The Art of Statistics: Book Review
I first became aware of Sir David Spiegelhalter after seeing him being interviewed on the BBC about the statistics behind the pandemic.
His clarity around the numbers and the compelling way he spoke about the data resonated with me and I discovered that he had written a number of books on understanding stats better.
David is also the Winton Professor of the Public Understanding of Risk in the Statistical Laboratory at the University of Cambridge. and is the current President of the Royal Statistical Society.
It’s safe to say that stats are his thing, then.
We’re highly aware at Customer Thermometer, that all our customers have a clear need to understand what their data is telling them.
It’s easy for all of us to be misled by patterns in numbers, either accidentally or through manipulation. It’s also easy to get cause and effect the wrong way around.
So this month, we chose Spiegelhalter’s book, The Art of Statistics: Learning from Data to review. While it’s not CX focused, some of the principles behind statistical analysis and pattern recognition, and the pitfalls therein, are incredibly useful for practitioners.
Especially now, when customer behavior is changing so much, and understanding pandemic-related data is critical to our business futures.
Before we kick off, despite statistics’ dusty reputation, it’s worth saying this is a fascinating book. Spiegelhalter has chosen really interesting real-world topics to make his points – ranging from who the luckiest Titanic passenger was through to how serial killer Harold Shipman could have been caught much earlier if different data was collected and analyzed.
So you’ll learn a lot about interpreting data but actually enjoy it. With that, grab your hand sanitizer and let’s get started…
The Art of Statistics: Top 5 takeaways
- We all need to be more data literate. We live in a world driven by algorithms, data, stats, headlines, social media and more. The way that data affects us all cannot be understated and yet the vast majority of us have no training in this field and it often appears impenetrable to the outsider. Headlines across all our publications are often totally wrong because the journalists themselves don’t understand the underlying statistics, especially the calculation of relative risk, says the author.
- Cause and effect are not as easy to untangle as it first appears. The very important point that ‘correlation does not imply causation’ is made by Spiegelhalter in chapter 4. He investigates a headline in a paper claiming that having a university degree can increase your chances of having a brain tumour, and goes on to show that it’s simply a function of wealthy people having a high level of education being more likely to get their brain tumour treated and therefore registered in the data in the first place. We see this a lot in CX, where a particular agent might receive low customer satisfaction scores, and upon proper enquiry it’s discovered that he or she is dealing with the most problematic product that company sells. Thus the feedback is about the product, and not about the individual, who may be doing a heroic job supporting a poor product in the first place.
- Presentation of statistics within sentences or descriptions can have a huge impact on how they are perceived. The sentences we wrap around our stats and data in our blog posts or presentations carry weight and power. How your data will be interpreted can easily be swayed one way or another… Spiegelhalter demonstrates this by showing that a ‘95% survival rate’ sounds more acceptable to most of us than a ‘5% mortality’ and yet they are the same thing.
- Data needs to be handled with care and awareness. We can also ensure this happens by carefully interrogating the data brought to our attention by others. We all have a role to play here; it’s important not to just take an output or a graph without considering what went in, and why things are so. As Hannah Fry says in her New Yorker article on the book, “In large groups, the natural variability among human beings cancels out, the random zig being countered by the random zag; but that variability means that we can’t speak with certainty about the individual—a fact with wide-ranging consequences.” This is why a fighter jet cockpit designed for the “average” person doesn’t fit a single one.
- The data and the stats will never be truly accurate This was one of the most useful takeaways of the book in my view. We all, I think, have the tendency to see stats as ‘absolute’. That if well researched and put together they are somehow “right”. However, the author is at pains to point out that statistics can point up findings but because the universe of data, the method of collecting it, the method of visualizing it and interpreting it all have inbuilt issues, they can never been entirely “right” and one must be acutely aware of the nuances. He says, “it is almost always an imperfect measure of what we are really interested in.” A really important point to remember when interpreting customer data. Hannah Fry again, says, “A mathematical analysis of what it is to be human can take us only so far, and, in a world of uncertainty, statistics will never eradicate doubt.”
The Art of Statistics: Top 5 quotes
- “Algorithms can display remarkable performance, but as their role in society increases so their potential problems become highlighted. Four main concerns can be highlighted: Lack of robustness, Not accounting for statistical variability, Implicit bias and Lack of transparency.”
- “With the growing availability of massive data sets and user-friendly analysis software, it might be thought that there is less need for training in statistical methods. This would be naive in the extreme. Far from freeing us from the need for statistical skills, bigger data, and the rise in the number and complexity of scientific studies makes it even more difficult to draw appropriate conclusions. More data means that we need to be even more aware of what the evidence is actually worth.”
- “…Many people have a vague idea of deduction, thanks to Sherlock Holmes using deductive reasoning when he coolly announces that a suspect. must have committed a crime. In real life deduction is the process of using the rules of cold logic to work from general premises to particular conclusions… But induction works the other way, in taking particular instances and trying to work out general conclusions.”
- “Whether the statistical work is good or not so good, at some point it must be communicated to audiences, whether fellow professionals or a more general public. Scientists are not the only people reporting claims based on statistical evidence. Governments, politicians, charities and other non-governmental organizations are all competing for our attention, using numbers and science to provide an apparently ‘objective’ basis for their assertions.”
- “Statistics can bring clarity and insight into the problems we face, but we are all familiar with the way they can be abused, often to promote an opinion or simply attract attention. The ability to assess the trustworthiness of statistics seems a key skill in the modern world…”
If you liked this, and you want us to send you the best CX and business thinking from around the web each month, sign up for our Lucky Sevens email here.
No spam, no nonsense, just the 7 best things we’ve seen, once a month.
Looking for more book reviews? Check out some of our latest here:
- Fanocracy: Turning Fans into Customers – Creating fans for your company who become loyal customers, from the pen of David Meerman Scott.
- The Ultimate Question 2.0 The original source material for the customer advocacy measurement that’s come to dominate global business KPIs.
- The Culture Code: The Secrets of Highly Successful Groups What makes a team more than the sum if its parts? Daniel Coyle investigates.