Remember your audience don’t understand Statistics

This month, we are going to focus on Statistics, starting with remembering many others don’t understand Statistics.

Appropriate use of Statistics is key to robust Analytics & Data Science work, as well as communicating the findings of that work. We have previously referred to a wide range of statistical techniques when sharing about Data Visualisation, Analytics or Data Science. However, it’s time to focus on this topic.

Now, more than ever (due to the coronavirus pandemic) it is vital that everyone understands the statistics that are being shared with them. As ever, a good place to start is to look to ourselves and what we can do differently. So, as well as reading on Statistics this month, I’ll focus this post on what we need to remember. How easily Statistics can be misunderstood or misrepresented even in our own work.

I hope that the resources I share can help you remember this challenge and be more careful in your communications.

Why it matters that you tackle Statistical Illiteracy

When speaking professionally about data topics I remember that the average numeracy level of UK adults is equivalent to an 11-year old. I suspect the average statistical literacy is much worse. Yet, too many leaders & analysts will pepper their communications with statistical terms or assumed understanding of probability theory.

Sadly, the current global pandemic has been a reminder that this can be a life & death issue. Accurate understanding of the R rate, the bases of forecasts or even confidence limits can be essential to policy decisions.

This article by physicist Carlo Rovelli, published in The Guardian, nicely summarises the importance of this topic. As he says, a better understanding of statistics and probability would help us all. You may not think that misinterpreting statistics in your business is a matter of life & death. But, pause for a moment & think, what’s the worst that could happen if the results of your analysis are misinterpreted? For many analysts, it could lose their business enough money to cost others their jobs. It matters.

How could you communicate Statistics better?

I hope that remembering the seriousness of this issue prompts you to want to communicate stats better, in a way that others can grasp. When seeking to improve any communication skills, it can often help to see & hear a proficient master of that art. Someone we can both learn from & identify some of the tricks that work for them when communicating effectively.

I can think of no better earner of that title than the 2020 winner of the Faraday Prize, Prof David Spiegelhalter. Even the wonderful host of BBC’s More or Less, Tim Harford, calls David the greatest living communicator of statistics (and interviews David in the video below).

A great way to observe & learn from David’s approach to communicating Statistics is to watch his Faraday Prize acceptance talk given in Nov 2020. In this video, he talks on the topic of “Communicating statistics in the time of Covid“. It is packed with so many useful tips to address the challenge raised in Carlo’s article & applicable to communicating statistics in your job too.

Are you also misled? How can you avoid that?

Whatever your level of statistical education, there is always a risk of being deceived ourselves. Perhaps we are tired or our guard is down, or we are trying to concentrate on too many things at once. So, it can help to have an initiative checklist of potential pitfalls.

Such an aide-memoire or check-list could also help us when advising others. Watch out for the risk that your slides confuse audiences or that your stakeholders misinterpret some of their own data. There is no perfect list to achieve that, but I’d like to share two resources that I’ve found helpful.

First, is a helpful guide from the Clearer Thinking blog. I recommend using it as a quick prompter for your own mental checklist. In it, Holly Muir shares 4 parts (each highlighting a number of common mistakes or tricks to spot them):

1. How can visual information mislead?
2. How can the way data was collected change the conclusion?
3. What common mistakes are made when interpreting data?
4. What questions will help you make sense of data?

The other resource I recommend is by the Data Visualisation communication expert whom Prof David Spiegelhalter cites in the above video, Alberto Cairo. Given the critical role of our presentation of data in helping others understand the key insights from any data, I am recommending a book that I have previously reviewed on this blog.

In the brilliant “How Charts Lie“, Alberto in his always engaging style, summarises the risk of statistical deception (via misleading charts) under these common threats. That is, Charts that lie by:

• being poorly designed
• displaying dubious data
• displaying insufficient data
• concealing or confusing uncertainty
• suggesting misleading patterns

Watching out for all of those threats can both help improve your data visualisation & protect your audience from some of the above statistical thinking errors. I recommend checking it out & sharing it in your business:

How are you overcoming a lack of understanding Statistics?

I hope this post helps kick off our focus on Statistics this month. How are you countering this problem in your business? What has worked for you in increasing statistical literacy amongst your stakeholders or team?

As ever, beyond the theoretical best practice, I am keen to hear from hands-on data & analytics leaders (as we do in our podcast). So, if you feel motivated to share your experience I’d love to hear from you.