17 Burst results for "David Spiegelhalter"
"david spiegelhalter" Discussed on The Economist: Babbage
"A fortnight ago. We ran a competition to win a copy of David Spiegelhalter book on statistics. We've received hundreds of clever and comical entries and my long suffering, and honest, producer and I are reading them now and we will announce a winner soon, stay tuned, and we'll have more books to win upcoming shows. And finally, to space Amazon's founder Jeff Bezos last week. Give it presentation on his vision for the future of humanity. This is at high school, and I want to highlight this quote the earth is finite, if the world economy and population is to keep expanding space is the only way to go. I still believe. What happens when unlimited demand meets finite resources. The answer is incredibly simple rationing. That's the path that we would be on. The goodness. If move out into the source system for all practical purposes, we have unlimited resources, so we get to choose do we want stasis and rationing or do we want dynamism and growth. We're out in the sources. We can have a trillion humans in the sources, which means a thousand Mozarts thousand Einstein's, this would be an incredible civilization. What could this future? Look like where would it trillion? Humans lip. Alon mosque lookout, there is a new billionaire space cowboy in town to discuss Jeff Bezos astronomical. Visions, I'm joined by the economists technology editor Tim cross. Hello, tim. Hi, Ken, Tim. What is Jeff Bezos trying to do so his Otomat goal is I, I suppose, is sort of similar to Elon Musk, which he wants people to get off earth and into space reasoning is slightly different. So musk has always said he wants us to be a multi species. In case anything happens to earth. It's nice to have, you know, self reliant colony Omar's a backup justification was a bit different. And he said, well, if you look at the rate at which energies, for instance is rising rising about three percent year project that food into the future. And you compounded up on pretty soon, even if which to entirely soda power, we need to cover the entire planet and solar panels. So he says, essentially earth is gonna run out of resources if we want more are loads out there in the solar system, and that, so we should go, so he unveiled plans for. Mood. Lander sort of the first phase and said he would Sal about a billion dollars worth. Amazon stock year to fund is rock company blue origin to sort of get this going and then in the more distant future. This is where he sort of diverges from Muscat, because musk wants to build a settlement on Mars. These was talking about living, not on a planet, but in space itself. Okay. And Howard humans living space itself will, so he revived an idea that's been kicking around lease the nineteen seventies impulsively bit earlier than that, which is to build sort of gigantic space stations, essentially, Hugh, cylinders in space and to live on the inside surface of those. So there was an American physicist back in the seventies Giradi Neil who sort of sat down and work through the, the very broad brush engineering strikes, and he came up with this idea what you would build a Senator, that's about five miles wide, and about twenty miles long. And then you would fill on the inside surface, you would put housing you get to these things, you would stick together end-to-end, you'd spin them opposite. Actions. Have they stabilize each other? And the spin would provide something a bit like gravity at the surface, you feeling with. I'd essentially, you get a absolutely colossal version of the s that was so big. It would have whether inside and people could live in these things, and it's just floating in space. It's just floating in space. It it's it's powered by the sons so you have this array of mirrors, and sort of big windows running down the sides of the cylinders. So you sort of be the sunlight in provides a day-night cycle all the tryst. He comes from solar power. The idea, I think, is that the ecosystem cement to be sort of self contained, so they need either know very little stuff, bringing in from earth. What problem does the solve? Well, so this is the question. So these like I said, frame this in terms of energy, use resolves consumption basically said, you know, if we want life to continue getting better at the rate, it's been getting better. Our results is going to go up, and eventually it's going to cross the ability of the planet to provide. And so, you know, I guess you could call it the sort of capitalism, implies space flight school of thought the only problem. I guess without is if you look at the population trends of the UN is projecting what seems to happen when people get rich enough, they tend to start shrinking families, and they often treating down below replacement level. So lots of advanced country. Now worrying that populations of falling not rising and the UN's own predictions is the earth's population will peak sometime next century and then then start to decline. So it might be that this three percent, energy gross figure the bees is using doesn't continue indefinitely into the future. Let's look at the technology itself is this technically feasible, it seems to be. So, like I said, this, the physicist, who I don't wanna say invented these things, but he sort of sat down and rigorously thought three of them. Gerald aena. He published a book about it called the high frontier nineteen seventy six and in terms of the broad brush engineering. Yeah, I mean, we have metals that are strong enough. We have a good enough understanding of physics to know that, you know, you can solve all the problems with rotation. Rotation is a reasonable they're not perfect analogy for, I would have real. Gravity on, if you make the cylinders big enough, most of the sort of weird, wonky effects. That would let you tell it from real gravity become kind of too small to notice the design works on paper, lots of things were paper. The trick is actually building the real world and all kinds of difficulties with no doubt read ahead. I mean, we talked earlier about building a self contained biosphere. So we don't really have any idea to do that said that I think Hugh applied enough brainpower, and we're really determined for some reason to make these things reality. There's nothing in the lose of physics. So you couldn't build something like this. Whether there's something in the laws of economics say it will be to ruinously expensive. That's another question. So what sort of timeframe is Jeff Bezos thinking for the moon landing, he's thinking pretty soon. So the first the very first phase of his plan is to go back to the moon and it's American policy at the moment to go back that by twenty twenty four how seriously, we should take that next? We've heard that several times before we hear from Obama. We had it from Bush before him, but beezus Cain twenty twenty four month. You've been optimistic but, you know we've been to the moon. We know. How to do it. If we really, really wanted to we could probably get back there within a few years, the moon bay stuff. I mean I you need to make traveling to the moon routine, which it's not you then need to build the moon base with the asteroid mining or whatever it is. And to build the cylinders themselves when even bees also I think he's talking hundreds of years when it comes to this stuff he's not talking within his lifetime. So in some ways, he's taking a sort of attitude towards his legacy rather than something that he's going to achieve in his lifetime. Yeah, he is the analogies. You often hear these themes. You know, people started building cathedrals in the middle ages, and they need right? Well, they wouldn't be finishing lifetime, but they built in any way and you can always fall back on the document, which is what happens when he's just press the space gets, and they say, well, look in the very long term, the choice between spaceflight and extinction. Because in hundreds of millions of years time, the sun is going to become too hot for life on earth to carry out and so a scene. We haven't nuked us to dust by that seeming, something around looks somewhat like modern humans, there are going to have to leave and find. Homes elsewhere because the eleven we have forever. So it sounds like it's a good way to hedge once bets if you went into look sort of eight hundred billion quarters. And will prime members get there quicker. Maybe they'll get next century delivery guarantee. You fantastic. Tim. Thank you very much. Thanks can. And you can read more about space colonization in the upcoming addition of the economist, and if you're still on earth, and having traveled into outer space, yet, you can take out a subscription. Just go to communist dot com slash radio. Offer to get twelve issues for twelve dollars twelve pounds. And that's all for this week's please read us on apple podcasts. It does make a difference. I'm kind of kooky and in London, this is the economist.
