Numbers don’t lie, but…

Are they really saying what we think they are?

During the lockdown, out of curiosity and as part of an article I was writing about whether putting salt on the tongue helps to relieve the pain of having pepper in one’s eyes, Moses and I carried out a little survey. All participants had to do was answer a simple question: ‘Do you think putting licking salt helps to relieve the pain of having pepper in the eyes’? When we were done crunching the numbers, we found as expected that most respondents agreed that the salt was helpful. However, as we looked more closely, we found an interesting trend — women were more likely to respond ‘yes’ than men. As a matter of fact, 80 per cent of the respondents that said YES were women. Moses and I looked at each other while we spoke about this like we had found our Eureka moment. To us, it was evidence that women were more likely to believe stuff like that than the other gender. We had the numbers to back it up. It took some more reflection to see what was wrong with our conclusion. First, our sample was too small for us to conclude, we had responses from a little over 20 people. Two, our sample was not diverse enough — most of the respondents were female, so it made sense that whether it was YES or No that got the most responses, the women were going to be well represented. Plus, a disproportionate number of the females that responded were friends who most probably had the same ‘worldview’ and were exposed to the same influences. So while the numbers were not lying, we could have misled ourselves by thinking deeper about the number and what they might mean.

Where did they come from?

About a year ago, an ad on BBC that claimed 68 per cent of future jobs hadn’t been created. 68? Really? Not 70? I was shocked by the degree of unnecessary precision. They said it with so much authority that one would have thought they travelled to the future, took stats and came back on earth to make comparisons. BBC didn’t mention how of these new jobs existed in the future, or even when this future was. With all the unknown variables, how is anyone able to come up with a percentage?

Are they necessary?

“90 per cent of Tarok women are fair in complexion.” Why not ‘most’ instead of 90%? Could it be because humans have a bias for numbers and know their opinions are more likely to be accepted when they throw a few stats around?

Can they explain what they claim to?

There are certain things that cannot be expressed as numbers. How can you say you love your children 80% more than you love your wife?

Bottom line

While numbers help us to understand the world around us, we should not feel compelled to accept a sentiment just because it has a statistic tied to it. While numbers do not lie, they may be misunderstood or twisted. We should ask ourselves, are numbers sufficient to answer the questions we are asking? Where did they come from? Could they be saying something else?