This week I'm joined by Karl Mamer to discuss the data behind three well known urban legends. Did a large blackout in New York and surrounding areas result in a baby boom nine months later? Do subliminal messages affect our behavior? Is placing beer alongside diapers a recipe for generating more revenue than these products in separate locations? Listen as Karl and I explore these claims.
Data Skeptic
Data Myths with Karl Mamer
(upbeat music) - The Data Skeptic Podcast is a weekly show featuring conversations about skepticism, critical thinking, and data science. - Welcome to another episode of the Data Skeptic Podcast. I'm pleased to be joined today by my guest, Karl Maymer. How you doing, Karl? - I'm good, Karl. How are you doing? - Very well. Many of you will hopefully know him from, if you didn't know him before this, checking out when I appeared as a guest on one of his many podcasts, The Conspiracy Skeptic, where we talked about the Bible Code. But you're rather prolific podcaster, besides that, I learned from conspiracy skeptic a couple mentions about soul survivors of your time spent living and teaching in Korea, which I've been going through the back catalog of. And somehow I never knew about until recently, but have been enjoying Ask a Canadian. So I hope anyone interested will also check that out as well. - Ah, yes, yeah, yeah. I mean, listeners of conspiracy skeptic, 'cause they always call them the Korean questions, or, you know, sort of aware that I spent time in Korea, but yeah, I don't talk too much about Ask a Canadian on conspiracy skeptic. Not that I'm embarrassed by it, but, yeah. The premise is I'm, well, I'm the Canadian, and then sort of a panel of sort of four Americans, honest to God Americans, some from even the South, which in Canada is actually anything sort of South of Ohio, but, you know, but these are actually real Southerners. And they sort of apply me with questions about Canada and sort of what's the Canadian view on this international crisis. And it's basically a comedy show. - Yeah, and we've got another sort of funny, I don't know what you call it, if it's just a, well, obviously it's just a coincidence, but a bit of interesting synchronicity that if you go back and listen to when I had Louis Zachian to talk about the game science dice, he mentioned a certain young woman writing a biography about him. - Yes, yes, well, it's a bit of a long story, but I mean, I'm a Dungeons and Dragons nerd from Wayback. I'm 48 years old, so I started getting to D&D about 1980. I believe, this is, they were just starting to release the advanced D&D manuals, first edition, so like Monster Manual and Dungeon Master Guide and stuff like that. You know, I was kind of a big fan of, how to pronounce this name, I was probably not Zachie, but I don't like that, I was even correct. - I think he said Zachie. - Zachie, okay, right, yeah. I'm not sure if I encountered him at a couple conventions, but you know, I was sort of aware of his game company, and you know, back then, sort of a lot of his games were just kind of like, you know, to a 13 year old, like sort of space opera type games and superhero games, and it was just all Star Trek and that sort of stuff. When you look back, it was sort of, you know, it was all sort of typeset on basically typewriters and hand-drawn art, and it was just, it was just gloriously amateurish, but so much fun. So rolling forward, before Wikipedia, there was sort of a online encyclopedia called Everything 2, and everything 2 kind of encouraged a bit of a hitchhiker's guide to the galaxy, kind of, you know, where you could play fast and loose with a lot of the facts and stuff like that. Anyway, so I was big into everything too, but like in a lot of like social situations, these kind of online things, people, I don't know, little cliques and cults form and stuff like that, and I guess, you know, people sort of, there was sort of an anti-carle faction because I just wanted to go on and write funny, you know, nodes and stuff like that. But anyways, to make a long story short, I sort of noticed that women on the system were kind of left out. They're sort of not kind of like, you know, picked on, and you know, they're right up sort of, you know, deleted or, you know, voted down and stuff like that. So I thought, I'm just gonna delete my account and just create a female account, and then just go on as a woman user and then just be left alone. - Yeah, no drama. - Yeah, so I don't think anybody ever really sort of, you know, sort of figured out why does a 22-year-old woman have this in-depth knowledge of like late '70s role-playing games, but I was left alone, and I could just, and eventually I wrote up a history of Luzaki. Around that time, it was sort of like, you know, your research, you know, he was retired, he was retired from the Air Force, I believe, as a sergeant, but different cons, you know, he would be referred to as Captain, or he'd be referred to as Major, I think he was referring to as Colonel. So to me, I just thought he was doing this kind of for fun, you know, like every year he would, you know, now I'm a Colonel. Like it just, it didn't make any sense that these titles kept changing other than it was just kind of a bit of his own, you know, his own fun. Like, a lot of people refer to him as Uncle Luz. So he was this very, I mean, let's not talk about him, 'cause he's dead or anything, and he's still very active in the gaming world. But, you know, he's this very nice man. And so, yeah, so I just sort of just this offhand comment about how every year he would promote himself, and eventually, you know, awarding himself the title of Colonel, which I guess maybe for the last, 'cause I wrote this in like 2002. So, it seems like, maybe for the last 12 years, he has probably been looking for this woman, you know. So I figured out my old login credentials and stuff like that, and I updated the everything too right up about him, and corrected it. - Oh, that's good, yeah. - Yeah, he has a wiki-like system, yeah. - Yeah, so that's the, that was just really weird, long story, but yeah, you manage, you manage to set the record straight, so. - Well, it's good, yeah. I'm glad we got the facts out there, then. - All right, good, yes. - Yeah, I enjoyed that and, you know, you, I've noticed you've got a prolific amount of essays as well. I've been enjoying on your site. What's the URL for anyone who might want to read that or some of the others? - Oh, you know, I don't know, 'cause that was like an old website that eventually, I think it was like, I started on GeoCities, and then, you know, it was back in the day when I didn't really know how to sort of do like relative URL paths and how to sort of link up things properly. So I just kind of ended up dumping it on my, well, my main site, you know, yred.com. There might be a link to E2 or everything too, but there's a lot of broken links and all of that and stuff like that, 'cause they just kind of dumped it in there and never really got around to sort of properly linking it up and stuff like that. But if you have show notes, you can maybe link it correctly or something like that. - Yeah, for sure. I also was pleasantly surprised to learn you were an early Commodore 64 programmer when we talked last time, so. - Yo, yeah, you got just some credibility on my audience. - Yeah, I mean, basic language, basically. - Sure. You know, you're in a way, I'd say you're rather a Renaissance man and I was also recently surprised to find out your, I guess, graduate work, your thesis was all around urban legends and that's what we're gonna discuss in some part today. How'd you get along with that or interested in that? - You know, David Letterman, really. I just saw a, yeah, when long, long time ago, this was being the early '80s, Letterman, you know, he said, "Oh, I'm gonna have on a guess." I wrote an author wrote a book called The Vanishing Hitchhiker and I was almost ready to turn the tea box 'cause I just thought it was about, you know, the death of hitchhiking, like, no one hitchhikes anymore and it's just gonna be some hobo or something like that, but it turned out to be about a guy who's sort of, you know, a folklorist who sort of documents urban legends which probably most of our readers are probably really familiar with urban legends. Snopes.com covers them in excruciating detail but basically, see, a friend of a friend's story. So the, you know, the Vanishing Hitchhiker about, I think that was what about, you know, people who sort of pick up this hitchhiker on the road and then eventually, they turn and the hitchhiker is not in the car and they arrive at a, you know, a road stop and talk about this spooky thing and then, you know, the waitress there is like, that was my son, you know, killed in the auto accident. Whoo, so a lot of them can be quite spooky but others, I mean, there's the urban legend for almost anything out there and consequently, there are some data, big data, data mining urban legends - Absolutely, I was even, one of the more famous of the urban legends is the one we can start out with, the Baby Boom Blackout. - Okay, yeah, the Baby Boom Blackout. So this sort of goes back to, believe it was called the Great sort of New York Blackout, 1965. - Of which Canada was also affected? - Yeah, yeah, a lot of sort of the Northeast 'cause that, you know, sorry, our power grid is really connected, especially in sort of Northeast. Something goes down and Montreal takes out, you know, it takes all of New York State, that kind of thing. You know, it was 1965 as the name sort of implies, ideas that during blackouts or other kind of disasters where people sort of are housebound and have no electricity that, you know, people, you know, kind of couples, you know, they get busy and then, you know, in order to an amount of people, you know, getting busy leads, you know, nine months later leads to sort of little mini baby booms. So that sort of story started in nine months after the 1965 Blackout Hospital. New York, I think it's the New York Times. They sort of had, yeah, it might have been like a series of articles about basically just sort of talking about, oh, there's this boom and babies and, you know, definitely goes back to the Blackout. But then probably five years later, people sort of, you know, they took a look at the actual data and sort of discovered, you know, talking to, you know, nurses and hospitals and going, oh, yeah, yeah, it's just, it's crazy, you know, that sort of thing. So it's kind of what you call it, you know, kind of confirmation bias. And you're primed now to sort of expect craziness in the emergency ward during full moons and you're primed to sort of expect an inordinate amount of births, you know, nine months after some disaster or something like that. So it just sort of seems like there are, you know, more crazy people or more births. But when you actually look at the hard data, it sort of is not true. Same thing with a member, her was at Hurricane Sandy in 2012, again, amateur, you know, demographers were all kind of again predicting, oh, there's gonna be this huge baby boom. And then I guess a couple hospitals in New Jersey, nine months later, they are actually reported, you know, an uptick in babies, you know, being born, which again, they attributed to Sandy, post-Sandy sort of baby boom. But when they sort of looked at the data, they sort of realized, oh, no, you know, this hospital just actually, the hospital just sort of opened up kind of more beds for, you know, for, you know, for expecting mothers. So, you know, they were just able, you know, they just had more capacity. So, yeah. And so, yeah, it's just one of those sort of classic things where, you know, if you suddenly do have, you know, natural fluctuation in the number of births, I mean, there are natural fluctuations. And then you suddenly have this big uptick and then people will count back nine months and sometimes they'll find something, right? They go, oh, yeah, there was that, you know, that blackout or was that storm or, you know, those sorts of things, but if they don't find anything, then they'll just attribute to a natural fluctuation. So, it's, yeah, yeah. - You know, it surprised me when I was looking into this, that you never see the opposite. It's never like nine months after 9/11, there was a decline that, you know, a morning country was less amorous. It's only a boom. - Right, yeah. Well, I mean, that's the thing too, is that, I mean, you could almost sort of spin it so no matter what happens, you know, people are gonna, you know, gonna get busy. So, you know, after 9/11, I mean, it just wasn't a lot of things to do for about a week after 9/11. So, it has this kind of like common sense that, you know, okay, you know, well, nothing to do is the more people are gonna have sex and stuff like that. But when you really sort of dig down into it, you sort of realize it's like, okay, well, you know, people who are already on birth control are just not gonna suddenly go, you know, powers out, let's, you know, let's throw the condoms out of the window or it's pills will work anymore. - Yeah, exactly, you know, so, and it's like, however much I try to tell my girlfriend that, you know, condoms are more effective when the television is on, it's actually not true, you know? So, and then, and then, okay. So, you know, but they're definitely gonna be people who are like, may have sex that night, who would not have had sex that night because there's just nothing else to do. But, you know, but then how many extra, you know, how many women that one particularly night, you know, are actually capable of getting pregnant in that timeframe, you know? So, there's a very small number of women and then, you know, a very small number of those women are actually gonna conceive and then, you know, and then something like, I don't know, 40% of pregnancy sort of end in miscarriages, you wouldn't really technically actually expect huge numbers of extra babies being born when you take all of those sorts of things in total. - Yeah, it's a good point. Even if we all, you know, as some nation or as a species agreed, let's all conceive on one day, it still would have that failed pregnancy rates and the variability of gestation would make it hard to pinpoint nine months, exactly. - Yeah, yeah, exactly. - But there is something, I think that's part of maybe the recipe for a good urban legend. It has to have some sort of intuitively plausible, you know, if you just think about it on the surface, it sounds right, I don't know if you're working with urban legends, is there a recipe like that, are there kind of a couple of high points you have to hit to have a good one? - Well, I think a lot, a lot of it. I mean, sometimes it's making you seem like you're in the know, you know, you've got the inside track on something kind of clever and that's almost sort of the way I then transited into conspiracy theories because, you know, again, people who believe in conspiracy theories, you know, they put themselves in sort of the camp of the, like the army of light, you know, we're out to fight the big evil. So, yeah, so a lot of urban legends basically sort of involve, you know, like you have certain sort of inside information and a lot of urban legends are just really fun and they're just fun stories to sort of tell. We are a storytelling culture and, you know, more stories you have, hire your statuses in the tribe kind of thing, so. - One of the things I noticed, and this is maybe more growing up, that that was part of the secret to a good story that was told around the block is someone had some piece of information, there was always a cousin or, you know, someone two towns over that knew something you couldn't verify. - Yeah, exactly. - And the element of mystery that makes it sort of plausible. - Yeah, I mean, my actual honor thesis was, I sort of noticed a lot of, like, I sort of tackled the, like the horrific urban legends and I noticed a lot of the victims tended always to be sort of young women, like it never, the victim was never like a, you know, like a sort of jockmail or, it was always sort of like, you know, the most vulnerable person in society. - Yeah, maybe sooner. - Yeah, exactly. And I thought to myself, okay, well, maybe that, 'cause, you know, part of urban legends are meant to sort of maybe transmit cautionary tales. Like, you know, don't go smooching with your lover near prisons for insane criminals, you know? That sort of stuff, or, you know, fast food, maybe is not healthy for you, could contain rats. You know, so I thought to myself, maybe more horrific the urban legend is that that increases the believability. So as you increase the horrificness, does that increase the believability? So I sort of gave different, same urban legend, but sort of altered the horrificness to different subjects, you know, first year psychology students. But basically, I found, I found sort of no correlation that, you know, more horrific the urban legend was, it did not actually sort of correlate with more believability. - Interesting. - Yeah. Although one thing, it also did sort of give me a good insight into how, as you, the experimenter, you can very subtly manipulate the outcome, because, you know, I started talking to my thesis advisor, and, you know, sort of like, in the urban legend, it was like, to me, the most horrific one is the woman dies. - Right. - And the next one was she had to have, sort of very painful bone marrow transplant, and then, you know, or then she had to go on antibiotics for a week, or just go on some calcium vitamins for a couple days. So that's how, in my mind, I rank ordered it. And then my first is like, well, that's your assumption, but how do you know that's the true level of horrificness? So I had to then do a second kind of survey to, you know, to get subjects to rate the actual horrificness. And what I started to find is a lot of people were ranking a painful bone marrow surgery as more horrific than dying. Like a lot of people would prefer, yeah, a lot of people prefer to die. And when their answers did not meet my expectations, I found myself arguing with them. - Right. - Really? Like, really? Are you telling me you would rather die? You know, now that you met, you know, so it's just very interesting how these things are not in any kind of, you know, method or something like that. And, but these things can sort of wildly influence possibly experiments. So, which is why, again, we need replication. - Absolutely. Well, another good one that has some replication that we're gonna talk about is subliminal advertising. Much like the blackout is one of those that seems ubiquitous in culture. Everyone kind of knows that this is an idea. And I don't know if everyone's as skeptical about it as you and I will be. - Yeah, a lot of people probably heard of it. And some people might even probably still believe it. There's a notion that a movie theater during the movie sort of inserted like drink, coke, eat popcorn, like single frames into the movie. And then, so, you know, people weren't able to sort of see the single frame, but then at, you know, during intermission, which we don't have anymore. So the, this is maybe sort of slightly dating when this quote unquote experiment was run, you know, during intermission, then like, apparently this drove up popcorn sales and concession stand sales. It was actually sort of started, what was his name? His name was James Vickery. And he claimed he ran the experiment in 1957. And people back then were even actually kind of dubious. And so they asked him like, well, can you rerun the experiment? You know, can you share us the actual method? And he was sort of very, very kind of cagey about his methodology and his actual data and other, just one kind of people that take his word for it. And eventually, he was, later on, he had to sort of confess that, you know, he just made it all up, you know. And I think people might have tried to sort of repeat it, but they were never able to sort of kind of get a repetition. So, and that largely killed the whole subliminal advertising thing until 1973, a sort of journalism professor or communications professor named Wilson Brian Key. He came out with a book, a subliminal seduction. People are probably familiar with some of these ideas. In that, he sort of claimed a lot of amazingly crazy stuff. So he claimed that whiskey ads in the ice cubes and whiskey ads, they put like skulls in the whiskey ads or, you know, offer ads of Ritz crackers. They imprint the word sex on Ritz crackers. And this apparently drives sales or desire to fill deathwishes or sort of a desire to, I don't know why he would ever want to associate crackers with sex. But yeah, too many crumbs. Yeah, but one of his kind of crazier ideas was that Play My Magazine kind of wanted to sort of titillate kind of like the homosexuality in all men. And so a lot of their models were actually not sort of naked women, or actually naked men, just sort of cleverly. This is like Playboy, they didn't sort of show anything other than sort of topless women, right? So sometimes like, you know, their cover ads, you know, that would not be actually a woman on their cover page. It would be like a man and a wig. And sort of lead us, you know, sort of confuse us and make us, you know, kind of get turned on by like, you know, a man or something. But yeah, many, many sort of crazy ideas. But I mean, you know, kind of based a lot in sort of pop psychology type ideas. But he did no controlled trials, like, you know, sort of sick people down and go, yeah, see anything in these ice cubes, ice, you know, like he didn't do anything like that. He just kind of would just himself, or him and some grad students would just like pour over pictures. And he would think he sees, oh, I see sex and, you know, these crackers, or I see skulls in these ice cubes. And I think it's his master work. It was called the clam plate orgy and other subliminals that media use to manipulate your behavior, which he sort of claimed, I think it was a Denny's placemat. They had kind of a artistic depiction of a clam plate special. And in that, you could see, like, there's a human orgy, a human orgy that also involved up to one donkey. This was his claim. So I hope we can put the explicit tag on this side. Yeah, absolutely. I'll be proud to say this would be my first explicit tag. So let's get right into it. All right. So yeah, so Brian Wilson, Brian Key, he actually made his way to up here to Canada. But sort of for a brief time, sort of taught journalism at the University of Western Ontario. He was like a friend of Marshall McLuhan. So I mean, that sort of helped him. Yeah, Marshall McLuhan was sort of a big, you know, because he was like sort of the god of mass communication. What was his famous line? You know, the medium is the message. And so, yeah, I think sort of his friendship with Marshall McLuhan sort of helped him just not get trumps of academia for having all these bizarre ideas. But yeah, I haven't seen that clam one, but I've seen at least a couple of the ice cube ones. It's funny, much like a lot of weird hoax type things. If you look at it cropped and tilted it just the right way, there's at least one that does look very sexually suggestive. But then when you see it zoomed out and at its proper angle, it's like seeing animals in the clouds at some point. Yeah, I mean, we know this today is peridolia. Yeah. And, you know, basically sort of like, if you're told to look for something or listen for something, you will, you know, you'll most likely kind of find it. Actually, professor in Toronto just sort of won an ignobel prize, whereas work on peridolia, like seeing, you know, Jesus Christ and toast or something like that. I wasn't familiar with that. What was his work exactly? Oh, I forget the exact title, but yeah, just just quite recently he was sort of awarded an ignobel prize, basically just for like seeing Jesus on toast or something like that, you know, peridolia, yeah. Yeah, it seems like we either see Jesus or sex. There's not a lot of, I don't see the state of California anywhere on any ice cubes and toast or whatever. No, exactly, yeah. But I think the most damage he came close to doing, he's actually dead now, he died in 1998, sort of surgery complications or something like that. But Judas Priest. All right, yeah, it was in the '80s. There were a couple of kids, you know, depressed, addicted to drugs and alcohol, broken families, that sort of stuff, and they were like listening to Judas Priest album and then formed a suicide pack and they took shotguns and they were like 18, 19 years old. And the first one, you know, put the barrel under his face and kill himself. And the second guy put the barrel under his face and pulled the trigger and he didn't actually die. He blew away most of his face, but he lived, or at least I think he lived. He eventually did die of sort of complications, but he lived for about 10 years or so. And so his parents sort of sued Judas Priest, sort of saying that, you know, there were these back masking tracks and there's one song they're not saying to us, he was saying, you know, do it, do it, do it. He has like, you know, kill yourself or something. So yeah, so this Wilson, Brian, key guy, you know, he's sort of testified, you know, at this trial that, yeah, you know, subliminal messages, you know, can have all these effects. And luckily the judge was kind of like, so you'd got any scientific evidence, something other than your opinion, no. So yeah, so the judge eventually, I think the judge is like, he ruled Judas Priest was not responsible, but he did kind of rule that you know, there are subliminal messages in the music, but there's no evidence that it can actually hurt anybody. So they did sort of throw the case out of court or award, you know, rule in Judas Priest's favor. - So did Key have any outside of his book, did he have any actual like empirical data that proved this out or was it all just sort of examples? - No, not one single reference. I mean, I remember reading this in the 80s and was just amazed at how, you know, where are his controls, you know. It was just, yeah. Like he was talking about how like, you know, even my grad students who have been, you know, primed, you know, about subliminal messaging, even, you know, they found themselves being influenced by these subliminal messages. And I'm like, yeah, 'cause they want to keep on being grad students working for you, you know, you know. That was sorts of things. Like it just, yeah. - I remember being taught, not in great detail, but in introductory psychology class I had as an undergrad that it was taken for granted, this was a thing. It all seems to stem from, I guess, mainly these two guys to sort of capture the minds of people. - Oh, for sure. I mean, 'cause I mean, it has, there's a certain plausibility to it, you know, and a certain lure to it. I mean, especially the idea like, you know, you can go to sleep at night listening to like, you know, your algebra tape or something and wake up in the morning and be able to sort of pass your algebra test or something like that. - Yeah, there's sort of a wishfulness. We would all like those things to be true. I found one study I was looking at where they had flashed different words. I guess they were positive, neutral, and negative. And this study ended up saying like, well, we found that some of the participants, some of the time were guessing that what they saw was negative. And that was the closest thing to a confirmation I could find, but it was so many, that it was just too many degrees of freedom for my taste. - Yeah, yeah. I think it's just, there's just really no data there to sort of indicate, you don't even think to it. Sort of like, what would Brian Wilson kill? Like he would even like involve, you know, hypnotism. Like, okay, so I've hypnotized the subjects and then shown them covers of Playboy magazine. And it's like, now you're adding in a whole other level of pseudoscience, you know? - Yeah, yeah, it's funny how to get layered very quickly. Well, the third I know we're gonna get into it. And one of my personal favorites is the, also ubiquitous beer and diapers urban legend. - Which, yeah. - I'm not clear on the actual origin of this, but I feel like it's one that most people are familiar with. - Okay, yeah. Well, the classic legend basically goes, these days that's mediated to Walmart. - Yeah, that is what seems to say Walmart. - Yeah, Walmart. So the idea is that, you know, 'cause maybe 'cause it's like, you know, Walmart, you know, they're so crafty Walmart. But yeah, the idea goes that Walmart, you know, they sort of data mined all their sales data and they discovered that, you know, young males or even any male will go to, you know, to Walmart, you know, to pick up diapers for the kid. And then we'll then buy a six pack of beer. So there's this connection between males, especially at night buying diapers for the kids and then treating themselves with a six pack of beer. And then the story goes, Walmart then started to put diapers and beer right next to each other, which then, you know, drove sales to sort of all new levels. And, you know, it has that kind of like, plays into sort of our biases of, you know, male fathers, you know, kind of, oh, you know, the wife sort of sends them out, you know, to go pick up the diapers. And they're kind of like, oh, you know, very reluctant to do this. And this is, you know, a bit of a maskulating job. And then they get there. And then they're like, I've, you know, I've done such a good job. I've, you know, I've bought these diapers. I'm going to reward myself with a, you know, with a case of beer or something like that. So yeah, so that is the urban legend as it exists today. And I heard, I first heard that I think in the late 1990s, but by a woman co-worker. And I think she was involved in sort of, you know, a data analysis and stuff like that. And so it was kind of something, I guess, that sort of passed around like marketers and data mining types. But when, you've heard this, obviously. - Oh, absolutely, yeah. So often I can't even place exactly when. - But I would say sometime in the 90s, well, I first encountered it. It's intuitively plausible it. And it also has this other feature I really like in a way in that it has great explanatory power for what data mining is. Because that's the type of insight a data miner would love to find because it's actionable. A business can go ahead and, you know, move the diapers and the beer near to one another. And revenue can go up. It's a simple thing everyone can understand without talking about the complexities of how data mining is executed. It's believable that you could find this. - Yeah, exactly. - Unfortunately, it doesn't happen to be the case in this particular situation. - I mean, there are cases where it's like, you know, you don't put peanut butter on one side of the grocery store and then jam on the other side of the grocery store. 'Cause I mean, people will not then buy them or they'll go to some other grocery store where they put the two things you want together kind of thing. So I mean, there are definitely these kinds of purchases that sort of reinforce each other. And you try not to make them very difficult. There's kind of a whole debate about like milk in grocery stores, you know, that milk's always at the back. And there's sort of one side that sort of says, you know, 'cause that's a very popular item. So they put milk at the back. So it kind of forces you to walk through the whole grocery store and then, you know, you'll buy other things till you get to the milk. And then, you know, you go cash out. So one thing, you know, that's why they don't put milk at the front of the grocery store. When you actually sort of really delve into sort of, you know, how does shipping milk work is they never want milk to ever leave cold storage. For every, you know, like minute, a milk is out of cold storage warming up. That reduces its shelf life and really kind of severely. So the reason milk is at the back is because that's where you can build kind of cold storage facilities in the back and you can just bring the cold truck in the back and they call it the cold chain. You cannot ever break the cold chain. So that's the reason why it's at the back. But a lot of grocery stores actually still even put milk at the front of the grocery store, but that's really why it's at the back. It does, of course, have the added effect that, yeah, sure, as you walk through the grocery store, you have to fill up your bass. You might start filling your basket with other stuff. That is a happy byproduct, but for the most part, it's just a way of keeping milk cold and having longer shelf life. Like it's sort of coming back to the beer and diapers or as people in the UK call them nappies. - Oh yeah, I'm reading that as well. - Yeah, when you sort of reach a store, you'll see a lot of times like diapers, brackets, nappies. What if diapers mean something else in the UK? It might mean like sweaters or something like that. - Yeah, I hope somebody tweets that at me 'cause I was at one point working with a company, we had some people in London and one of the people there said something about her cubby which sounded really awkward in context. So I asked for a little bit of an explanation. She said, well, the part in the back of your car where you put groceries and whatnot. Okay, well, what do you call it? Well, I call it the trunk to which she laughed hysterically saying, you mean like an elephant? (laughing) - But as in a lot of urban legends, it's something that they have a basis in some kind of fact but have sort of been widely mutated. But this one does have a basis in fact that there was a company called TerraData and they were contracted by a drugstore chamber. Not a huge drugstore, chain, OSCO drugstore. - Yeah, I know how. - And okay, oh, okay, maybe it's not familiar to me but we're just starting to get, what was it, Rexall? Do you have Rexalls there? - I don't know Rexalls, nah. - Okay, well, Walgreens are always like, my girlfriend loves Walgreens. They've got some sort of mascara there or something you can't get in Canada. But anyway, so this chain of drugstore, so they wanted to kind of find these sorts of associations and their idea was honestly to sort of find these associations and then make the whole drugstore just as efficient as possible for the consumer, so things that they want to buy together, they don't have to trip across the store and stuff like that. So what they found was that it was between five PM and seven PM that when they were looking at sort of even they called baskets, what cash register they take, what is on that cash register take for each sale, between five PM and seven PM, they found like a sort of statistically significant combination of beer and diaper purchases. But only between a very limited number of hours and they believe it didn't really depend on the day, it was all every day of the week, but they didn't know the sex of the person buying, whether it was like men or women. But they never actually even did anything with that actual data, like they didn't sort of go, oh, beer and diapers, so that's between five and seven PM, let's put them together. So the whole, the store was able to sort of drive and crease sales, putting these things together, that is sort of a later addition. But yeah, so there was sort of a correlation, but no one was able to figure out the sex and no one ever really attributed any kind of explanation as to why there was this correlation. - Yeah, it's almost a strange one in the, you have to wonder if this makes a challenge I've seen, you know, in my own career and the work of other data scientists is sometimes you're brought in, there's a lot of money on the line to find something like this and you feel bad coming back and say, well, we didn't really find anything, here's our invoice, so. - Yeah, yeah, exactly. - In a lot of cases, I see people kind of on fishing expeditions and five to seven PM is an awfully specific time range. Also, the time people get off work and go shopping in a lot of areas, so I don't know how to take it in context. I mean, obviously don't have their data to verify anything, but it seems kind of spurious to me. What do you think? - Yeah, well, I mean, what do they call it? P value surfing, where you just start with this, yeah, big, big data set and then look for statistically significant correlations. And then, and I think you commented on this too, but you like, you'll see a lot of times like an insurance company wants to make them get some free publicity newspapers, so they'll crunch their data and they'll be like, people who are born under the horoscope sign of Taurus are the worst drivers and, you know, Libra's are the safest drivers and you go, okay, well, you know, I mean, if you're gonna sort of make that kind of comparison, I mean, there has to be a worst and there has to be a safest, but I don't think what they do is they don't sort of go, okay, we're gonna divide a random, we're gonna randomly divide a data in half and then we're gonna see if there's any significant correlation between, you know, the horoscope, sun sign and safety and then, okay, so we, oh, Taurus is our worst drivers and now with our other data, you know, the other half of our data set, now we're gonna, is that predictive of and then do they find the same correlation in the other data set? And then if they do, then they can go, oh, okay, you know, that lends credence to the hypothesis, yeah. But I don't think they'd ever really tend to do that. - Yeah, I don't see a lot of that, or even more so extending it and saying, okay, we've cross validated and now we have this belief that it's, did you say Taurus is our worst drivers? - Yeah, like Taurus. - If that's true, then we ought to say over the next six months, we're going to see this many accidents from Taurus's and this many from Libra's or whatever the other signs are, I don't know them, but it should have some predictive power and I never hear studies that way, like, oh yeah, we moved the beer and we saw this much increased revenue. - Yeah, yeah, that's the other problem too. Again, like the Hurricane Sandy thing, right? You know, it's like, oh, you know, people are like, there's gonna be a baby boom and then, you know, very few people then, like nine minutes later, actually go and look and, you know, so you do simple Google, so it goes, oh yeah, no, didn't happen, but then of course, you know, like people who are really committed to the idea will just go, well, Taurus is so warned have just improved their driving. (laughing) - Right, yeah, it's almost like absence of evidence for the conspiracy is evidence of cover-up. - Yeah, exactly, yeah, there is no, people can always rationalize any outcome that contradicts their hypothesis, right? - Yeah, I actually hope to do some other episodes in the future about these grocery store type problems because when I look at it, I tend to think Occam's Razor is the most explanatory. Your great example about why milk is in the back as a very intuitive explanation to me, but it's such a complicated space. You know, moving the beer to be near the diapers probably means you moved it away from, you know, the heavy stuff, so the guy who might've bought a more expensive bottle of some fancy tequila or something is we're not gonna see the tequila, it's incredibly complicated problem, I guess, to think of it in a global way. - And ethically, do you really want to encourage drinking and parenting, like is that something? - That's a good point as well, yeah. - Yeah, we keep diapers by the crack pipes. (laughing) Is that real? The most ethical thing you can be doing, we make money. - There's kind of a part, it harkens back to what you were saying earlier about urban legends being, you know, you're in the know, that gives you an added advantage and I see some of that in the data world where someone will say, well, we crunched 30 petabytes of data and we ran this analysis and we came out with this result and I see that sometimes we're reluctant to people to say, well, prove that to me a little more because I don't know about those 30 petabytes of data or whatever else. I guess I'm wondering if you find that that's, can spawn a new breed of urban legends if that's part of the recipe that makes a data mine are capable of generating new ones? - Yeah, I don't know, I mean, I'm not aware of too many, I'm a real data mining urban legends other than that one, but yeah, I mean, that's one that I kind of heard and just kind of believed, you know, like, oh yeah, it makes sense. - Yeah, I've seen it in textbooks, so for what that's worth, it's a pretty prolific claim. - Yeah, but I mean, one thing I have found is that sort of in the absence of good information, people will like, like, why is this thing happening? If they do not have sort of a good explanation as to why a certain thing is happening, people on this bizarre need to interpret it as like, it's happening to make me a victim, you know? - Uh-huh, yeah. - Yeah, people have something this weird need to sort of make themselves a victim, even though it kind of has this weird negative outcome. - Yeah, I've never thought of it exactly that way, but I think that's quite astute. I hear a lot of times people will say stuff like, oh, well, they're tracking me because they want to sell me more things. And the flip side of that argument is, they want to put things in front of you that you actually like more often, which sounds like a service. So I'm always doing a little porn on that, you know? I mean, I'm all in favor of privacy, but it's hard to frame something correctly sometimes, I think. - Yeah, exactly. I remember I had a friend who worked for an advertising company, and my friend does not have a lot of needs in life. Like, he likes to read the newspaper, he lives in his, you know, one bedroom basement apartment, and that's what all he needs in life. And so, he found himself working too much of this advertising company, and then sort of went to his boss and said, you know, can you cut my hours? Like, I don't need all this money. I just want to work, like, instead of 40 hours, I want to work like 30 hours. And they're like, oh, okay, you know? And, but then his other co-workers, you know, are married and have mortgages and stuff like that, they saw, you know, they could not conceive of somebody wanting their hours cut. So, the only way, their only conclusion was that my friend was earning so much more money than them, that he was able to sort of cut his hours and still, you know, maintain this, you know, lavish lifestyle of a house and two cars and three kids and stuff like that. And so, their interpretation was not that, this guy just doesn't need a lot of money and is maybe just a little lazy. Their interpretation was, we're being screwed over here. - Yeah, I think that is very thematic of a lot of these. A lot of the urban legends I was interested in as a kid, you know, these and others, even if we went into more of like the, you know, the hook is in the handle of the door, kinds of stuff, it's hard to really, or at least, I think it was hard to ever track these down before the internet was around. And I thought that just the ubiquity of these affordable devices we carry in our pockets to fact check things would have tempered a lot of the urban legends because I ought to be able to Google for that and at least find a local newspaper. It's not like back in the day when you'd say, "Oh, you're just, the search engines don't know about it or something." - What's your sense of that, has the internet made urban legends go on the decline? Or are we just in a new era of different types of them? - Well, I mean, it definitely thinks of something like snopes.com, I think that is, it's really easy these days to sort of nip those kinds of urban legends in the bud. Obviously, things that are sort of very political, like, you know, Obama's the first president and not salute soldiers properly, you see all these things that a lot of those things I think are, you know, or George Bush is the first president to do something terrible, that you can pass people those, you know, links to those snopes things and they're still, they're still not gonna back down from that kind of stuff, things that are very sort of political, but, oh, I mean, around the fringes, I think, you know, snopes.com sort of does a lot to sort of kill those sorts of things off. And obviously, you know, things like Facebook and stuff, I'll immediately see people passing on links to snopes and sort of, you know, killing a lot of these sorts of things in the bud. And people, people generally almost all seem to have collectively agreed that, you know, snopes is the final authority on these urban legends. - I think it's a great resource and I've lost many an evening, just wandering through articles. - Yeah, yeah. - So do you think that sometimes I see these things passed around, you know, there's always a game of telephone, the players change, the names change. It's Walmart, it's Target, it's whomever. But every once in a while, I'll see precise figures creep in. Like, I think it was, who was the guy we were talking about earlier, who did, who falsely claimed you to the experiments in the theater, Vicry, I think. His claims had percentages attached to them. It was like, Coke sales rolls 18% and popcorn rolls, 58%. Did you at all, and when you were looking at urban legends, look at the presence of numbers, whether they hurt helped or were irrelevant? - Well, I mean, I have read, and be skeptical of this, but I have read that when people lie, their stories tend to be really specific, a lot of weird specificity in their stories that sort of are lies under the belief that it makes it sound more true. - Yeah, no, I've definitely seen that as a feature, definitely. - Actually, one of the urban legends, that's kind of an urban legend that I actually, I actually predicted was, you know, I lived in Korea and so I sort of follow Korean news and stuff like that. And one of the Korean newspapers has, you know, has sort of like, you know, learn English kind of feature. So they will take news articles and different articles, and then sort of run the English and then run the Korean translation. So for fun, they took something from the onion about Kim Jong-un being the sexiest man alive. And they took that from the onion and then sort of did a Korean translation and used that as sort of a light and fun way of sort of teaching English, especially teaching American senses of humor, which is very radically different from Korean senses of humor. And the Korean, they make it quite clear that this is a work of parody. Americans love parody and stuff like that. And I was looking to some going, you know, other news services are gonna pick up on this and they're not gonna realize that this is parody because they're not gonna be able to read the Korean, but they're gonna, it's kind of in their feed and they're gonna see a Korean newspaper reporting this and they're just gonna pick that up and throw that into their news feed without realizing that it's parody. And sure enough, it just sort of went mad and among like Asian online newspapers and stuff like that, Kim Jong-un has been declared sexiest man alive by this American newspaper, The Onion, and as if it was completely true. And I was not at all surprised when, you know, a few days after this thing running and I think it was like the Korea times or something like that that it just hit all of these sort of Asian newspapers as if it was a true and factual story. - How funny. Yeah, I think there's a great study for someone to do here. Maybe a good grad student's project of how bad information starts from one source spreads and whether or not, you know, a post of snopes can cut off a little tributary of spread or whatnot. I'd love to see some work on that. - Yeah, I mean, you know, the best piece of advice I always give is like, if you're searching on the article and all that you're finding is basically the same article as being repeated by all sorts of different sources, slightly rewritten, then, you know, be very, very, very sort of skeptical. - Yeah, as I do with all my guests, I like to ask for the benevolent and the self-serving reference to references to things can be anything, a book, a website, whatever you like, that you think the listeners might enjoy. - Oh, let's see. Well, I mean, I guess if I have to offer snopes.com as something you'd enjoy your ultimate time sink, I suppose. - Yeah. - The only thing I would snopes is I like to sort of like kind of do a control A, control C on web pages and then just sort of paste it into a Word doc and then create a PDF and then throw it on my Kindle. So I don't spend a lot of time sort of at the computer, kind of like reading blogs and things like that, but snopes makes it really difficult to sort of copy their content and paste it into Word and stuff like that. So you have to kind of turn off some JavaScript or something like that. But they say, look, we spend a lot of time researching these stories and our only revenue are these are banner ads. So we just can't have our material sort of floating around out there. - I suppose that's fair. And your self-serving reference, what can people find something that benefits you as directly as possible? - Oh, I guess, I guess, you know, my conspiracy skeptic, yrad.com/csforconspiracy skeptic. - Well, it's been a pleasure. Thank you so much for being my guest. I hope everyone listeners enjoy and well, if they're not already, check out the conspiracy skeptic as well. - Okay, great. Thanks a lot, Kyle. - Thank you, Kyle. - Thanks for listening to the Data Skeptic Podcast. Show notes and more information are available at www.dataskeptic.com. You could follow the show on Twitter @dataskeptic. If you enjoy the program, please leave us a review on iTunes or Stitcher. A review is the greatest way to show your support. (upbeat music) [BLANK_AUDIO]