Statistical ‘sleight of hand’ abounds

17th March 2017 By: Kelvin Kemm

The statistician-general of South Africa is Dr Pali Lehohla. We have known each other for years and have engaged in many interesting discussions.
Recently, when commenting on statistics in the modern world, he made the point that, with modern technology, data is now available in unprecedented amounts and that this is termed the Data Revolution.

However, data is reasonably useless stuff. Lehohla said: “Data is unstructured observations and you require statistical science, which, in the main, deals with the collection, classification, analysis and interpretation of numerical facts or data.”

What he was pointing out, and what I have said before, is that data is not much good until it has been interpreted to extract from it the information that resides in it. This is something like sifting beach sand to find coins that may be hidden in it.

Lehohla stated that, over centuries, statistics has emerged as a science-based knowledge system. Unfortunately, many people forget that important fact. It is not the case that anyone who can add three numbers and then divide by three is a statistician.

Imagine a nursery school class of 19 children who are all aged six. They have a teacher who is aged 26. Now ask the question: What is the average age of all the people in the classroom?

There are 20 people in the classroom, counting the children and the teacher, and their average age comes out at 7. But what use is that bit of information? The figure of 7 is mathematically correct, but none of the kids is 7 and the teacher’s age is far away from 7.

This simple example illustrates the type of unreasonable things that some people do with data. That is why data is not much good unless it has been professionally analysed by a statistician, who knows how to interpret the data.

Lehohla also warns of the tendency to project data analysed by nonprofessionals with the objective of projecting sentiment or prejudice. He warns of the danger of an “unleashing of tabloid editors and demagogues to provide their own ‘truth’ of what is going on across society”.

Lehohla mentions the term ‘post-truth’ politics, which refers to the declining authority of statistics – and of the experts who analyse them – in favour of anyone who mentions his or her own home-made stats.

Some of these types of stats are very transparent to those who know what to look for. It is a sad state of affairs.

People say things like: “This electricity source will provide enough electricity for 100 000 homes.” But they do not tell you what a ‘home’ is or what it consumes. Sometimes they are slightly more cunning and say ‘average homes’. This can be an ‘average’ like my kindergarten class illustration.

Another stunt is to say that an electricity source produces a number of kilowatt hours of electricity a year. This is a quantity, not a rate. Electricity is sold on a rate basis, not a quantity basis. This is like saying: “Joe drinks 150 beers in a year.” That sounds okay? But now, if I add that, actually Joe only drinks beer five times a year, and each time he drinks 30 beers, then that changes the picture a great deal.

Quoting the electricity ‘amount per year’ is a trick to cover up the ‘rate’ aspect. The ‘average house’ comment is also a trick, and it is intended to make you not realise that it is not your house that can be powered, but only one that has two rooms and three light bulbs.

Every time that I see the ‘number of houses’ sleight of hand, I whip out a pocket calculator and check and, sure as nuts, it is data manipulation designed to fool the reader.

Lehohla says that an international team of experts compiled a stats document called the Cape Town Global Action Plan, which is scheduled to be tabled at the United Nations Statistics Commission this month.

It is said that some people use statistics like a drunk man uses a lamp pole – more for support than for illumination.