We live in a world of ​‘post-truths’, where it is increasingly difficult to separate fact from spin. 

We assumed that being data driven would reduce biases and help us make better and more informed decisions. 

However, more data and more sophisticated analytical tools have not helped to the extent we would have liked, particularly in the humanities. 

Instead, selective representation of data, facile interpretations and convenient extrapolations have become the norm, all designed to fit a narrative. 

I have recently come across several instances where the same data was used to shore up two opposite arguments. 

1. Is this the best use of human intelligence?
2. How does Mathematics help us get to the truth? 

The elegance of Mathematics, at least the parts that I read and understood in school, is that it is built on a set of foundational truths that are universally true. That 2 + 2 = 4, or that a circle can never be a square, is my limited understanding of Mathematics.

Euclid’s system, the basis of geometry taught in schools even today, is a set of principles (axioms) that are absolute truths. I do not claim an understanding of non-Euclidean Mathematics where this belief could come under question.

Why do we then hear that often numbers represent half-truths? A recent tweet which I thought was quite insightful said, ​‘Truth is almost always the combination of two opposite half-truths’.

An HR friend of mine once told me in half jest that there are four kinds of lies in increasing order: lies, dark lies, statistics and résumés. That statistics comes second worst in this ignoble list is a sad truth.

Why then does data (which is essentially numbers across time) not lead to the universal truths that we see in the real world?

First, data needs to be backed by rigour. It should lend itself to scrutiny and proof. Theorems in Mathematics are accepted only when they carry proof. In fact, when Sir Andrew Wiles, often considered to be the greatest mathematician alive, proved Fermat’s Last Theorem, it was rejected the first time. He had to return with the correct proof two years later.

How much of the data we see in today’s surveys and research reports can we vouch for? Often, the statistical rigour is compromised or tailored to suit a narrative; for example, the sampling method, the sample size, the way questions are structured, and data collected across time. The controversy on the new GDP series a few years ago is a case in point. Perhaps, oversimplification is another aspect of lack of rigour. Most of us could be guilty of this. For instance, while assessing performance of people, I often see managers making facile arguments of extraordinary or severely abysmally poor contributions based on shaky data, only to support a recommendation.

Second, we encounter several issues in the presentation of data. If you torture data long enough, it will squeal what you like to hear! For example, the deliberate mixing up of correlation with causality, slicing the time series to show convenient truths and presenting comparative figures only to suit a point. Often sophisticated data charts are no more than just optical illusions.

During the COVID-19 crisis, when real time data was a problem and historical data did not exist, mathematical modelling such as ​‘the R factor’ were used to project trends. While it undoubtedly helped in planning a response, the spread of an epidemic depends on far too many complex factors, and any mathematical model has limitations. Should the presenters be more mindful of this?

Third, there is spin. This takes many forms. For instance, quoting only from selective sources, exaggerating a trivial point because it fits the messaging and ignoring key findings that are inconvenient. The debate on ​‘jobs data’ in the recent past was an illustration of how truth could be perceived differently, depending on which end of the political spectrum you were. In addition, demagoguery. A wise man once said, ​‘A lie told over and over is perceived as truth’. And we have leaders in business, society and geopolitics who have the capacity to convince anyone, anytime on anything, leading to catastrophic effects.

It is utopian to believe that we will have one version of the truth always. However, the mark of an educated mind is to put data to test and search for the truth.

A sound knowledge of Mathematics helps, and our schools can prepare children to learn to question, become positive sceptics and not rest till they arrive at the truth.

I heard Uday Kotak, the celebrated banker once say, ​‘If something is too good to be true, it is too good to be true’.

We would like teachers to enable students to relentlessly pursue truth using Mathematics as an ally and At Right Angles (AtRiA) is a humble endeavour in this direction.

About the author:

Sudheesh Venkatesh is Chief Communications Officer and Managing Editor, Azim Premji Foundation.