Denise RQ Numbers can feel really scary. I bet that if I asked this room how many of you were afraid, or didn't like, or were bad at maths and stats in school, I'd get at least 50% of hands up? Probably more. But numbers are becoming more important than ever.
We've all heard the word "big data" or "data science" and really, we're just at the infancy of our numbers and our data gathering. But so what navigate to this website? What does that mean for us? Well, it means that government, businesses, advertisers, media, and even some scientists are going to be using numbers and data to sway your opinion rightly and wrongly. And with enormous amounts of information, we run the risk of misinformation, or even more worrisome, disinformation. So what does that mean for the non maths-and-stats lovers? Are you just going to be bamboozled everyday? No. We're going to learn how this kind of maths isn't maths. It's simply critical thinking. It's learning how to spot the lies, the damned lies, and the statistics. And I'm going to show each and every one of you here today how to do that. So let's take our first situation: OJ Simpson, one of the most infamous criminal trials in American history. OJ Simpson was an NFL superstar accused of murdering his ex-wife, Nicole Simpson, and her lover, Ronald Lyle Goldman. During the trial, one of OJ's lawyers, and not the Kardashian but Alan Dershowitz, a famed legal mind in American law, made a very convincing statistical argument. He said that while it was known and definite that OJ abused Nicole, that it didn't matter. It was totally irrelevant. Because four million women in the United States are abused by their domestic partner every year, and of that four million, only one in 2,500 are then murdered by the domestic partner that abuses them. So the chance that OJ murdered Nicole is only one in 2,500. And you can't convict someone on a one-in-2,500 chance. But what the jury, and most definitely, the prosecution should have picked up on is that the statistic that Mr Dershowitz stated was completely irrelevant. It's what I call a "sample relevancy diversion." Let me explain. In Mr Dershowitz's situation, here's a group of living women who are abused by their domestic partner. What's the chance that the domestic partner that abuses them will then murder them? One in 2,500. But that's not our situation at all. At the time of the trial, Nicole Simpson was dead. She had already been murdered. So she's actually in the group of murdered women who were abused by their domestic partner. And of murdered women who were abused by their domestic partner, nine in ten of them were murdered by the domestic partner that abused them and not by some other member of society, some random murderer. So the chance that OJ killed Nicole was 90%, not one in 2,500. This is a sample relevancy diversion: point us at a group that's similar to the one we're talking about, state some statistic about them, when really, it's not the group we're talking about at all. In our case, the diversion sample was the living women who are abused by their domestic partner. The real sample, and the one Nicole was in, was murdered women who were abused by their domestic partner. So, now we've maligned lawyers and ruined any confidence in our justice system, let's move on to the media. I was sitting in a cafe, eating a bacon butty. It's the one Scottish dish I have truly adopted. (Laughter) And while I was smothering it in that delicious Brown sauce, I read that my chance of bowel cancer would increase by 20% if I ate that bacon butty. That was the newspaper headline. All of a sudden, there's some government guy on TV, talking about the "true perils of bacon," that we need to tax it, and every news outlet, as you can see, was also talking about that. But let's take a little bit of a closer look. While the headlines were fuzzy, the actual study said that if you eat two rashers of bacon every day, your chance of bowel cancer will increase by 20%. That's really scary. But the question that you should be asking is that it increases 20% from what? Well, regardless of how many bacon butties you eat, five in 100 people will suffer from bowel cancer in their lifetime. And if all 100 of those people eat a bacon butty sandwich every day, yes, it's true; your chance of bowel cancer increases by 20%. But what's 20% of five people? One person. So if all 100 people eat a bacon butty sandwich every day, instead of five in 100 getting bowel cancer, six in 100 will. Which is awful. Really awful. But it's not nearly as alarming as when the newspaper splash across a headline that you're going to increase your chance of bowel cancer by 20% if you eat a bacon sandwich. This is called "relative risk." Your chance of getting some disease is one in a billion, and a researcher says if you increase your wine consumption by two glasses a week, you increase your chance of the disease by 100%. But the chance of getting the disease is still only two in a billion. I think I'll keep drinking that wine. (Laughter) OK, we can't trust the media; we can't trust the justice system; I'm just going to win the lottery, buy some island in Barbados, and get away from it all. Somebody has to win, right? Right? Someone has to own that island? And as you can see, Camelot, the operator of the National Lottery, makes a nice buck off the back of our belief that somebody's got to win. I'll pay my two pounds, I will take my chances, I want that island. But what really are my chances? One in 45 million. That means nothing to me. I have no idea what that means. Do you? I don't think so. I mean, you could say you're more likely to be a movie star than win the lottery, or you're more likely to be killed by a flesh-eating bacteria - which is true - but that doesn't help me quantify 45 million. But David Spiegelhalter came up with a great example of how to quantify it using bathtubs. OK. Imagine you're in a bathroom, really imagine this. You are staring at your bathtub. It is filled, brimming over, with dry rice. Nothing fancy; just the dry rice you buy at the grocery store; brimming over. And you take one single grain of rice - I'm holding it in my hand - and you paint it gold. And you bury it somewhere in that bathtub. Bury it, deep in, cover it all back up with dry rice. Now you get people to walk into your bathroom door, pay you two pounds to close their eyes, stick their hand into that tub, and miraculously, with one lucky dip, pull out that one golden grain of rice. It's a pretty good business if you can convince people to give you two pounds to do that. (Laughter) If we really understood what 45 million meant, would we still play the Lotto? I did buy a Lotto ticket because the EuroMillions is 130 million this week, but I knew better, right? What all these situations and examples show is that you don't need to bring a calculator everywhere. You just need to have a little bit of critical thinking in the back of your mind. So if someone says something about a group, you say, "What group are we really talking about?" Someone says you have 100% increase in a disease if you eat a Mars bar; "100% increase from what?" Or, it's one in a million. Find some way to quantify to you what a million is. Spotting the lies, the damned lies, and the statistics isn't math. It's simply critical thinking. Really what this all means is that if something doesn't feel right - a number, a statistic, a study - it might not be. And now you have the tools to ask the right questions. So when someone says, "It's the truth. The data proves it. The numbers show it." It may not be the truth. Thank you.
0 Comments
Leave a Reply. |