"david spiegelhalter" Discussed on Monocle 24: The Monocle Weekly
"They're always nurses are always the top and doctors a second. But after that becomes professors and scientists so is pretty good and a lot better than joy. Lists politician who dine on the Balsam next to state agents and used car dealers. So although everyone quits, Michael, but Michael guy was being unfairly quoted that's fake news itself. I think because everyone cut him off by his sentence when he said, everyone is not enough of experts, but she went on say some extremely sensible and with which I agree. He went on to say say that the experts people have had enough for those who make predictions, and then don't take accountability full them when they go wrong. And I think was fair enough. That's fair enough. I suppose as I mean, not quite taken out of context to this sense that there is a disconnect from a technique Crecy. But maybe that is well among some the may be I mean, there is a well, the has been extremely vocal minority with what you might call fairly Antezana views, very deeply skeptical views about the claims made about whether it's vaccines climate change or anything that's about this expertise. What these scientific -chusetts saying, and they have very vocal and never. Prominent on the web. If you Google vaccines up become their websites and things so that's the Golden Globes very organized. And that's fine is not the majority and any sort of actual survey. Everything it says that people want more evidence based policy they want to hear about the evidence. The are interested in in evidence in why something might be the case they want to know the argument because they don't want to be bombarded with just numbers and facts and technical, but they want to know why people believe things one of the examples that comes up in your book as of miscommunicated statistic, this idea of the bacon, sandwiches and Kloss one cuss engine his bad as cigarettes special. So it might be how do you and have you worked to kind of back to communicate that kind of could you explain? A wonderful example, which do dry out quite lucky bacon was cast fide by the W H is a class one carcinogen this you say in the same category as us special cigarettes. It didn't mean as was reported by some of the media there's dangerous cigarettes. What it meant that the evidence? They were constantly genyk walls, very strong and believe it, but you can is Jenny full stop. I haven't stopped eating bacon. I.
"david spiegelhalter" Discussed on The Guardian's Science Weekly
"The final chapter of your book looks at how statistics might be done better. What needs to change? And do you think any of that will help boost people's trust in stats? Yeah. When it comes to trust. I made the point. I always quite Banus, Nora Neelam. This is that you know, we shouldn't try to be trusted what we should be trying to demonstrate trustworthiness the office for national statistics. Now, it's number one pillar in the code of practice statistics. And the government is trustworthiness. So it's up to the statistics community to demonstrate that trustworthiness, and that means transparency about the limitations of the data. But also transparency about how we shouldn't terported. Well, I think that there should be the strongest attest to voice at everyone's responsible for this from the funders research from the peer reviewers and authors editors everybody is responsible. Got that part on journalists. Partly got that part to play in improving the trustworthiness of what eventually. Gets passed out for the consumption of the public or politicians. We've all got a part to play. I think just the realization of the problem is is the first step, and I think things are happening at the moment are very good. Many. Thanks, David Spiegelhalter for joining us this week. You can find linked to his new book the art of statistics on our web page. Just go to the guardian dot com slash podcasts. And navigate to this episode of science weekly. If you want to get in touch with us, you can the Email address is science weekly at the guardian dot com. Thank you for listening. And until next time goodbye. Great podcast from the guardian. Just go to the guardian dot com slash podcasts.
"david spiegelhalter" Discussed on The Guardian's Science Weekly
"Hey, Google speak to the guardian briefing. Welcome back to science weekly. I'm nichole Davis. Before the break, we discussed the golden ages the districts, and how it transformed medical science next. We turn to today and look at the state of two sticks in science. Recently in the journal nature. David Spiegelhalter and more than eight hundred other signatories argued that scientists shouldn't use something co statistical significance to evaluate their findings, and these eight hundred signatories have said this is really awful. I mean, people have been saying awful Frazier's. But it's reaching this some six hundred crisis put statistical significance is a mathematical measure in nutshell. Scientists use it to see if any difference observed between two groups being studied Lille or whether they're simply down to chance start with p valley's peeve developed by Fisher one hundred is a brilliant idea of basically measuring the compatibility between data and the hypothesis. So a lot of what's knoll hypothesis setup, which says usually drug doesn't have an effect. There's no factor. In other words, we're just status quo, and then the piva you. Data and the people who is high capacity today trees with the effect the probability of observing something as extreme as you did if the null hypothesis which peeve alley is used to describe probability. That is the likelihood that a large different seem between the two groups might be down to chance nother was they help scientists to figure out whether these are important on not but David says boiling results down to important or not could mean will losing out on crucial information. People would do all this enormous data collection in this very subtle analysis, and this and then they make dot cultivars into a wider at significant or not significant and that I caught me would be bad enough to suddenly to try to take all out subtle science and decide this is a true discovery or it's not is it damaging though in does it lead to people getting research. Crohn's full work that's nonsense. I wasted taxpayer's money. Or is it damaging in terms of people thinking drugs work when they don't don't work when they do. What is actual? Impact on the world on society within drugs. I mean, it's a tightly regulated area. And so in a way, they've got that sorted. I it sort of works. I think much broader scientific areas where but also of testing medical interventions and so on and as I say this to is two things that can go wrong. I is and that you might decide something significantly less than twenty five decided significant and think that means you'll ninety five percent shorts. It's case, and that's that's wrong. So in other words, they're full discoveries being made all the time, and you could say that the reproducibility crisis in social sciences particular neurosciences as well is you to these Bassett number false discoveries. But the other the other he can run the other way is that you get you get false non discoveries that you you say, oh, this like this treatment of mentioning a things get perfectly what treatments could be perfectly effective. Just get thrown away ruled because they did. Have a significant faint does something need to replace the typical significance than or you saying you get rid of the statistical significance problem by just saying not actually saying what your study concludes in a way, you can include a law without the conclude with trying to say it's turning it into a yes, no discovery nothing and conclusion, and that, you know, with another we should we should have p values. We should have estimates. We should have confidence that divorce, and particular we should be saying what does this evidence had to what we already know? And that's taking out says, it's like more basin perspective is that what would a reasonable person based on the currently vailable evidence. How would this change their mind? I think that'd be much more proprio you to to describe the importance of of a piece of research rather than trying to turn it into a. Yes, no issue. Some say this. Yes, owner page is just to duck Sionist, and it could be why some increasingly distrustful of stats and science reportedly because a number quashed at us doesn't seem to reflect oven reality. Let's say you're told that unemployment has declined since the recession of two thousand and eight, but if you're struggling to find work, and so your friends, this number doesn't feel credible is no what you're experiencing. In other words numbers are just too simple to capture the lived reality. This is what academe ick William Davie says in a guardian article from twenty seventeen I really recommend reading it. If you want to understand more about his take on how individuals can feel a disconnect between statistics and their reality. We've put a link to the article in this week's episode description at the science weekly podcast pages at the guardian dot com. In some read it to and put this point. But a clash between reality and statistics today. So I rewrote that school. And when I first read it really annoyed me, but reading some good points here in the first thing is to realize this is only about ficials desisting in the migration unemployment GDP and things so this is about statistics in the political discourse. So that's the first restriction. And that's only one area stance. The other thing is that the the evidence is that people have not lost. Trust officials to cystic remains extremely high in this country. Elsewhere. The point he made though, which I with valid points. I think with a couple of them first of all is that these monolithic constructed statistics GDP, quoted, a a nation level, and particular migration, you know, extensive migration improved EP actually that is not a good argument to use for somebody in an area of the country, which. Brexit campaign showed very clearly who might not feel that's has been benefited from. And an and there's some what's the quote, somebody said was that of dressing meetings. Oh, well, they may be your GDP. Not my GDP really encompasses this idea, and he's up to your right that quoting national figures, whether it's crime migration unemployment actually that. That's not helpful for people to understand that you know, what's going on in around them. And I think there has been a problem in the past of of studied statisticians get into quoting some big national figure, although I do more blame politicians nothing decisions for doing that. And that the deconstructing those numbers to fine granularity is terribly important. The other point that he made was that. This was re this was written before the big Cambridge scandal with saying that the big corporations and in our Facebook, Google things like that. Then in the ultimate you quoting Cambridge. I liga. They are now taking control of data. You know, the all the routine data being collected social media and elsewhere. This was all in private hands. And that this was going to dominate the the official stunts. I think, you know, this again was a reasonable view. But even in two years things have changed massively. First of all the demise came John? Let's go and the the considerable pushback now in terms of the responsibility of of data and things going to change in the future. With the news is endless new organizations sent of data ethics innovation they love licensing country. Elsewhere, and I think people in the future will look back at the time. So he's describing Mike oh that this wild west of data hired people operate in that way. So I think the controls and the corporations will grow enormously coming back to what you were saying about trust in new saying that people actually still do trust official figures. But there's no doubt that we see examples, maybe. This is just because visibility is changed. But we see examples of people dismissing figures, whether they're official not these can be they may be migration figures, they may be crowd size figgers at inaugurations or even marches and things like that. There is a sense. I think that people are withdrawing retreating into sort of isolated realities. There is less of this share reality. I mean, you're talking about extremely vocal subgroups of people. Even if one of them is Donald Trump, you know, these are still minorities of people who have strong views these large voting population. It's always been a big distrust of the federal government in the US in the way that hasn't been in the same way in this country in the byu, very happy really negative thing that started developing in this country, but in this country, you know, I think they're all vocal groups in Amazon's anti statistics things that the strong as has in America. So they're all focal groups, but there's no evidence if you look at the trust ratings, but everything of that to trust in in scientists and professors nationals remain incredibly high very high. Indeed politicians, really love. But that we have seen situations where people just flatly refused figures that are put in front of them, whether it's about economic growth, whether it's about immigration, whether it's about vaccines, and the successive flaxseeds stuff a lot of people. And I think this is growing maybe it's from a small base. But it I don't know it. I have no proof of. Yeah. I see you anti fax seems to be getting into countries. I I'm surprised it's getting into Denmark Japan. This would terrify me if I was the statistician where my work is supposed to lay down the facts. Brought a lot of interest into into fact checking and correcting misinformation people say they don't believe the numbers. Because of the motivations because of what they believe they don't like what they're being told to do this a spacious skeptical, they general destroy. You know, skepticism fat science, whatever like that. And so this comes as reasons, I don't believe numbers. And there we disagree with any argument you make and so it's not a matter. Just correcting the numbers in a sense. It's it's it's much. Even that does is is quite tricky to change someone's mind, when the this is a know Masai, cultural tribal identity that people have adopted, and I think that's the big problem. There is that, you know, wonderful social media, and what it's done is is is nailed minority voices to group together into these tribes and to reinforce them beliefs in particular ethos, and to then reject everything that could question this. That's that is quite a difficult thing to counter, and I think, you know. The reductions tribes is something that is way beyond statistics and evidence quantitive evidence. And I think that's generally shoot of concern for the future. The
"david spiegelhalter" Discussed on The Guardian's Science Weekly
"The the guardian. Statistics subject, many struggle to gloss. But not professor said, David Spiegelhalter, an of never stopped having that in a way, you know, emotional engagement in the importance of the subject of understanding about uncertainty probability chance, and what we can learn from the information around us, but how much can data and statistics really tell us. Some argue that statistics are losing their influence and ability to objectively depict the world around us. Anyone who collects data and tries to learn from it knows immediately the amount of judgment, and the Manta discretion that goes into the the planning of the study the collection of the data and its interpretation. This is not a cold objective rule-based exercising. Anyway. I'm nichole Davis and on science weekly. We ask why some people beginning to distrust the testing. I'm professor David Spiegelhalter. I'm from the university of Cambridge. And I'm chair of the Winton center. Full risk evidence communication David came into the science weekly studio to discuss his new book, the art of statistics with Ian sample. Why write this book this time? We hear a lot about data and society today. It's it's a fantastic time. They know the so much more information available, and this is being used in all sorts of ways as being used for technology in our phones as us for recommendation systems on the internet, but there's still a really strong role for using data tries to gain knowledge to understand the world better. And I think the idea of what can we learn from this numerical information, I'm is more important than ever. And yet the lessons a lot of them have not really been learned. And as I. Tight in the book. I think that's because it's actually quite tricky. It's the only book I've ever read that congratulates the reader for reaching the end, this stuff isn't easy is it. No, it's not. I always joke that you know, I get asked. We know why do people find probably statistics difficult, and then you know, after working in this area for forty years. I finally concluded that it's because proteins mystics really are quite difficult. There's some very subtle ideas around. Tim's the technical tools that get used. But there's also a great subtlety because this is not mathematics. This is not starting from set of principles and deriving some results by set of rules. There's always judgment involve there's always knowledge involved. And so that's what I sent. Why call it the artist assistants tell me the wind. Did you first get a tasteful statistics? And what got you hooked? If that's what you are. I am hooked completely. And it happened quite early. I did maths it university and pure maths. I and I really liked it. But the by the second year but hoffy through the second year. I suddenly realized my brain could not cope with anymore. I just bang my head on the ceiling of distraction. I just couldn't do it. And you couldn't I just stopped. So I do and I was doing some stats and that she's very mathematical stats. I didn't really like that either. But I I suppose like so many people had an inspiring teacher who actually had a passionate view about the importance of statistical inference, the ideas, what is probability, and I've never stopped having that in a way, you know, emotionally gauge -ment in the importance of the subject of understanding about uncertainty probability chance, and what we can learn from the information Arandas, what we might get sense from you. And it might be crazy to get a really high level of helicopter view of what you. Think statistics is done for the world. Like, what was the world like off the statistics came along with with this new power? We'll sit chain. I I mean, my background lodging medical statistics is total revolution in medical statistics. Before the second World War medicine and the chance was anecdote. It was what school eminence based medicine. You know, you're a poet in my experience, boy, we found XYZ works. And then you know, off the second mobile with the development of randomized trials, which chemical some agriculture, then since the first randomized trials streptomycin if thing late nineteen forties, and then also development of academia, g with discovering link during lung, cancer and smoking is extraordinarily Harry. It is totally revolutionized the practice of medicine, and that is really based on statistical analysis, which is why still medicine is the area where stuff decisions given. Absolutely equal status with with senior with medical researchers.
