Statistics are very difficult to get the hang of. It’s very easy to see something when it’s only an illusion. A good example of this is a numerical shift from being basically nothing to barely anything. For example: Imagine you have a volume setting that goes from 0 to 100. It’s currently set at one, which is basically nothing. You can’t hear anything at all. You turn it up to five, so that if you focus really hard you can hear the faintest noises. Now it’s barely anything. It grew to 5 times its original size, an increase of 400%, so that it’s 500% of the original, but it’s still hardly noticeable. If you only looked at the numbers 5 times, 400%, 500%, you’d think there’s been a massive surge in volume, but that’s not the case.
This statistical misdirection sometimes shows up in discussions. For example, in her awful transphobic essay, J.K. Rowling wrote: “The UK has experienced a 4400% increase in girls being referred for transitioning treatment.” Wow, that sounds like a lot! Except that the increase was something like 40 girls in 2009/10 — only about 0.0002% of all girls! — to 1,806 — only about 0.008%. (These numbers were obtained by looking at the number of children in the UK in 2011 and assuming they were half boy and half girl. While rough, demographics don’t change that quickly, so these numbers come close to reality.) That is still well below our estimates of how many people are actually transgender; in the USA, a similar country, about 0.6% of the population is transgender. In this case, the increase is from basically nothing to barely anything.
It’s also important not to categorize something as barely anything when it’s more than that. Recently, Paul Krugman wrote on twitter that after 9/11 “there wasn't a mass outbreak of anti-Muslim sentiment and violence,” a statement that many took issue with. Some pointed to hate crime statistics which show that the number of hate crimes against Muslims spiked from 28 in 2000 to 481 in 2001 — a roughly seventeenfold increase, or around 1700% of the number in 2000. Of course, this seems to fall into the mistake of thinking an increase from basically nothing to barely anything is particularly notable, given the number of non-Muslims there are to commit hate crimes. The mistake here is to think that hate crime statistics vindicate Krugman. Hate crime statistics only capture so much information and do not measure how much anti-Muslim sentiment has grown. They also rely on accurate reporting, something the United States doesn’t really have. It’s not particularly hard to find a Muslim person who lived in the United States at that time who can attest to how the culture changed to be more virulently Islamophobic. While the desire for hard numbers is respectable, we can talk about how a culture has changed without them.
There are three types of lies: Lies, damned lies, and statistics. Statistics can be useful, but you have to handle them carefully.