This blog has MOVED!

Please visit for the most updated content. All these posts and more can be found over at the new URL.

Saturday, November 26, 2011

A pet peeve

Excuse me while I step up to the soapbox.  

There are certain companies that continually use *statistics in a way that I feel is misleading and slightly ridiculous.  I value transparency and disclosure, and running a business in a way that is above reproach.  This is why, even though I think the average consumer probably sees through such silly antics - it still annoys me.  Really really really annoys me.  

*Really, it's stretching it to even call it statistics - probabilities?  percentages?  the notion of confounding factors?

What exactly am I referring to?  Let's use this as an example.

Suppose I say that 75% of Renegade boot users finished the 2011 Tevis this year.

This is in fact a true statement.

Considering that the Tevis had an overall completion rate of 69% (abnormally high due to some changes that took place on the course due to snow in the Sierras), you might conclude that riding in the "majik" boots give some sort of advantage......

Now how about I tell you that of 4 riders using Renegade Boots, 3 finished.

I am in fact, telling you the same information (and, to the best of my knowledge, my numbers are real-life correct) but it doesn't sound nearly as impressive eh? 

Comparing percentages between 2 different groups (all tevis finishers, tevis finishers in a certain hoof boot) imply a big enough sample group to make such a comparison valid!

And of course - we are leaving out all sorts of other factors - whether people who decide to ride Tevis in boots have prior experience on the Tevis trail, whether they are being paid or sponsored to ride in the boots etc.

When I have data that could be interpreted a myriad of ways, I chose to give people the raw data so they can come to their own conclusions.  When given a number that has been calculated, I am putting a certain amount of trust in the group that is presenting the numbers, if raw information is not being given - after all, there are lies, there are damn lies, and there are statistics.

Let's try a couple more of examples -

At tevis, I have a 0% rate with shoes, 100% rate with boots.

Thus boots are better than shoes.  

Compare my overall completion rate with shoes [76%], and without shoes [91%]

I am almost 20% more likely to finish a ride if I am in boots (although one of those shoeless ride completions was done truly barefoot)

Or Farley’s rate with and without: (80%), (90%)

Farley is more likely to finish a ride if she is in boots (ignoring the fact she has twice as many rides in boots than shoes)

Do you see how ridiculous this is? 

Are these statements in fact, completely true?  Yes

Are these statements misleading?  Yes

Each individual has a responsibility to make sure they aren’t being duped by the statistics or calculated numbers - BUT companies that continually provide misleading statistics are a personal pet peeve. 

I don’t like feeling like a sucker, and I doubt you do either.  So please, think critically and don’t take ANY statistic or calculated number at face value, and please, unless you can check the source and see the numbers behind the calculations, and assumption etc., don't pass along stats that are misleading - Thank you!

FYI - A discussion very like the one above is continually re-visited in vet school, as we spend a substantial amount of time discussing journal articles and the statistical significance of their results.  In the foundations block, we discussed how companies manipulate stats and numbers (with real-life examples and products...) for an entire afternoon as part of the foundations block.


  1. Well said. This is something that infuriates me as well and is seen all too often in nutrition studies sponsored by companies. h and let's not forget that they generally run 3-10 studies before they get one with numbers that they are willing to report. All those other studies end up in the ether...

  2. In library research, I've noticed that at least 40% of statistical numbers are completely made-up.


  3. AareneX - you are naughty naughty naughty!!!!!!!!! :)

  4. *like* Aarene

    I'm so glad to hear that yall are taking a long hard look at this in vet school!

  5. Have you seen the book "How to Lie with Statistics?" It's a super short, funny read, but it goes over things EXACTLY like this. It's an especially good book if you're in an advanced degree program, reading peer-reviewed articles that spout all sorts of stats - you'll be able to look at the numbers with a bit more skepticism. Or if you're just cynical, like me.

    'Cause like 95% of equestrians are cynical, you know.

  6. It seems like I've heard of the book as a recommendation before. I will definitely look it up - it seems like it would be a relevant and fun read!


Note: Only a member of this blog may post a comment.