"david spiegelhalter" Discussed on Amanpour
"Of Ninety-seven percent thought. So what is the lesson then from this, oh to be very cautious about when people make a claim that one thing is causing another one of my favorite examples was a study in the US that says, you know, sodas cools violence, and will they did the interviewed a whole lot of teenagers. You know, how many sodas do you have a week and violence on you? And the the kids who have more side is way more history of violence and they soda must be causing the violence. And I think why why why is it? I. To make a joke about you could be the other way after hob as violence. It makes you very thirsty. Dean, maybe you feel like a soda again. No, this this is this could be an issue. What are the most serious ways? They can't come up that confused the public perception. Oh. Back scenes and autism. If to get on the subject that of course, it's not the tool contested. There is people get the vaccine the vaccine to bite the same time that no kids might be diagnosed with autism. But you know, they can say that they actually this week we've been discussing in Neil, stay of the easels aren't Greg banned from public spaces because they don't have the herd immunity because they haven't been vaccinated. So there's an association between the age of getting vaccine and the age of which people get diagnosed with autism. And. Then turned into a courtroom. And the president of the United States has sought of lent his support to this idea. So how dangerous is it? When a leader of the of the world, does you know correlation and causation on not someone to lecture the leader of the world. But some it is an important issue. And again, just something you should try to recognize when it's being dumb when an association in time just like those crazy figures just an association over so many years. That doesn't mean that one thing is causing another. It's been puffy lodge perky sensible, people are not nothing, but a cool that feelings can be really manipulated by people using data said, David Spiegelhalter. Thank you for joining us very much and next week turned from uncertainty about numbers to downright resistance to cold hard facts in his new book, dying of whiteness physician. Jonathan mezzo details. The dia consequences of some right wing policies for the populations. They claim to help from gun control to healthcare mezzo finds that some white Americans would rather die than betray the politics of their identity. He spoke with Harry, strain of awesome, what is dying of whiteness, dying of whiteness as a story about the politics that claim to make white America. Great again, if you're a working class white American and making your own life harder, sicker and shorter. I spent seven or eight years. Over the course of my research looking at the everyday effect of what happens if you're a working class white American, and you live in a state that had policies like cutting away healthcare and blocking the Affordable Care Act. Allowing the easy flow of guns massive tax cuts that defended roads, bridges and schools. And what I found was that those policies that were supposed to make your life. Great ended up from a health perspective making your own life as dangerous as risky as did secondhand smoke or as best as our car crashes, these policies themselves really functioned as risks to your own health. So what is whiteness in the context of this book or what I look at is really what I call racial resentment linked to whiteness. And so what I'm tracking is a story of the ways that politics that are anti immigrant anti-government progun kind of about what are called backlash politics that are couched in a kind of racial resentment. This resentment that basically minorities or immigrants are taking away. Villages that are due to white Americans. What I track is the ways that those things I work their way into state level policies and then into national level policies. You're separating out that this is not because people are racist. But that the policies have racist component. I interviewed many people over seven or eight years for this book, I certainly encountered many examples of overtly racist sentiments, no, no doubt about that. But ultimately what I found is that the health risk to working class Americans came not from their individual racism or intentions. I didn't really try to assess what's in somebody's heart. I didn't know. And I think that's very complicated. What I found was that the risk came. If you lived in a county city or state that had these kind of racially anxious or racial backlash policies that dictate your health in a way these backlash policies. I mean, this is a long time coming this is not just from the election of Donald Trump or a specific event. What we're seeing? Right now is that these these tensions that have been brewing for quite some time. These concerns again about immigration about about taking away people's firearms about the overspread of government ABC. You're absolutely right. They've been brewing since the forties fifties and sixties. But right now, I think is a particularly urgent issue because these policies that have been localized just to either the extreme, right or two particular states now are impacting national policy. And so the implications are much broader before
"david spiegelhalter" Discussed on Amanpour
"It. Would you agree that the world could use a little more kindness? If so then you're going to love butterfly coins. They're real coins that feature a beautiful butterfly on the front and attracting number on the bag. You do something. Nice for someone then give them the coin and tell them to pay the kindness and the coin forward. Each coin includes a story page at butterfly coins dot org, which shows its journey on a map along with notes pictures and videos, so here's the coolest part. The smallest of kind acts done today can lead to amazing life changing events in the future. You can literally watch the ripple effects of your kindness spread through the world forever. It's simple. It's fun and insen credibly rewarding. You'll be starting. A lifelong legacy of kindness with each coin. Join the movement today and help us change the world at butterfly coins dot org. Remember to create an ad like this one visit pure winning dot com slash. I'm Biagio Messina. And I'm joke seen we're the producers behind HSEN television documentary series unmasking killer. Join us as we explore the identification capture and arrest of Joseph James, the Angelo the alleged Golden State killer in a special ten part podcast series unmasking killer, all new episodes premiering Tuesday. February twelfth subscribed today at apple podcasts or wherever you listen to podcasts. Now, it's often said that there are three kinds of lies lies damn lies and statistics. But that's slogan is roundly rejected by my next guest and renounce status Titian, so David Spiegelhalter in his latest book, the art of statistics taking aim at dodgy data in the age of alternative fives has never been more important, and while statistics could be baffling and even misleading the Cambridge University. Professor tells me that it's all about the context and the quantity in this masterclass in math. So David Spiegelhalter, welcome to the program. So your new book the art of statistics. What is it for is it for geeks isn't the lay laypeople the novices when it comes to start and maths and the like, I guess it's record wide audience for people who want to understand better all the numbers in the news, everything they read and who wants to essentially teach them, so. Selves, the principles of statistical analysis, and that's signs to be almost a bit. Dull and many people will have done some fairly tedious stacks cool sits in their education. I think this tries to be a lot more interesting buying gauging it engaging people with with solving problems. Everything I do about solving problems using data to solve problems and using that to illustrate statistical crystals why a stat so important, and why do people not know enough about them? And that's coupled with this kind of a sold on the nature of facts and evidence in the famous people have had enough of experts. We're we live in each of data from the traditional dates of officials statistics and scientific research. We've got this massive explosion data science machine learning terific citing time and in this age of data. It's important to realize that the data doesn't speak for itself. You don't just stick the three comes on. So it requires skill and understand. Funding and cat to to ring out the meaning from the data to draw the correct conclusions with the appropriate caution and humility, so let's just blowback just assert that data and doing what you're saying. It should do put it in context ring out the meaning affects just about every aspect of daily life political life scientific life, right everything everything from the decisions we make about a health finance families are future. And of course, how we vote. It all depends on people giving us information. And a lot of that information is in terms of numbers big something is this numbers on it. Just tell you big something is is it small is it big. Should we worry about it shouldn't we worry about? And again, I think people are desperate in this age of fake news and everybody competing to be the expert and peddling whatever sometimes rubbish people want to know where they can get the best and how they can actually trust the data and the stats. Exactly. It's all to do with trustworthiness of. Of numbers. And usually when I hear a number. I'm very suspicious because I think that person's trying to persuade me something they trying to change my mind, usually, they're trying to make you frightened about something might you anxious to to influence your opinion. Then all actually just informing me. And so I think it's essential skill in order to be able to understand how numbers can be used and misused and be able to question them. Well, you go through a whole set of a framework for how to read and address and how you've come to your conclusions. And we have as a guide several pages of some of your points. And we're going to start at the beginning with algorithms, and we're going to look up there and see the first picture, and this, of course, is called doll who you call the luckiest person to survive, the Titanic disaster. Tell us why. And how you'll stat look made that conclusion. Okay, the Titanic. You may not know. But within the data science community is a lot of interest in the Titanic because there's a very nice messy data set of the passengers showing how old they would that class. They were traveling how much they pay for the ticket. They name and etcetera etcetera, even some of them that cap it on just become a sort of a big competition to try to build the best outgrow that can predict who survived
"david spiegelhalter" Discussed on This Week in Machine Learning & AI
"Hello. And welcome to another episode of twin? We'll talk the podcast y into interesting people doing interesting things in machine learning and artificial intelligence. I'm your host Sam Sherrington. In this. The second episode of our neural series. We're joined by David Spiegelhalter chair of the Winton center for risk in evidence communication at Cambridge University. And president of the Royal statistical society David who was an invited speaker at neuro presented on making algorithms trustworthy. What can statistical science contribute to transparency explanation and validation in our conversation? David and I explore the nuance difference between being trusted and being, trustworthy and its implications. For those building systems we also dig into how we can evaluate trustworthiness which David breaks into four phases. The inspiration for which he drew from British philosopher, Nora O'Neill's, ideas around intelligent transparency. Enjoy. All right, everyone. I am here in Montreal for the nerves conference, and I've got the pleasure of being seated with David Spiegelhalter David is chair of the Winton center for risk and evidence communication at Cambridge as well as president of the Royal statistical society, and he was one of the invited speakers here at nervous talking on making algorithms trustworthy. David welcome to this week in machine learning. I thank you for having me. It's our pleasure before we jump into the topic of your talk. Please share a little bit of your background. And how you got involved in statistics machine learning and kind of the confluence of the to exactly what I'm gonna start to station as you can tell and I was around in one of the l- some of of a in the nine thousand nine hundred eighty s and I was very interested in computer, aided diagnosis such as it was then and interested in statistical purchase to that using simple basin methods of the. Regressions standard stuff. And and that was exciting time. And I got very interested in this new idea of Beijing networks and graphical models. And so in the nineteen eighties really worked to develop this think the Laos and Spiegelhalter outgrow them was for exact propagation Bayes networks, we did a lot of work and that and then went into beige in graphical modeling developing the bug software for Beijing car, mock trae Monte Carlo analysis, and and so on and worked all the time in this sort of intersection of Michael machine, learning and. Pistons for those ten years be much to do with communication. I've got a job that involves communicating statistics risk evidence. And now, we got a center this strange center in the Mets department Cambridge with a gang of psychologists and communication specialist. XP people web designers. I'm very interested in producing trustworthy material that communicates numbers and statistics risks predictions, and so on okay. That's really interesting. I was wondering what the meaning of risk in evidence communication was almost anything to do with numbers. Better than public communications to ticks, right? Right. Right. Okay. Fantastic. And so you're here at nervous talking about making algorithms trustworthy. What does that mean? This your trust is very important. I've been very influenced by this wonderful philosopher in the UK, Nora kneeled who studied canton is come up with this very important idea, which has been very influential organizations developers system shouldn't be trying to be trusted though. This the wrong objective to try to be trusted. Well, they should be doing what we all. Should be doing is trying to be trustworthy and the Louis to earn that trust because that is within our control to be trustworthy. And this idea of being trustworthy has has has a big impact in the UK the national statistics code puts trustworthiness. That's number one object. Why is that nuance important between trust being trusted and being trustworthy, being trusted something you want? But other people can only offer to you big, trustworthy, something within your control got an that means really analyzing what it means to be trustworthy. Okay. And so what does that mean from a statistical perspective or how statistic? Wchs inform trustworthiness. I think that in the talk break, trustworthiness algorithm, or any social system into two components that the system itself should be trustworthy claims. It makes should be trustworthy. You should be able to rely on them. Or if you can't rely them. It can tell you. How confident today's the other thing is that what is very important is that the claims made a Bank to the system trustworthy by the developers by the commercial entity, whatever. So you not any believe the system, but you're going to believe what said about the system. Now, what that leads you into very quickly is the importance of the valuation and in my tool I draw in analogy with the hiney developed evaluation phases. You say drug development and pharmaceuticals which I've worked in that era for decades, and they're just very briefly full phases. Phase one is safety on a few healthy people face to proof of concepts done on some selected people to try to optimize the sich. Thanks three other big controlled trials, which. You actually compare it with a comparative that allows you to Mark at the drug and face for is post marketing surveillance, and what I did was draw analogy with developing our them. So they're gonna go into practice that phase one is just the digital testing that people do in on a set of test cases face to is labar tests where you actually compare it save a doctors if you go to medical system and do the user centered design for the interface and vice threes field tests where it actually goes that. And we have vanished. What is impact is which might be beneficial? But it could be harmful. You never know what side effects of might have. And phase four is the post once the thing is I monitoring to make sure it's not degrading, and that is not making the stakes. And so what I'm saying on the whole when read about evaluations, they ran to get phase one. Just accuracy test cases, some of the moving into face to comparison with diagnostic systems, medical experts and things like that. Well, was nothing phase three what? Actually is the benefit impact. When these things put into practice society properly evaluated, and I think that the claims about a system to be trustworthy than you need a much more rigorous evaluation in order for claims about assistant to be trustworthy. You need to have a much more rigorous evaluation. My sense is that we're very far from that today. Exactly. This field is developed so wonderfully. So this stuff at the conference is so amazing. But it still rule out fantastic technical capacity, very early stage because when these things don't moving into society, you find people saying they come on onto demont. It's not it's not immediately Elvis that. This is going to be good thing all areas. And so I think you know, this area is due to mature into something which which does rigorous evaluations. It's interesting. So one of the controversies at last year's Noor ups, then nips was kind of a call for increased theoretical rigor around deep learning in particular. But our current approaches to in general. This is a call for rigor also, but very different one one from Marvis statistical perspective. Rear tentative of what does it mean to actually implement this Jimmy both because you need the rigorous sort of internal analysis in order to demonstrate what it says is trustworthy because the part of the trustworthiness. Of course, this will be explanation is to be able to cite wise, come up with its conclusion justify that conclusion. And from the other statistical perspective, I type very strongly. Statisticians are obsessed with uncertainty getting the Arabs right with much concerned with the uncertainty. We are at the point estimate. And so that's what we bring. And I think again if a claim is going to be made on especially when it's making some uncertainty. Look like confidence you got to understand what that means. You could have been relying on the on the claimed confidence of of of of what is what is silent. Algorithm comes up, and you talk to you provide examples of this the kinds of claims that you envision. This kind of model being applied to and you know, what you'd expect to see or you've seen in kind of passing claim through these filters in the tool kit various examples at different phases. Some statistical ideas can come in just the phase when you're preparing algorithms on your database to decide which is the best one by ranking algorithms and how using some bootstrap methods on all the tests that you can get probability that any algorithm is actually the best rather than just producing a simple league table. Again. It's been a lot of statistical league tables, and essentially taking them part because just because something happens to rank best on one particular data does not mean, it's the best algorithm for the Fulton. Just because the football team is top elite doesn't mean it's the best team because as a wish involved, and we have all the good at trying to put a number on. So there's that us the face to again, the recent critique of systems that have made comparisons with doctors saying diagnostic systems, which Ranchi being. Pulled apart because of the lack of statistical rigor and in. Very good that got to that stage actually doing the very well doing to the standard of Riga, the wom- would expect. The face three. I talk about an trials involved in photography. Stick system is a terrible system. But she helped when is put into practice, and it's because it wasn't because what the system the computer is saying is because it just change the culture of data collection and encouraging people to make early diagnoses and be confident about that work. There's all sorts of unintended ways that systems benefit, but also unintended ways in which they might home. And so I went to those those those applications, but then went on to this idea of transparency, which is at eight is element to transfer trustworthiness, and this philosopher, Nora Neil's, got some great things to say about transparency. She thinks transparency is could be can be really dangerous. It's not an end in it self, especially in the sense of disclosure in that in a you can be very transparent, and yet nobody understand what's going on. If you could release the code that right transparent, but people hopeless hopeless. So she's really pulled a pot transparency and says what she's making this appeal for intelligent openness, which means that any information you give this really good checklist information. You give accessible people are going to be guest. It is gonna be intelligible to understand. It's got to be usable. That's going to meet the needs. And it's going to be accessible. We think somebody needs to be able to check the working not everybody. But somebody had you know, needs to get a check the working if necessary. When using learning methods. That's really quite a challenge to Canada. I did making what I thought some very nonce being done by Google deep mind with in London with morpheus hospitalizing ice Ganz in which they liberty train. This networks provide intermediate steps mention that so that it doesn't just go straight to diagnose. It's a pry ballistic diagnosis is putting into mediate steps, which seems really cool thing of and that's because that projects have to be very strongly influenced by the clinicians themselves and want that that's the way they used to thinking of act and they wanted to map the way of thinking, it many people claiming that you don't necessarily have to make a trade off in performance in order to get a much more interpretive model that actually in this Voss numbs of models Voss of options is very similar performance, especially especially as actually a lot of the differences in performs illusory that was what I talked about. Okay. And so the actually the struggle to among the great spice of models. You can use to choose one that actually Nabil's much more transparent much better explanation makes it more, trustworthy because people can see the reason you ran off several of the qualities that Nora Neil outlined. But they're all very subjective and seemed to be in some ways at odds with this the typical rigor that, but no exactly, but that's nothing. I think. Psychologist. Yes, I've very by this. I'm not even gonna try to define exactly what explainable or interpret transparent means, but you can be quite rigorous about you'll evaluation of some of these aspects example in the interfaces for the systems, we build we evaluate three things in which thing have an impact on people cult. If do they understand it the behavioral what does it do to that that behavior when their intentions and the effective how does it affect their emotions, and we wanted to measure all of those and they could be completely different. So he's very important to get that feeling of what do people get from it, for example through surveys? Yeah. Yeah. So we psychologist would do actual direct face-to-face interviews on people. This is the face to evaluations within the laboratory we get patients in getting to talk through a system should my tracking as well. See how they using it using it and then valuating on these things. The metrics off quite difficult to do. The satisfaction with which the decision has been made is quite a quite a tricky thing to you need to try to be able to do it. So, you know, these rather loose things. It is worth the effort of trying to measure them as accurately as possible I used as not an idea system. We've we put a front end on predict which is woman newly diagnosed with breast cancer who turned like what other therapies to have apart from surgery which space. Fatty basic statistical analysis of data full-size in case, isn't it produces survival because fifteen years for women, and then looks at will be the fact these are personalized to various attributes of the tumor women, and then says that survivor we change if you take particular therapies and those defect to the therapies will that teachers based on two nickel trial cools data from randomized clinical trials, and that's fine. Our ideas that the system which is currently used by doctors, and it will be used by doctors talking to patients or even by patients themselves and support groups, but we using exactly the same system rule these different groups, and that means very good explanation facilities, both the terms, but it will so ways portraying the risks to patients, and this is serious stuff. This is the chance people are gonna be alive in ten years. But with very careful use of Wooding, an imagery and even the color and exciting the we can. And the point about this is that for explanation that the two things one size does not fit all different people have different needs different levels of understanding about numbers and graphics. And so what you need is both multi laid explanation. The very simple level of the top three too much deepen a maths in. If you want to you can see a PDF with all the Matson, you can get the code if you really want to. So you got all his lead vaccination vertically, but also horizontally, so when we're explaining fifteen years -vivor we can provide botch arts and survival curves. And I wanna raise tables, tax, etc. All of those options, what people prefer to see. So you've got vertical and horizontal explanation. Choices is no correct way to do it. But you can try to evaluate all of it only. Some people wanna see the stuff at the bottom, but they should be able to see it because that's poverty, assessable openness, that example, is a compelling one. I find that oftentimes dealing. With with physicians. There's this there's a presumption of trust or trustworthiness that may work for a lot of people. But sometimes you want a little bit of data, and they're not always prepared to train people making much melt everybody, you know, I just completely off top of my head. I'd say about half people quite propelled still go along with the very paternalist point of view. Thank you very much. Tell me what to do just tell me what to do want to know anything else, but an increasing proportion. Aw, asking questions and actually wanting to size of their rights. I've got friends who have used system that we've been working on in order to challenge that doctors saying, okay. Tiny benefit. I know I'm gonna get terrible side effects. I'm not going to have yet. And they've used to chums. It's in power them empowered them. I think very valuable not only that. But in the UK now, there's a much stronger legal structure on what must be explained to people in order to give informed consent for treatment. And they know those oldest should be explained to people we're providing is actually some of the tools to doctors to to carry out that work. That's okay. There's a thread in the community around taking ideas from adjacent fields like electrical engineering, the idea of datasheets or model cards. Some folks have called them in and basically different ways, documenting the characteristics are biases of different data sets or systems. And it sounds like a part of what you're doing is a similar idea. But applying ideas from. Clinical trials and the statistical methods associated with clinical trials, and and medical and pharmaceutical field to the way, we talk about and communicated around a systems and machine learning models and not just this refund. Larry being VO for years and building stick systems the people and then both evaluating them. And and putting them into practice. I'm one of the crucial things about the evaluation that people get really obsessed back to this. Pedantic way that's not decisions tend to operate. It is that the probabilities must be meaningful. If you say seventy percent probability for something or seventy one hundred John's gonna mean that meaning that some are at the number of all the times, you say that it should happen in seventy percent of the time the under calibrated probabilities in other words, the uncertainty. The accuracy of the county is important as the accuracy of the number right now is very statistical idea, and very I think it's very important because otherwise you get these grossly overconfident things. I'm ninety nine percent. Sure. That this is the diagnosis grossness leading really is terrible. So. That's a I think a very another very pool way thing that can be brought from statistics, which is two lots on how to evaluate the calibration probabilities Testa ticks to us and so on and check that element of trust within the kind right, right? This all Kalsa mind, at least in the US. I don't know if it's more in the UK when you're adverse advertising pharmaceuticals. There's like you have your thirty minute ad your long, Brad. They shouldn't have to do. I mean, that's just regulation. What you what should we know? But at the same time, you know, that's kind of a summary of data. She. That's that's not trustworthy communication like having to sign getting some software and those terms and conditions sixteen pages of terms of conditioning. Yes, that is not intelligent openness in no way. Is that excessive usable comprehensible? Comprehensible bright breaks every rule and it's terrible. That so the communication. A little, but is a complete sham in terms of good communication. I agree. I agree at the same time. It is wants to better than you know. Yes. Or the computers. Exactly. The time is the worst systems all proprietary systems that are used in courts to decide about recidivism risk Obama. Right. If shocking proprietary that putting done transparent, you know, I d what he is being used in them. I mean, that's absolutely shocking again break every rule, you know, everything I'm trying to I'm talking about his broken by that kind of system. So key takeaways from your talk. Oh, yeah. By six testicle ideas experience in other as a lot to offer a luck. But also again, not just taking from you know, from statistics taking ideas from philosophy and psychology caracul testing that that really in this maturing disciplines unbelievably important discipline. I think could take a lot more a kind of very much in line with some of the key themes that I'm hearing this year's nerves. You know, it's in fact, two of them one is the importance of fairness transparency etcetera, and the other is kind of the importance of interdisciplinary approaches year kind of bringing. Because this wonderful what going on this one featured the fat L social impact statements this. They've got poly because they do not identify transparency as an objective. They've learned. So that transparency just a means to a name. It's no good just being transparent less. You buy a Nora Neal's ideas, what transparency mean? Well, we'll definitely provide a pointer to nor are Neal and your work as well. Of course, the tool up on Facebook as well. Fantastic fantastic. Well, David thank you so much for taking the time. Thank you very much roskin. All right, everyone. That's our show for today. For more information on David or any of the topics covered in this episode. Visit twiddle AI dot com slash talk slash to twelve. You can also follow along with our neural series at AI dot com slash neural twenty eighteen as always thanks so much for listening and catching next time.
"david spiegelhalter" Discussed on Talking Machines
"And you know, I if I look at the scale of the program so skimming through the posters. Yeah. I mean, this just directly in my area, there's mole material than I could possibly assimilate. And actually, I am concerned about this the so what factor around a lot of. Tuck nowadays because this so often the use of see reviewing dozens of a great job on the standing whether something's really important problem will not because you just don't necessarily have that experience in your reviewing body particularly in the growing field. Because a lot of your reviewing body is only just starting to understand the context. I do struggle a little bit nowadays when I'm at conferences to get into things because sometimes you hear something anything. Okay. That sounds like it's this combined with that which is fine. But you know, you're not like, I'm not sure why you should cat machine learning is at its best. When people across fertilizing, amazing ideas to do very cool things. But in some sense, the best very best ideas the ones that you can explain quickly. But no one ever thought of before they often very harshly reviewed because if you've explained them quickly people don't think, oh, I never thought of that the full. That's amazing insight. I think oh that must have been simple. Right. That's obvious. Now. Now, you've explained it. So clearly explained it now I understand how could that be forward? Yeah. Absolutely. Yeah. And I think. So so I personally struggled a bit. I don't know how the people manage. I think people put a lot of time into what they going to look at. And and for that reason. I mean, the program is always interesting in terms of what's and a lot lot more. As we go on is relying on the suit of the vita talks, the Brian lectures, very I always enjoy listening to David Spiegelhalter speak. He's so great on public understanding of risk. He's based she's based here in Cambridge. But he has all these mechanisms because of a David Spiegelhalter, I jumped out of an aeroplane. Well, if you are looking for a justification to fling yourself out of an airplane or or other. Talk. I I am super excited about the talk one of the invited talks the investigations into the human AI trust phenomenon. I think that that will be really cool. I think that's professor Howard. And then also this is the second year for the competition track. Which was a lot of fun to sort of watch unfold last year, and I'm really excited to see how they have changed it or we the winters. I went sort of questions are up for this year. I'm pretty sure that the questions are already up and probably have been for a while. But it's always exciting to sort of like see it unfold live, and there's a number of like one thing I noticed a little bit because I've just been invited to a couple of things there's a number of fringe events to the conference. That's exciting. I don't know all to them by being smaller workshops where people are just getting together. Isn't that something like the g seven's there? Something always already there. The president of the g seven isn't. Will be counted. As a fringe event has. They Canadian government is hold anything. So that's really interesting yet. The so candidates host seven conference on artificial intelligence on December the sixth year right at the same time as nips. Yeah..
"david spiegelhalter" Discussed on KGO 810
"A moderate drinker you're would be far better served by buying a fire extinguisher, a bike. Helmet then, quitting the sauce, in an extended critique of the, Lancet study British statistician David Spiegelhalter calculated that on average it would take. An incredible four hundred thousand bottles of, gin to prompt a single extra health problem among moderate drinkers with these kind. Of numbers Spiegelhalter was particularly critical. Of the study's conclusion that public health agencies should quote, consider recommendations for abstention Roach Spiegelhalter quote there's. No safe level. Of driving but governments do not recommend that people. Driving come to think of it there is no safe level of. Living but nobody would recommend abstention manners a lot, of crap science, in the. World I try, not to be cynical I'll. Try to be skeptical not cynical It gets harder, and harder Coming up in a moment, the science between hot fees New, research is revealed while here. I, am talking about crap science hit. You with this yes I am, capable of appreciating irony it's the quandaries of modern life but I am assured by certain members of the Armstrong and Getty team that, this really is revealing in Sting the, science beat the hind y ladies like to pose in front of their phones and get hot hot photos On, a somewhat related topic oh we don't really have time do we got like two minutes Michael Very, quickly We talked about women in the workplace in Queen, bees yesterday gals who bully other gal's in the workplace and how they can make life miserable Here's, you know I'm going to keep everybody anonymous because they didn't Mark off. And I just didn't want us I don't wanna get anybody in trouble but I'm a fifty two year old female for. The last. Five years been able blessed work from home before that can honestly say it's been a fifty two year. Old female for, five years shut up I had one job got along with the three other women, who worked in the same room as myself other than that not a women can be. The most horrible creatures to work with drama drama drama gimme, a room full of men to work with any day over one female co worker I love working in a male environment rights another gal. Women are to, Patty that said the reason why average women don't like working with quote pocket popular women said they're generally good. Looking, men treat, them entirely differently men lose their brains and think with other body parts am I being petty. Or truthful now you're being truthful there's, absolutely true to that Queen bees and nursing a male nurse working in. A busy intercity hospital easily five one female to male work environment I'm surrounded by Queen bees can confirm they are way. More ruthless. On each other than the other than the males additionally there often more vulgar and sexual than men I. Work with the, cetera backstabbing gossip teaming up on each other click against click click against individual clicky, means all while smiling through their ever present Rb f- resting bitch face shivers in terror I'm a woman. Who worked in a large office with many women in hated it. So much gossip, backstabbing at center I.
"david spiegelhalter" Discussed on KDWN 720AM
"To ten pm but research by lancaster university so just people are watching on demand services later into the evening the study was carried out to examine the impact of internet traffic on electricite consumption as demand on the national grid increases they say evidence from nearly four hundred devices show peak hour from ten pm that's a small study only four hundred devices but i have to tell you it is something that i think a lot of us have noticed before they even came up with this they say the author state that to the extent that this traffic is associated with viewing films the programs is it just a mobile devices are used to prolong hours of tv watching and maybe the main tv set is turned off so they say this there was a study back in two thousand sixteen where the author david spiegelhalter from the university of cambridge sin couples are less interested in sex because they watch more tv in bed he blamed figures showing falling rates of sex on the mass connectivity we have compared with just a few years ago but the tv used to close down at half past ten i don't remember my parents staying up to watch tv except for the late show the late show was on late i mean i remember friday nights in the eighties i remember watching the dukes of hazzard and then i think dallas i forget what the friday but i remember that was like friday night lineup and then unless they switch to a different day macgyver i think was on wednesdays done done done trying to think moonlighting was tuesday nights thursday night the mustsee tv did it happen till the nineties that's when we started to get seinfeld and a lot of good shows i think cheers so thursday night started to be more of the sitcom i used to watch the huxtables the bill cosby show i forget what day that was that might have been a thursday i forget what most of the shows were seven eight o'clock seven o'clock is fivetime then people now skipped seven o'clock eight o'clock so eight to ten and then i think the night show the late show was either eight thirty or i'm sorry ten thirty or eleven thirty i forget but in the olden days you didn't have tv so sex would be your entertainment that's why people have bigger families kids will go to bed parents would go and have sex and then fall asleep then they'll get a good night's sleep you didn't need ambien sex paul sleep.
"david spiegelhalter" Discussed on TalkRadio 630 KHOW
"Of tv watching perhaps after the main tv set has been turned off you know my guess is if people are watching their tablets in imbed they probably may not even own a traditional television the study is published in the journal of energy research and social science now who in the hell came up with combining energy research and social science into a gerbil and someone's paying probably a thousand dollars a year for that quarterly journal i get it for free because while i'm cool and i love energy research and social science the research could support a warning from professor david spiegelhalter from the university of cambridge who says couples are less interested in sex because they watch more tv in bed i'm not sure i think the real answer about why people aren't having as much sex as they used to is explained in this philosophical clinton i'll tell you i can't relax our place i felt like a few drinks at a bartender says surprise me showed me a naked picture my wife the truth and my wife would i would never have to address because stop laughing that's actually not very funny because i kind of feel that way like when i get mecca i like turn the lights off when i tell you what my wife does have sex scream especially when i walk in on them i got no sex life ten years ago my wife put me on hold the last one made love to my wife it was ridiculous nothing was happening i looked can't you think of anyone either i know my wife cheats on every time i come home the perez quick out the window such life is nothing my wife caught me out to once a month coming out at once a month oh i'm lucky to she cut are completely you have sex with my wife is like magic soon as i get in bed she disappears the front door wearing a see through negligee trouble is she was coming home smoking that's another one trying to stop smoking that's a beauty cigarettes my wife flow we made a deal my wife and i we only spoke after six say packbell since nineteen seventy five what bothers me is my wife she's up to three packs a day i don't know i find it funny will you need to get a life we've stopped it to quit smoking so my wife's up to three packs a day wall street journal has a story about shopping for surgery in colorado and i love it here's a.
"david spiegelhalter" Discussed on WNYC 93.9 FM
"The headline not many david spiegelhalter is the professor for the public understanding of risk at cambridge university's statistical lebar tra bich tricky and is challenging and particularly as ludlow skill is working at what one not being told by the dog that didn't bark in the night and that source who've detective work does require some practice when we looked into these ships verse his car's statistic it turned out that what we weren't being told was well quite a lot the calculation in two thousand nine was more of a thought experiment and if the ships in the calculation used the fuel that was the worst quality ship fuel allowed in the regulations and if the world's cars used the cleanest gasoline that was required anywhere in the world then it was true that was dumped to james cole bits of the university of delaware the original source of the claim and it turns out that the headline worlds fifteen largest chip submit as much pollution as all the cars in the world should really have read sixteen of these massive ships the ones that don't get exist would admit as much sulphurous all causing the world if they use the dirtiest possible fuel which they went and all the causal using the cleanest petrol available which they armed fix that for you and no it's not as catchy number three get the backstory of all the statistical claims in the world face particular stat for towels showed up in front of you dressed to impress why are you saying it where did it come from who wants you to hear it and what do they want you to do about it he's david spiegelhalter grants remarks with this wonderful statement that he would never join any club that would accept him as a member said the slight paradoxical statement that defalut because the club comunity good of it would have him and i've got that slightly cynical view myself about stories are here in the news weekly health stories because the only reason there in the news and i'm hearing this because they're extremely different they go against accepted wisdom and so when they newsworthy which for me almost certainly means that they're wrong and so i've got what i call the grants or principled the very fact with i'm hearing story these reason to not listen to it that's not to say every surprising claim as.
"david spiegelhalter" Discussed on KQED Radio
"A bill that he of any event on the day i sent him to record himself after ten or fifteen minutes of not seeing a single other car on the road the came around the spend there's of manure wagon being driven by a tractor right down the road and yes we the air at all happens the manure wagon is going down the road tractor has just pulled out on the road and now i'm passing by and all three happen that the exact same place it just happens what is your question then is it why does that happen or is it a question is why does this happen to me and does it seem to happen to others as well a k these are these quite complicated then if you're just your father get the thing but you get to the bottom of this week cult of a man who he thought might just have the answer i'm professor must be professor david spiegelhalter we did you say you're professor surrey does that mean you've been knighted yes yes you you're a night you're the first night i've ever spoken to i'm already knows nothing but caved stuffed full of this hundreds is three a penny the probability of meat talking to a guy yeah exactly raise the very high data professor sir david his a statistician were pretty dull within when i ran my old man's questioned by in his response i get caught little stories like this people contact me and they get very anxious and fat because they feel it somebody's going on in their lives that making things happen all the time they getting signs from the environment to rang them okay near term is it's noticed synchronicity synchronous you knew all these big gouda faculties of chike.
"david spiegelhalter" Discussed on RadioLab
"Creating a dangerous situation first of all but then it's just this fire and then it defuses and you could sit on that stretch of road for six more years and not find that happened in the key thing here my dad says is c tim adler witnesses this sort of thing always athens to me a lot all the time i would say it happens once every two weeks we have the possibility of an event on the day i sent him to record himself after ten or fifteen minutes of not seeing a single other car on the road the came around the spend there's of manure wagon being driven by a tractor right down the road and yes we ear it all happens the manure wagon is going down the road at tractor has just told out on the road and now i passing by and all three happened at the exact same place it just happens what is your question then is it why does that happen or is it a question is why does this happen to me and does it seemed to happen to others as well a these are these quite complicate peripheral tissue father get the thing but to get to the bottom of this we called who man who he thought might just have the answer i'm professor must be professor david spiegelhalter we did you say you're professor sir does that mean you've been knighted yes here here you're a night.
"david spiegelhalter" Discussed on The Science Show
"The next time is bound to be tales haha no government that's the kind of ask the question that will would you suspect that this is a boss coin or is seventy eight heads out of one hundred in possible well digit felling cambridge code david spiegelhalter whose written wonderful books on the statute everyday life and they really fascinating ones i rather like his his many morph cnn wayne if you go to china issue of the obama legacy mongering micro more flesh right you the battle would have lost half an hour of life by going to beijing and breathing in the air and getting on a motorbike as equivalent i'm here to talk to you about how to rephrase the arctic which hugh hunter from trinity college in cambridge will talk about next month on the sancho nahrin meddling with maths it's tricky trickier separatation as we know every day is to the weather he has said he collaborator with chaos imagine that you're looking at a pendulum as simple rod attached to a friction less pivot allowed to swing back and forth you lift the tip of the pendulum and let it go as you watch it swing he know that you will always be able to predict the behavior of the pendulum this is because we can write down a straightforward equation that captures the physics of the pendulum system and we can solve it for any time in the future we know exactly where the pendulum is going to be.
"david spiegelhalter" Discussed on WNYC 93.9 FM
"And writing a key skills that we would ensure that children have and we wouldn't tell him that they not have leading personal writing pierce and we will work on that and i think the same has to be traced numeracy there are certain basic numeracy skills that are important for everybody and it's not like hey just say not an and this station how did numbers may fail we get anxious did you have a bad time at school numbers i did never get it now people should acknowledge more than two petitions on worse but to say that this is an intrinsic characteristic of may an old world in with us just the way i am is i think about playing well on poetry unfit while elsewhere were no actually you could do something about it you could make a bit of effort on everyone's going to become amassed professor who wants to be david spiegelhalter is a math professor modern that he's a professor for the public understanding of risk at cambridge university and he does not stand for people letting the numbers push them around fightback show the figures who's boss you can do something about to an actually will improve i think the way in which you deal with the world it enables you should be able to kortrijk quadrupling told otherwise you're just open to manipulation by people people find it difficult to critique them to ask questions of this judge is a big number or small the major culprit in this all people who are using numbers who are telling us numbers because very often they do so in a way that is either unhelpful or even deliberately coercive when someone gives you a number you accepted almost automatically as something that is true you don't question its authenticity unless you know they're very well sixty percent of americans earth functionally enumerate if you believe that number this is an example of what i'm saying i just me that i was gonna say where does that come from one of the few because a number has a specificity early news unchallengeable no one can question whether to plus choose for really challenging a number that's given to goes against the grain.