Julien Balbontin/THE VARSITY

A woman over 40 is more likely to be killed by a terrorist than get married; people only use 10 per cent of their brains; one in five children is approached by an online predator.

These are just a few of the most commonly repeated statistics that can shape how we understand the world. In fact, statistics are routinely used — and in almost all cases, required — to support everyday arguments, drive policy, and frame reality.

Given the importance of statistics, it is troubling how often we fail to ensure the validity of the numbers we cite. For instance, there is absolutely no scientific evidence that we only use 10 per cent of our brains; yet, the myth is reproduced in mainstream media — see, for example, the summer thriller movie Lucy, which revolves around the protagnoist “tapping into” the other, apparently dormant, 90 per cent of her brain. Clearly, we need to check our sources before using them to form arguments.

Once the numbers are established, however, there is still more work to be done. A student writing a paper on child sex crimes in Canada for instance, could have a credible source from China, but erroneously generalize the context specific numbers to a different country. More insidiously, partisan think tanks or researchers are prone to producing results that confirm their own political leanings.

As such, we need to ask critical questions of the research process. This will help us gain a deeper understanding of where these numbers come from and what they actually mean. By extension, we will then feel more confident about whether and how we can apply them.

Consider the statistic “one in five children is approached by an online predator.” Which regions does this statistic apply to? What qualifies as an “online predator”? What age bracket does this statistic actually cover? How and when was the data collected, and by who — is there a bigger agenda influencing this research?

Perhaps what is most crucial to consider when evaluating statistical arguments is causal claims, which are often conflated with correlation. For example, we often hear that women earn 77 or 78 cents for every dollar that men earn. Although this statistic may be true, people tend to associate this gap solely with discrimination based on gender.

Undoubtedly, sexism does contribute to the gap, but there is actually a much more complex blend of economic, social, education and cultural elements at play. The wage gap differs greatly by occupation, region, and race. Lisa Maatz, the VP of the American Association of University Women, says, the 78-cent figure is “not an apples to apples comparison.” So, when assessing the root cause for a statistic, we need to bear in mind it is often impossible  — particularly in social sciences, where environments are harder to control — to account for all the related factors.

This critical reflexivity is particularly important for evaluating statistics, but should be replicated more broadly when receiving any sort of information. Climate change deniers routinely cite a petition signed by over 31,000 scientists as proof that there is no consensus on climate change; but these signatures were completely unverifiable. Reading this petition (and the articles that report on it) without questioning its credibility leads people to being misinformed and establishing opinions based on fabricated evidence.

As students and citizens, and especially with the general elections coming up, in October, we need to be well informed about the world around us. This involves not only reading articles, data, and listening to news, but constantly questioning the credibility of these sources and critically thinking before making a consequent decision. The same applies when writing a paper or presenting on a topic  — almost every argument uses statistics to support a claim, but ensuring that you know what these figures mean and where they come from is the most critical step.

Naveli Gandhi is a student at Woodsworth College studying public accounting and economics.

Like our content? Subscribe to our weekly newsletter

* indicates required