain or stress. There is no doubt there are ‘problems’ with a proportion of all dogs in society but defining that proportion is extraordinarily difficult.

How much more difficult it is then to collect meaningful statistics about genetic abnormalities. The Kennel Club’s research into the reasons why dogs die, makes almost no mention of genetic disease – because, by and large, genetic diseases are not in themselves life threatening. If they were, they would not in any case, as the dogs would die before they could reproduce.

We can certainly identify many genetic conditions but because they are identified does not mean that they are necessarily debilitating. In fact, some may be little more than an inconvenience. Many humans live with being short-sighted, have mild allergies, asthma or are hard of hearing (all of which have a genetic component) with little or no inconvenience. The Cavalier King Charles which went BIS as the CKCS Championship show might well have some indication of disease on a CAT scan but it looked calm and composed and cannot be compared to the poor creature shown on the programme which was clearly in considerable distress.

In its introduction to its Purebred Dog Health Survey (2004) the Kennel Club says ‘The results of this survey and particularly the breed-specific analyses should be interpreted with caution.

This sensible warning was totally ignored by the Pedigree Dogs Exposed programme if it was even looked at. Almost all the statements in the programme were vague. ‘Up to’, ‘estimated’, ‘probably’ and ‘may’ were words attached to any figures. The statement that ‘up to 30% of Cavalier King Charles Spaniels may have the condition’ is not in inaccurate in itself but as the KC’s ‘estimate’ is that the condition ‘probably’ affects about 2% of the breed, it is hugely distorted.

What we do know is that one of the most knowledgeable geneticists in the world of dogs, Dr Malcolm Willis has stated, categorically, that the inbreeding coefficient in the UK pedigree dog population as a whole is in the region of 4% percent and that less than 1% is the result of a parent to offspring mating. This is a breeding coefficient of 25%. To put this in perspective, a cousin-to-cousin alliance, which is not at all uncommon in this county and overseas, has an inbreeding coefficient of just 6.25%. These are the facts.

So any statistics should be handled with care – perhaps all summaries of figures should have a government health warning!


Explore posts in the same categories: Uncategorized

One Comment on “Lies, damn lies and statistics”

  1. davidcavill Says:

    Can you be more specific Are you taking about the way statistics are used to distort the facts or the ‘facts’ that were distorted in the programme to which this article refers?


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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

%d bloggers like this: