 Sharks, thankfully, are restricted to the oceans. They can't get up on land because they don't have legs, and they don't live in the sky because that's jet territory. So sharks are pretty exciting, right? And shark attacks are even more exciting. Statistics less exciting. Just putting this out there is going to take a little while to get to the sharks. I've ranted a bit before about math anxiety in a version, so if you're currently scrambling to close the window as quickly as you can so that you don't have to hear about mathematics or numbers, please instead click here. It opens in a new tab, and I promise we won't talk about many numbers at all when you get back. So statistics are not very exciting in the same way that an n-wrench is not very exciting. You can use statistics as a tool to do some really amazing stuff, but the tool itself is not really going to take your breath away. Don't get me wrong, data can be beautiful, but I wouldn't really go out of my way to frame the Excel ANOVA function. Okay, maybe I would, but I'm weird. As far as mental tools are concerned, statistics is really one of the most important means we have of understanding the world. It's kind of the interface point between the pure floaty logic of mathematics and the chaos of the real world. Actually, chaos is a great example. You've probably heard something about the second law of thermodynamics, which says something along the lines of, in any reaction, entropy always increases. But it's not like matter knows that it has to be more entropic over time and changes itself to be that way. It's actually a purely statistical process. There are just more ways for matter to be disorganized than there are for it to be organized. So given enough time and enough reactions, matter tends to be more disorganized. Statistics is also how we know anything about science at all. A scientist can't just take one measurement and say, okay, it's 10, then knock off for lunch and get published. A big part of science is recognizing that weird stuff just happens sometimes. Maybe your hand was in a weird place, or maybe the lighting was weird on the gauge, or maybe at that particular time somebody on the floor above you was jumping up and down about their grant getting funded. The only way to really hone in on a particular value is to measure it many, many times, and then to run a statistical test on that data to figure out what it really is. However, anyone who knows something about statistics knows that unlike more pure mathematics, it can get very subjective, and that it takes a lot of focus and mental energy in order to figure out exactly what a particular statistical test means. That can be a big problem for just about anyone, not just because many people don't like math to begin with, but some people will use statistics as part of justification for their arguments and treat them like a hazardous substance to be handled at arm's length. Yeah, it's a 13% chance. Sample size, I have no idea. That can lead to fundamental misunderstandings about what that data actually means. Take this example. In 2013, you were more than 20 times more likely to be killed by falling furniture than you were to die from a shark attack. My first gut reaction to hearing that is, my god, the armoire is trying to kill me, sharks are misunderstood, peaceful creatures, I don't know what to believe anymore. And if I didn't think hard about it, I would go on believing that everyone should replace all the shelves in their house with tiger sharks, just to be safe. Would you mount them on the wall with hammerheads? There are several warning flags in that statistic that should jump out at you. First, it's a simple ratio, and ratios can be very misleading. There are usually fewer than 20 shark-related fatalities every year, which isn't really a lot when you're talking around 60 million deaths in 2013. So even a really large multiple of that small number still isn't going to be that many people. It would be like if I had a penny and you had two pennies. Technically, you're twice as rich as I am, but you still only have two pennies. A second red flag is that it's a direct comparison between two very different things. That can be very convincing and dramatic if it's done right, but frequently a media producer will just pick through facts and figures until they find one that's offensive to intuition. The farther apart two things are in a statistical comparison, the more likely it is that there's some factor that's not being taken into account in that comparison. The shark furniture statistic is missing something really important, the sample size. If you really wanted to get a sense of how dangerous sharks were compared to furniture, you wouldn't just be able to compare how many people died each year from both. You would have to look at all of the times those things and people interacted. I'm around furniture every day for extended periods of time. In fact, ever since I got an Xbox controller for my computer, it's even less likely that there will be a given point at which I'm not interacting with furniture. All those interactions are pretty safe. My couch hasn't tried to kill me, but given billions of people and millions of couches, something weird is going to happen on someone's moving day, and someone's going to die. On the other hand, people and sharks are not frequently in the same place at the same time. It's a pretty big ocean, and they're both relatively small compared to it. Odds are you're not going to run into a shark in your daily routine. Considering how few human-shark interactions exist, even 15 fatalities starts to make you wonder if maybe hanging around with aquatic predators might be dangerous. Once you start really thinking about statistics in detail, especially things like this, like the relationship between sample size and incidence, you might find yourself getting a little frustrated with people and news agencies who don't. For example, it's common knowledge that in the original Star Trek series, if you're wearing a red shirt, you're as good as dead. Red shirt is even a colloquialism for somebody who's expendable and destined to die, like stormtroopers. At first blush, that seems totally accurate. Over the entirety of TOS, no fewer than 24 red-shirted crew members die. That's more than twice the number of the nine gold-shirted fatalities, and almost three times the number of people wearing blue uniforms. However, if you consider how many crew members in each color we see in the series, the story changes significantly. Yes, 24 red shirts die, but there are 239 of them to choose from. If you're on screen and wearing a red shirt, you've only got about a 10% chance of being blasted out in airlock. In contrast, nine of the only 55 gold-shirted crew members bite the bullet. That's around 13.4%. You're actually more likely to die wearing the same color as Kirk than you are wearing a red shirt. Of course, you're best off wearing blue, but we all knew that. Have you ever encountered a truly frustrating error in the interpretation of statistics? Please leave a comment below and let me know what you think. Thank you very much for watching. Don't forget to blah, blah, subscribe, blah, share, and don't stop thunking.