Statistics measure, tell stories and enable us to make sense of the world. In our data rich society only the statistically literate will survive!
What do statistics mean to you? To me, statistics are a language; the art and science of making sense of data. It has never been easier to gather and analyse statistics, so why does it seem that it’s getting harder to pin down the truth?
I have perhaps an old-fashioned view that statistical information can help us to understand then improve the world. This view is beginning to feel antiquated in a time where numbers are being chucked around willy-nilly with casual disregard for the truth, with one aim only – to simply peddle a message, grab a headline, or sell a brand. But is this a robust conclusion?
Let’s look at the evidence. Triangulation is a statistical technique that enables validation of data through cross verification from two or more sources. So, let’s examine some recent heavily reported events: the UK General Election 2015, the EU Referendum 2016 and the more recent US Presidential Election 2016. All three were brutal in their misinterpretation and miscommunication of data in order to prove an often unpleasant point. All three were heavily reported on social media, and for all three the eventual results contradicted most polls and data-driven forecasters. They were all also arguably system 1 elections for the most part, with the rhetoric pandering to the lowest emotional denominator. Neither the words nor the statistics appeared to matter, which is potentially why both were misused and often abused!
A lot has been written as to the whys and wherefores, which I won’t go into. It’s the pastime of statistical mudslinging that particularly gets my goat — it’s statistical bullshit! Is it ever fair in love, war and now politics to misuse statistics?
Unfortunately, statistical bullshit is in the ascendance, due in large part to the rise in social media, nature’s home for attention-grabbing headlines that rarely let the truth get in the way of a good story when stirring up emotions!
But really, it’s not that difficult to check statistics or facts, especially when there are independent fact checkers readily available such as Full Fact or Snopes, let alone your favoured search engine!
Statistics are not weapons to be lobbed at your opponent, but tools to help us make sense of the world around us. Facts still matter. It’s incumbent on all of us as intelligent human beings to be statistically literate. But what does this mean exactly?
Statistical literacy is essentially the ability to accurately find, use, apply, understand and communicate the story contained within statistics. Statistical literacy is a crucial competence in an information rich society. Both governments and businesses contend that these data skills are crucial, and that the provision of accurate and trustworthy statistical information strengthens our society. The Centre for Economics and Business Research estimated that 58,000 new jobs a year would be created in the UK in the big data marketplace between 2012 and 2017. Across the EU, ‘big data’ is predicted to contribute an extra £147 billion per annum to GDP by 2020.
Statistics are part of our everyday life, from the moment we get up till we go to bed. The amount of data produced each day is now in almost unimaginable quantities. It is produced by governments, the civil service, financial institutions, opinion polls, campaign groups, scientists, newspapers, brands — the list goes on. We can’t avoid it. Statistics surround us but, they are neither inherently true nor false. They cannot lie. But, like everything else, it’s how they are used that determine their validity. They can be hyped and sensationalised, misused (either deliberately or naively) and can be used to support contradictory arguments, or even lies!
What’s wrong with these charts?
We need to be very careful on how we interpret statistics. It’s up to us to work out where the truth actually lies. The cynic in me says take them with a pinch of salt until proven otherwise, especially those that sound too good to be true. They usually are!
Statistics are sly. If you’re not accustomed to their subtleties and deceits, you may be none the wiser. But, the bottom line is that the more statistically literate you are, the lower your risk of implementing faulty statistics in your own marketing. Those that get it will succeed. Those that don’t will get left behind.
How can you use statistics more wisely? How can you work out whether to believe them and what they really mean? You need to question the data.
- In what context was it collected?
- What was its intended purpose, and why it was collected?
- Am I clear on how it was asked, and how it was collected?
- How recent is the data?
- Is it from a credible and reliable source?
- Is it realistic given what else I know?
- Is it of suitable quality for my requirements?
Then you need to question your own biases, or at least recognise that they exist. Remember, we’re all heavily influenced by our inherent biases. We tend to believe statistics that confirm our personal beliefs, preconceptions, hypotheses or our view of the world, and are blind to or reject those that challenge these views, especially if made by people or companies we dislike or mistrust. This effect is even stronger for emotionally charged issues and deeply entrenched beliefs, especially if we trust the source. I should probably mention that I have a Masters in Statistics, so maybe, just maybe, I’m a little biased!
How statistics are organised can contribute to how they are interpreted. Tables and graphs are commonly used to present results. Tables provide the detail, while graphs show relationships. But, bad conclusions can be drawn from good data! What are seven common pitfalls to avoid?
- A correlation does not a causation make
- Comparing apples with oranges
- There is more than one type of average 
- Possibility is not the same as probability
- Confusing statistical and practical significance
- Blips are common and don’t imply a trend
- Knowing the difference between a percentage and a percentage point change
Finally, simplify how you communicate statistics. Break down big numbers and put them into a context your audience can relate to or find meaningful. Have a clear storyline and make sure that your message is spotted, remembered, and most of all understood!
 P.S. There are four – mean, median, mode and midrange!