“It’s science’s dirtiest secret: The ‘scientific method’ of testing hypotheses by statistical analysis stands on a flimsy foundation. Statistical tests are supposed to guide scientists in judging whether an experimental result reflects some real effect or is merely a random fluke, but the standard methods mix mutually inconsistent philosophies and offer no meaningful basis for making such decisions.”
Odds Are, It’s Wrong: Science fails to face the shortcomings of statistics
BY TOM SIEGFRIED MARCH 12, 2010
Read the whole article; it’s very good, and makes my head spin. If we can’t trust statistics in science, what is the basis for drugs being approved by the FDA, or the stated risks associated with everything my teenagers do, or how much pumpkin pie can I really eat before going over my daily limit of recommended fat?
That article was written in 2010, which you may think is old news, but How to Lie with Statistics has been around a very long time. That’s a book on how statistics can be used by companies, politicians, anyone with an agenda, to manipulate people with “hard facts”—a.k.a. math.
To be fair, Siegfried’s article is not about scientists purposefully fudging numbers, but how they generally don’t understand how to organize their data properly. They’re not bad guys (like certain advertisers), but still misleading the public because they don’t know what they’re doing. Maybe that’s worse.
So what can we do? How can you teach your children to understand the numbers thrown out in the media? Reading that the percentage of the U.S. population that will get the flu, on average, each year is between 5% and 20% on WebMD doesn’t mean much unless you understand how average is calculated (mean? median? mode?), what was the sample size, and maybe most importantly: who funded the research to get the data?
That last question may be a tough one to find. But understanding how statistics work, and how they can be manipulated, is doable and important. First, you might want to brush up on the knowledge. You could read Stephen King, but for something more encompassing, you need to stretch your brain. And no, I’m not talking re-reading the mind-numbing statistics textbook from college somehow still in your basement.
The free site Udacity has an accessible and comprehensive statistics course (my sixteen year old is taking it right now). There’s also The Manga Guide to Statistics, reviewed here, and the popular series take on it: The Complete Idiot’s Guide to Statistics. I like that one a lot.
The resources above are good for teens and adults. For the younger set, it isn’t hard to bring up how statistics work because they are everywhere, everyday. Next time a percentage of something is thrown out as evidence to do, or not do, (buy, or not buy), explain to your child what that number means, and how it could have been manipulated. Make them aware that just because math seems straightforward, using statistics may not be.
To make it fun, have them conduct a survey with their family and friends. It can be about anything (“what’s your favorite pie?” at Thanksgiving, for example.) Your job is to help them word the survey to get simple results. Then, help them make some fun graphs and play with their data. That hands-on manipulation is the best way to learn. I did it with my own kids. We asked people about tea! You can check out all our fun graphs here.
Hopefully, by the time they become scientists, their understanding of organizing data will be enough to trust their research! Especially if that research tells me I can eat all the pumpkin pie I want. Let’s all become more science literate by understanding how statistics works.