Statistics flatter only to deceive

The jury is still out on who is the real owner of the quote: “In God we trust. All others must bring data”, but in all probability, it would be a statistician. ‘Stats’ explain the world by adding accuracy and certainty to its affairs. A news story that says the coronavirus impact will shrink global GDP by 1 per cent would add a lot more value than a headline that says ‘Covid-19 to dent global GDP deeply’.

Numbers are charming when they fall in the right place, and can be alarming when they get cocktailed with bad news. For policymakers, planners and businesses, numbers are crucially sacred. They complete their story. Plato may have said a good decision is based on knowledge and not numbers. But number mavens and data doctors won’t agree. For them, numbers never lie. They represent the truth. Two plus two equals four, not six.

In recent years, the advent of modern computing and the rapid emergence of allied segments such as big data analytics has added an extra layer of intrigue and many layers of intricacy to the world of numbers, making rapid number-crunching a glamorous vocation. Statisticians and mathematicians are today paid more than ever. And their services transcend disciplines. Some statisticians are like Francis Galton, a cousin of Charles Darwin and an obsessive number lover, who said: “Nothing in human nature is indeterminate. Anything and everything can be measured.”

Understanding numbers

In the post-truth era, where (manufactured) numbers (data) become tools for propaganda and spreading hatred and fake news, making sense of the numbers one encounters every day in media (mass and social) and other walks of life becomes extremely important, as the numbers today hold the potential to make or break careers, social harmony, policies and other programmes. There is a number for every occasion, idea and individual. From Wall Street-Dalal Street products to Unique IDs to most recently Covid casualty data released and crunched by multiple agencies, numbers are omnipresent — and are becoming omnipotent in human life. Far from adding accuracy, numbers are now used to control, conquer and con.

In this context, understanding statistics becomes an essential evolutionary skill. In Something Doesn’t Add Up: Surviving Statistics in a Post-Truth Age , mathematician Paul Goodwin offers us an exciting handbook to understand the world of numbers. “I now realise that my mathematical colleague who argued that there are two worlds — mathematics and waffle — was wrong,” writes Goodwin in the introduction to the book. “There are two worlds, but they are the world of reality and the world of numbers. And the second world is usually at best a simplification of the first and at worst a gross distortion of it.”

Goodwin’s book has 11 chapters, apart from a pithy introduction and an appendix which details a few questions one must ask about a statistic. This is a well-written and a well-edited book. Clearly, it targets everyone — those who love and believe in numbers, those who don’t and those who don’t give a damn about statistics; because today, you may love or hate numbers, but you can’t ignore them. They are like algorithms. “We should be particularly concerned when the world of numbers is a serious distortion of the reality it aims to represent,” writes Goodwin. He says this distortion can be deliberate — marketers, politicians and campaigners often exploit a range of numerical conjuring tricks to sell us untruths.

But not many are bothered; for most people numbers are “uninteresting, untrustworthy, mysterious or indigestible”. And the result is they accept misleading numbers or refuse to accept useful numbers. Goodwin’s first chapter — Rank Obsession — details many examples of this fact. Goodwin exposes our deep penchant for ranks and scores and illustrates why in most cases they don’t mean much, and weightage and points are just a matter of absurd numerical tweaks that may reflect biases of those who prepare the numbers. Most models are infested with inaccuracies and errors, which often a commoner’s eyes may not catch. “Sometimes ranking is a matter of judging whether an apple is better than an orange,” notes Goodwin.

The truth about stats

Each chapter in this book is a goldmine of wit and wisdom. This reviewer’s favourite is the second chapter — Perilous Proxies — where Goodwin discusses the art and craft of using proxies.

For starters, in statistics, a proxy is a variable which is used to measure an “unobservable quantity of interest.” The variable is not directly relevant to what it is used to explain.

The gross domestic product (GDP) is an example. Goodwin discusses how the GDP has ceased to become a clear metrics of calculating the size and strength of an economy owing to grave inconsistencies and anachronistic errors in measuring it — such as where “the outputs of industries that pollute are valued on an equal basis to those of non-polluters”. These days, people use Google, Facebook and Wikipedia to plan work, to communicate and to travel. “Years ago these activities would have cost money, but now, because they are free, they do not register with GDP,” notes Goodwin. “We are generally better off because we have access to these services, but GDP ignores them.”

There are many similar examples. The book busts several myths around numbers and stats while maintaining the utmost respect for data-based, fact-based scientifically-charged enquiries. This is not a cynic’s manual. This is an intelligent person’s guide to the baffling world of numbers.

If you are one of those hapless souls who burn midnight oil vainly trying to crunch Covid numbers released by governments, media and every other stakeholder in the calamity, to arrive at good-for-nothing, acid-reflux inducing numerical results, your therapy should start by reading Goodwin’s book. There is a 50-50 chance of your worries vanishing, because nine out of the 10 things he says in this book are just great.

Grab your (e)copy. This is one of the best things you may do during quarantine times.

COMMENT NOW