Numbers Be Damned

January 20, 2022 | By Shi Wen Yeo MC ‘23

image description: a close-up of a child’s hands drawing a line with a red pencil. They are writing on a piece of graph paper, with several numbers already written down.

This piece was written at the Veritas Forum 2021, an annual writing program offered by the Augustine Collective. Students from various universities work with writing coaches to write articles about virtue in the sciences or social sciences.

---

Memory is like a haze that gradually sharpens into focus as one grows up. One of my earliest, albeit haziest, memories is of Saturday afternoons when I was nine years old. For any child that age, Saturday afternoons were synonymous with pee-wee baseball games, time spent hanging from tree castles or playing in the sand. For me, Saturday afternoons were the longest afternoons of the week. Instead of basking in the gentle sunshine, I spent those afternoons suffocated by harsh, luminous lights that dangled from the ceilings of my “tuition center.” And I was not the only one. Like me, thousands of nine year olds around the country also spent most of their weekends in centers like mine. The goal in mind? Passing the “Gifted Education Programme” test at the end of the year. Many years ago, a group of people got together and decided that a test administered at nine would be a good way to determine which children were more statistically likely to be “gifted” or not. And so my parents’ hopes were pinned on my frail little back and a couple points on a test. And a belief in the system of statistics that governs it all.  

There is no doubt that statistics have taken over our lives. In my head, this almost looks like a world where everything we touch is made of binary code, zeros and ones, much like in the movie The Matrix. Here, it would be cruelly ironic to quote a statistic to support this claim, but I can venture fairly confidently that almost all of today’s journalism quote some kind of numbers in their writing, either to buttress their claim or refute another. Throughout the pandemic, we have been bombarded with information about vaccine efficacy, and we make decisions about which vaccine to receive based on one percentage point differences. Our hearts sink when the Curve mounts steeply and our spirits soar when the Curve flattens. As I found out in my short internship stint at the World Bank last year, most countries have entire departments dedicated to the collection and analysis of statistics. A concept that promises to divine the future, that commands the worship of millions of people, and that is ubiquitous in almost anything. By this description alone, one would undeniably categorise statistics as a religion.

There is also a gaping gap in journalistic writing about statistics. In preparation for writing this piece, like any good student, I very meticulously and studiously looked up articles on statistics by major news outlets. And this was a very interesting and telling process—when I searched “articles on statistics,” the search turned up pages upon pages of actual statistics. Dashboards, infographics, articles peppered with percentages. But nothing about the hows and whys of it all. 

This is a problem. Most concepts that are so influential, like religions, are subject to intensive global scrutiny. Mention God in any piece of writing and be prepared to deal with a deluge of cynicism. One cannot help but wonder why such cynicism is not also fairly applied to statistics. Blind acceptance has become a norm, and when presented with statistics we revel in the calm assurance that these numbers on a page give us, without questioning the methods, much less the significance of those numbers as a whole. Yet, I would contend that there are good reasons to be wary of statistics.

On a very basic level, there are actual reasons to be suspicious of statistics when they are presented at face value. Last year, having completed two semesters of advanced econometrics, I recall much more of the class being about the limitations of those statistical methods than about the methods themselves. Given grossly oversimplified datasets, a lot of the work involved using overly-specific case studies to generalise to larger trends. Beautifully linear models were imposed on trends that were clearly non-linear, and, in the papers that I read, certain statistics were cherry-picked to tell a story which the writers wanted to tell, but that was clearly bereft of reality. Ironically, knowing more about the inner mechanisms of statistics made me even more skeptical of its power, just because I found out how vulnerable statistics were to manipulation. A simple study could be subsetted by time period, population or other characteristics to say something totally different about a population. Sure, there are some studies that are statistically robust, and there are some genuinely well-intentioned data scientists who control for probably every statistical flaw there could be in the model. But when presented with a statistic, the natural instinct of most people is to believe it, without any regard for the quality of the study. This is how a large part of the population is left vulnerable to the wily, manipulable statistics. 

