 So hopefully people came up with a few suggestions about What it would take to convince them of this sort of claim now This is a really nice example because it brings in a few of the ideas that we were talking about previously about What you would include in the experimental design how you could use experiments and science to inform everyday decision-making and One of the thing I mean if you put yourself into that position What would it take to convince you if if I prepared a cup of tea? Behind myself here, and I had tea, and I added a little bit of milk And I gave it to you and asked you did I add the milk to the tea or the tea to the milk and you were correct I don't think I'd be terribly convinced I don't know whether you were actually just guessing or whether you could legitimately truly tell the difference between Milk added first or second better probably I think would be if I had two cups of tea one with milk added first One with milk added second and then gave them to you and shuffled them around maybe and you had a sip of each and then made a guess So that's two, but even still I don't think that's terribly convincing And I think people probably share the same sort of intuition how many cups of milk how many cups of tea and How many of these sorts of things would it take to convince you that someone can tell the difference between tea added first or tea added second Is it two is it three is it four is it five? How many would it take to say okay? I believe you I think something's actually happening here Well, if we go back to Ronald Fisher's case the way that he set it up was He had eight cups of tea So we had eight cups of tea four of the eight had milk added first four of them had milk added second And he told the woman that this was gonna be the case We'll set this up. We have four here four here and we'll prepare these for you And I'm going to hide which of these are which and I'll present them to you One at a time you can taste them and then you have to guess whether it's milk first or milk second Now one of the important elements here is that he would have to present them Randomly now if you think back to episode three where we were talking about the stocking task where people had a very Right-handed bias in selecting the stockings that they preferred Then the same thing could be happening here If you just had for example all of the milk firsts on the right all the tea firsts on the left People might just choose the one And say all of the ones in the right are milk first And so in order to account for that what you need to do is shuffle them up You need to present them randomly to the woman But what do you mean by randomly because we've just learned that that we're quite terrible at coming up with random sequences So if if you were to give the the eight teacups to the woman in the sort of random order that you came up with you might just choose milk first tea first milk first tea first and She being being the same at picking up these non random things might go milk first tea first milk first tea first And she would do that same random order So it would look like she's responding correctly, but actually she's just picking up on the same sort of non randomness That you created that's exactly right And in fact Ronald Fisher pointed that out not explicitly pointing that out But he suggested when selecting this random sequence that you use some sort of objective device like a coin flip or the roll of a dice or cards or I can't remember something like things that you would normally use in games of chance able to do that a random number table he actually said because at the time they had those and So exactly and so you would present these randomly to the woman And she would have a sip and then she would make a guess as to whether it was milk first or tea first And he had 80 of these cups now I don't want to get into the mathematics But one of the elements that we want to highlight here is that he had eight cups of tea of which four were milk first Four were tea first and she had to get a Sufficient number correct in order for people around the table to believe her but the question is what is that number? What would it take to convince people around the table that she got it right? Well, what if she got two of the eight right three out of the eight? Well, he didn't actually say in the experiment when he wrote this up in the chapter He didn't actually say how many she got correct, but imagine that she got all eight correct that She could correctly distinguish between the Cups of tea with milk added first and milk added second. How how strange is that? Well, there's eight cups of tea I mean again, I don't want to get a mathematics here But if you were to actually calculate the way The possible all of the possible ways of doing that there are about 70 ways of Making a choice in that scenario, but only one way of getting it correct So there's a one in seventy chance which is pretty unlikely that she would be able to do that by by chance or That she would be able to guess correctly Without any sort of capable real capability of doing that. So it seems that that's sufficiently weird sufficiently Unlikely to be able to say that well hang on maybe there is something here, but you could set that criterion wherever you'd like It's not just eight cups You could demand nine cups or ten cups or twenty cups if you had twenty cups that she had to guess all twenty correct before you believed her the the possibilities of doing that are about I don't know one in five thousand I think of doing that correctly by Just guessing alone, which is it's pretty unlikely I would I'd be pretty inclined to say that she can do what she claims if she guessed all twenty correct or even the eight of those correct So I think that's a really nice example of the experimental method and and highlighting this idea of fishiness of Lumpiness and chance. Yep. That's exactly what scientists do Every day that they're designing experiments and they're trying to figure out Whether there's anything strange going on here and you and I are actually cognitive scientists, that's how we spend most of our days and and we some of the things we've been working on in our lab is Testing claims such as fingerprint identification So there are people who testify in court usually police officers about whether a crime scene print matches a suspect or not They're called fingerprint experts, but do they have genuine expertise? Are they better at any better at matching prints than you and I are? We'll find out in episode 10 Yeah, another one that that a colleague of ours when Wu is working on is whether honeybees can tell the difference between artwork by Picasso and artwork by Monet Something else we've been working on is that that flashed face distortion effect that we saw in episode 2 where the faces look Strange and distorted. We've been trying to manipulate ways to to strengthen or to to exacerbate that effect What and what creates it so again? We're just where we're setting up experiments and trying to convince ourselves and our colleagues and everybody else That there's something going on here or or maybe not and something one claim that's Really interesting me at the moment is wine expertise or wine taste testing So there's been a few articles in the popular press very recently Kind of saying there's actually nothing to this like wine tasting is is junk People can't tell the difference between cheap and expensive wine or even red and white wine and there might not be anything to Wine expertise, so I spoke to Tony Montanarchus about exactly this