Early last year I wrote a post on ecology and mathematics that was inspired by an online discussion happening at the time. Although comprehensive advanced maths skills are not essential to being an influential or inspiring ecologist, a good level of mathematical knowledge and understanding of statistical analysis is definitely necessary to create honest science and communicate the importance of your work to others.
But it’s not just ecologists who need mathematical common sense. Anyone who deals with, or is interested in science needs to understand the ambiguity of an average, or the difference between a regression and a correlation. In fact, anyone who cares about the society they live in should be aware how deeply statistics and data now influence the way we live – policies and decisions on anything from what product choices you find in retail stores to how much tax you pay are all based on data.
Why does this matter to us? Well, if those data are a bit dodgy, or haven’t been analysed and presented appropriately, problems arise. And when these kinds of data misrepresentations are used to fuel public opinion or inform government policy, there can be serious impacts on communities, individuals and ecosystems.
We saw it with climate change deniers, the tobacco industry, and the UK government’s attempt to prove neonicotinoids weren’t killing bees. There are many other examples, ranging from the nasty to the plain ridiculous – the dodgy data that triggered the implementation of government austerity policies around the world, tobacco companies trying to avoid plain packaging, the coal seam gas and tar sands industries, and the Australian government trying to withdraw World Heritage listing from one of the last wilderness regions on Earth. Usually the people that are most affected by these mistakes or misrepresentations are the people that have the least understanding of what went wrong.
How did this divide between research and the people that it ultimately affects, become so wide? Is it a result of poor education systems with mediocre science curricula? Is it because researchers don’t have time or knowhow to communicate beyond their own immediate colleagues? Or did we not even realise this was happening, because we put our trust in the Media to be the go-between, without realising what an increasingly poor understanding of science even journalists have?
So is the onus on schools, universities, teachers or supervisors to teach students these skills, or on students, scientists and anyone else to educate themselves?
One could argue for both sides of the story. Education systems are meant to educate, and some curricula do miss the mark…but ‘education’ is also a two-way street. Students need to be able to receive and understand the information to learn it…and sometimes that receptiveness comes after school/uni, along with experience and application.
I didn’t understand statistics after finishing my undergraduate degree, because I had only been exposed to Statistics the “software program”, not Statistics the “philosophy”. It wasn’t until I started my PhD that I discovered there was far more creativity and excitement to statistics than the limited world of R, linear regression and p values. That approach to science isn’t ‘wrong’, it’s just not the only way to do it.
There isn’t one statistical language, but numerous statistical dialects, depending on whether we are talking about sport, wildlife populations, clinical trials, finance or climate change. Just as the key to speaking a language is in understanding its alphabet and grammar first, the most important part of understanding statistics is knowing the context of the raw data, including how they were collected.
Of course, it is not possible to teach everyone advanced statistical fluency, and most people wouldn’t be interested anyway. A lot of scientists don’t even know how to analyse data outside of their own field of expertise.
But you don’t need to be a scientist, or even maths-fluent, to know if you’re being fed dodgy data. Just like you don’t need to be fluent in Spanish, Italian and French to know what hola, bella, and mademoiselle mean. We – scientists and non-scientists alike – just need to understand how to make some critical decisions about the data that are presented to us.
University and school curricula, from all disciplines, can equip students for this better by spending more time teaching mathematical and statistical philosophy and skills in understanding and interpreting the context of the data, rather than focusing on using a particular software program to perform a subset of ‘vanilla’ analyses.
Beyond the education system, there are plenty of journalists and writers to watch who do know how to spot bad data, like the ABC’s FactCheck and The Checkout, the Washington Post’s Fact Checker, and writers like Simon Singh (Trick or Treatment?), Michael Pollan (In Defence of Food, The Omnivore’s Dilemma), Guy Pearse (Greenwash), Ben Goldacre (Bad Pharma, Bad Science) and Andrew Nikiforuk (Pandemonium).
Mathematics is a foundation of life, not a minority pursuit exclusive to scientists. It is much more than knowing how to multiply and divide, although even mistakes at that level can have nasty impacts. If you care about health and nutrition, national parks, budget cuts, taxes, the availability of medical treatments, or conserving biodiversity, you need to care whether mathematically-sound, honest data are influencing your opinion.
© Manu Saunders 2014