Shortly after being admitted to one of the most competitive public middle schools in Singapore came a day in tenth grade. A teacher had personally come to my homeroom and asked for me. A series of hushed whispers descended upon the room—everyone, including me, knew that it was some serious matter. This teacher was famous for never leaving his office. I wondered what could be the issue, and stepped outside the classroom with a mix of trepidation and anticipation. Turns out, it was about a special history enrichment programme that I had applied to. I was rejected, the teacher said, because my grades had just missed the mark for admissibility. I gently pleaded with him, asking if he would please consider these essays of interest I had written, alongside my background in competitive debate. He shook his head and said, “I really can’t make any exceptions. You have to realise, you are just a statistic to me.”

I wish I was joking when I quote these words, or trying to be dramatic. Yet, these were the exact words he had uttered to me, a mere fourteen year old who was interested in history. It felt like there was a glitch of some sort in the Matrix—sure, we all knew that this was the harsh reality of the statistics-obsessed system. But did he really have to come down personally to deliver this stinging slap to my fourteen year old face? Was it really necessary?

Reflecting on this experience, I think another good reason to be wary of statistics is that they induce an anxiety that is deep-seated and far too unnecessary. Our world is already so anxiety inducing. Distractions, opportunities for comparison with others, consumerism, and other social ills are ubiquitous, even from childhood. Believing in statistics means also believing in arbitrary boundaries in society to sort children into. It means believing in the misguided notion that somehow an extra 0.1 increase in a grade point average makes one worthy of joining a history enrichment class. And it is this belief that leads to anxiety, because then people are hardwired to fret about every little point there is, instead of worrying about the bigger picture. There simply is no need for this extra anxiety in an already highly-strung world.

In the Sermon on the Mount, as described in the synoptic gospels, Jesus gathers his followers and dispenses useful wisdom on a number of things. One of the oft-quoted bits of this sermon is a segment where he indicts those “of little faith.” His argument here is a naturalistic one that, I think, levies an important criticism against statistics, especially those that claim to have predictive capabilities. Roughly paraphrased, He says something like—if the birds in the air and the animals that walk the earth don’t worry about tomorrow, why should we?

Whether or not you are an adherent of Christianity, I think that the sentiment of this sermonette rings so true in many ways. It addresses a very important impulse that statistics encourage, which is anxiety to know the things of tomorrow. Data scientists are obsessed with uncovering the predictive capabilities of models, almost “playing God.” The reality is that this venture is often highly futile, because there is so little in nature that is predictable. At the end of the day, those who look to statistics as a religion are grasping at straws because, while potent, statistics hardly ever tell the full story. They can only act as a crutch that seems reliable to those who crave an artificial sense of stability and meaning in a fundamentally unstable world. And for those who believe in God, this sermonette shows how misaligned with religion statistical models can potentially be. When we overtly venerate the power of statistics to explain and predict, we lose sight of the hand of God that infuses our lives with purpose and has sovereign control over it.

To be clear, statistics in and of themselves are not evil. A healthy example of statistics being used well, in my experience, is the Effective Altruism movement. In my brief experience with Yale’s chapter, I discovered ways in which members of the movement probe the methods by which data is collected and its representativeness, in order to make an informed decision about charities to which they can donate, or the ways in which they can best do good. This was a positive example because of the amount of rigour that went into their interrogation on the quality of the statistics and the impulse behind using the statistics—not to castigate or judge, but to discover ways in which blind benevolence gives way to inefficient altruism. 

Back when I was nine, I didn’t end up passing the Gifted Education Programme test. The stress got so intense that I actually just ended up going to the test center and filling in the bubbles on the test sheet in a pattern that I liked, without looking at the questions (sorry, Mom and Dad!). I did well, but not as well as my parents hoped I would have in the Primary School Leaving Examination four years after. (As I said, Singapore is a land of tests.) Do I regret any of it? Perhaps other than my poor parents’ emotional roller-coaster, I really don’t think so. 

At the end of the day, I am still convinced that all of us are more than mere statistics, and that there is only so much that statistics can explain or predict. In today’s society, for the sake of our own mental health, we would all be better placed to break away from the veneration of statistics and instead pursue a more holistic, qualitative appreciation of the world around us. Instead of accepting numbers at face value, a healthy amount of skepticism and a greatly lessened degree of anxiety would be helpful heuristics to start approaching statistics. Because ultimately, it is at the end of our own human comprehension where true meaning begins.