As we move from one crisis to another one common factor is the importance of numbers. But another common factor is the way they are routinely misunderstood by politicians, journalists, news presenters and others who seek to influence and report on events. Indeed, I blogged on this quite recently in my blog on Innumeracy.[i] But I have just read an excellent book which should be required reading for all these people. It’s only 137 pages and I read it in one sitting. But it covers this ground quite simply and effectively. Sanne Blauw is the numeracy correspondent for Dutch online news site and publisher De Correspondent. She has a PhD in econometrics from the Erasmus School of Economics and the Tinbergen Institute. The Number Bias, her first book, was a bestseller in the Netherlands and has now been translated into English.[ii]

She sets out to demystify the world of numbers, so that everyone can distinguish when they are being used correctly or when they are being misused. And so that we can all ask: what role do we want numbers to play in our lives? She argues that it’s time to put numbers in their place. Not on a pedestal, not out with the rubbish, but where they belong: alongside words.

We have been counting since time immemorial. The oldest written messages handed down contain symbols that refer to numbers. Hunter gatherers would have been able to remember all the information they needed but as the economy became increasingly complex, money was introduced instead of barter, it would have become impossible to remember everything. As states needed to levy taxes from large numbers of people, officials needed methods to record all incoming and outgoing payments. In the same way that you cannot understand each other if you do not speak the same language, you cannot enter into agreements if you use figures in different ways. For centuries there were countless different methods of measuring units. Now almost every country in the world uses the International System of Units but that standardisation has taken a long time to develop. Today throughout the world people speak of metres and kilograms, GDP growth and IQ points, CO² emissions and gigabytes. And so, the most widely spoken language in the world is not Chinese, English or Spanish but numbers.

These days you can’t open a newspaper without coming across a chart but the idea of casting figures into images is relatively new. Bar and line charts were only invented by William Playfair at the end of the 18^th century. Florence Nightingale used his ideas to draw attention to the grievous situation in army medical care because charts could explain a mountain of numbers in an instant. It was only in the 19^th century that the idea of the average was developed to apply to human beings, though this is a concept that is often badly abused. If Bill Gates gets onto a bus than the average person on that bus is a billionaire.

Other medical pioneers used statistics like Nightingale to save lives. But there is another reason why numbers are important. They help to keep rulers in check. But the ruthless abuse this. For years corrupt governments in Argentina ordered inflation rates to be manipulated. Stalin had a statistician killed because he said that the population of the Soviet Union was smaller than Stalin claimed. An independent statistical agency can prevent politicians from using numbers to serve their own interest.
 
The idea of the Intelligence Quotient was first developed in France and was intended to help children by developing a method to determine which children needed special education. Alfred Binet, assigned this task by the French Minister for Education, created a test with exercises of increasing difficulty; the last question a pupil was able to answer correctly would indicate their ‘mental age’. If this age were far below his or her true age, the child would require special needs education.  But the concept has been used more to discriminate than to help.

When the United States came into the First World War, 1.75 million recruits were given an intelligence test. Once the data were compiled and analysed the results were dreadful. White American men had the mental age of a 13-year-old; immigrants from eastern and southern Europe scored even worse. Right at the bottom black Americans had a mental age of 10.4.

For some this confirmed long-standing prejudices and would endure for decades. But the research had been hastily conducted in a slapdash matter, the rooms in which recruits took the test had no furniture, were badly lit and it was so crowded that you could not hear what was being said.  Many of the recruits did not speak English, nor could they read or write but of course this only emerged much later. The consequences of the research meant that quotas were applied keeping millions of people outside American borders. It even became legal in 1927 to sterilise someone by force and this practice was not outlawed until 1978.

Increasingly numbers were applied to abstract concepts. Even money is one such as the coins and certainly paper notes have little value in themselves but we all attribute value to them and hope and trust that the politicians will stick by the rules, which, of course, they frequently do not. The idea of measuring a country’s economy is relatively new, developed in the United States in the years leading up to the Second World War. The government knew the country was in a deep depression but had no real measure of it. First attempts to measure national income focussed on households and companies. But as the war approached and the government knew it needed to spend money on arms this would have meant a drop in national income. The solution was found in a different measure, the Gross Domestic Product. This would measure the total value of all goods and services produced in the country, including those generated by the government.

These days politicians and policy makers tend to forget that GDP is an invented concept and use it as an objective measure, for instance when the government uses GDP to argue for cuts.  GDP is not a concrete measurement like gravity. You do not make it anymore ‘real’ by sticking a number on it. For many governments economic growth is their main objective, but it may not be what people really value. A polluting industry, for example, is good for GDP, but bad for health and the environment.

What you measure is what you can count, but this may not always be the most relevant. As Albert Einstein is rumoured to have said ‘Not everything that counts can be counted, and not everything that can be counted counts.’ When the New Labour Government decided that people should be seen in A&E within four hours, there was widespread manipulation of the targets by hospitals. People were kept in ambulances for longer and checked-in post haste to meet the deadline.

What you measure will eventually be captured in one figure. This happened with emphasis on GDP as a proxy for a nation’s well-being, but the development of a country is about more than money. People should have access to good education and reliable health care, among other things. This led to the concept of the Human Development Index, which today is a popular measure to gauge a country’s development. This looks at three factors: life expectancy, number of years in education and income. In 2018, Norway came out on top with 0.95; the UK was ranked fifteenth. But here again a complex concept is reduced to a single number.

A number should not be the end of a conversation, but a starting point. A reason to keep asking questions. Which choices have been made during the research? Where do the differences come from? How do they affect policy? And especially: does the number measure what we believe is important?

Market research, including opinion polling, is a highly complex area with many examples of flawed research leading to wrong policy decisions and even potentially influencing election outcomes. Ms Blauw points out a number of difficulties:

1. The circumstances or questions are flawed
2. The survey excludes particular groups
3. The interview group is too small
4. Too few people want to take part
5. The margin of error is overlooked
6. A particular outcome matters to the researcher.

In my next blog I will continue to summarise this important book covering the criminal conspiracy of the tobacco industry to deny that smoking causes cancer for over fifty years; the challenge of big data and how the concept of algorithms is self-defeating to the point that even the programmers do not understand them; and how our psychology decides the value of numbers.

---