How to make the world add up

How to make the world add up

Show Video

[Music] hello and welcome to this event for the cambridge festival my name is david spiegelholter i'm the chair of the winton center for risk and evidence communication in the math department at the university of cambridge and it's my enormous pleasure to introduce you to tim harford and for the next uh 45 minutes or so we're going to be talking to talking together particularly about his new book okay so tim harford i'm sure you all know him um he's an economist an author best-selling author and a journalist for the financial times and because i think you know the way that many of us will know him is as the presenter of this wonderful radio program more or less who have been extremely busy over the pandemic so um i'm going to introduce you his is tim it's great to see him there hello hello tim and we should of course be sitting you know next to each other on a stage or whatever in front of everybody well we can't do that so we try to pretend that we pretend that we are okay so tim there's lots of things you know we could talk about but i'd like to focus on this book um which has got a is i want to back yes yeah we've all got yeah it's great from all good book sellers and it is a good book i've even read i've even said so on the on the back cover um and it's very interesting i you know i should say i'm a statistician i'm interested in statistics and numbers i'm interested in communicating them um and uh so first of all i'm just going to ask you why why did you write this book well for a long time i didn't want to write the book i i've been presenting more or less for 13 14 years now and there are a lot of good books about statistics out there i mean it's a wonderful book by some um spiegelhalter fellow called the artist statistics um you know whatever happened to him um there there's i mean terrific books by andrew dillnot and michael blastland who who first created more or less there are various classics in the field and i just felt well i'm not sure i've got anything new to say and then i over time i think the experience of fact checkings certain rather contentious elections and the brexit referendum made me realize maybe i do have something to say there were a couple of points i wanted to make the first was that it's very easy to get trapped into this um pitfall of constantly finding fault of endlessly fact-checking calling out lies calling out misleading information and losing track of the fact that statistics can be used to perceive the truth as well as to to spread misinformation that was important and the other thing that was important was the realization i think has become quite obvious in the last few years that a lot of what we think isn't to do with the facts uh it's not to do with the data it's to do with our preconceptions it's what we want to think or what we're afraid might be true or what our friends think and that's i think that's fine i mean that's just being human but i when i was writing this book about statistics i thought if i can address those two issues not to be too cynical and to think about uh human fallibility and filters and biases if i can do those two things then um maybe i can contribute something that others haven't focused on yeah and i mean that's what i find so interesting is that um you know it doesn't cover the statistical ideas you know that happens so often in criticisms mean media the mode and all those sort of concepts that we're used to using when we're when we as you said we are taking apart uh the the claims made on the basis of statistics and instead your chapters are you know you've got titles like ponder your personal experience and search your feelings and things like that it's it's very much a sort of human psychological view of our relationship with numbers yes by search your feelings it sounds like those star wars movies meets uh statistical thinking uh i mean actually i begin that chapter that's the first chapter in the book uh i begin that chapter with a story of uh of a venerable art critic called abraham bradius being presented with a piece of art in the 1930s and asked his opinion and he's he's i mean he's brilliantly um revered people regard him as as the the the expert on rembrandt and vermeer these great dutch masters and he opens this unrolls this painting he looks at this painting and and contemplates it and and he begins to shake with emotion he's just overcome with joy at this thing this discovery of a masterpiece by vermeer of course it's not a masterpiece by vermeer it's a forgery and it's not even a good forgery it's a rotten forgery and yet this expert falls for it and he falls for it because he's he gets carried away by his own theories and preconceptions and wishful thinking and the reason i wanted to begin the whole book with a story that actually isn't about numbers at all was because yeah our emotions are incredibly important and when we first see something whether it's a forgery of a piece purporting to be by vermeer or whether it's a claim in the headlines of the daily mail or the bbc or on social media when we see something our first response is often emotional and that's true even if we are experts even i'm sure david you you would fess up to this that you just instinctively you say oh that doesn't that can't be right i don't believe it or yeah that sounds about right and it's a gut feeling at first and we need to to notice that gut feeling and get past it if we're going to actually start thinking clearly but also i sometimes think those gut feelings can be rather useful i always say you know you can sniff a number and uh you know after a lot of years of doing this you can sort of get a feeling of what you know why this is a slightly smelly number or not or whether it it feels quite fresh and i really there's the idea of forgeries and everything i'm reminded of the story of hugh trevor roper and the hitler diaries when he you know famous historian who validated the these hit the diaries which were yes simple forgeries and uh ran in the sunday times as a great scoop and all this sort of stuff so because wouldn't it be wonderful if they really were but yeah yeah wouldn't it be amazing if it was true thinking you identified as beautiful you know we call it motivating reasoning whatever we want to call it and how of course it influences our our feelings i mean actually i must say a lot just like that story it really isn't about statistics the the vermeer story and a lot of what you say in the book you know it's not just it does certainly does not apply to just numbers i mean it's about critiquing claims that anyone could make about anything yes i was i was very struck years ago uh more or less was up for a prize from the royal statistical society of which you will know david but other people may not know that you used to be president of the royal statistical society and at this um this prize giving evan davis gave a short speech and he said something that really stuck with me he said that basically good statistical journalism is basically just good journalism you need to give people the context you need to tell them why a particular claim is important you know why should they believe it or or why should they be skeptical of it who says so what does it what does it mean how does it fit into the bigger picture um and these are certainly claims that a good statistician can help to answer using technical tools they're also questions that you can answer using common sense or by you know doing a bit more googling or asking around yeah uh so yeah it's it's it's the the technical details matter but for most people they're not the only thing that matters well for everybody they're not the only thing that matters for most people they're not the main thing that matters you can ask really smart questions and understand the world a lot better by asking fairly simple questions i i completely agree i've you know we both know michael blastland who is one of the originators of of more or less and as a journalist he's an english literature graduate and yet i i respect his ability to to tear apart a statistical argument as much as anybody else's and uh and i think this whole covered pandemic has also shown that you know i've done a lot of the standard taking numbers apart and i've not been using sophisticated statistical methods at all really for anything it's just actually basic ideas of can you believe the claims we know what you know what what are we not being told and all those sorts of things that you you show so well in your book yeah absolutely i mean it's it's been very interesting to see um the way that statistics were used really throughout 2020 um to try to make sense of of the pandemic uh and at first well there's this catastrophe in folding the world but from the point of view of a geek it was kind of refreshing uh having having dealt with you know trump elections and brexit referendums and and arguments where statistics were just it's just ammunition they're just ways for people to win arguments and no one really seemed to care everyone had their own view and there's always a reason to dismiss whatever the other side was saying suddenly we're in the situation where people really wanted to know the truth and they realized that there's a lot of things you could only understand about the virus by using statistics so how dangerous is it who's most at risk how many cases are there how does it spread what are the effective treatments uh how do these vaccines work are they safe i mean these are all statistical questions and um of course it quickly became polarized and people would start having the same stupid arguments based on cherry-picked data or no data at all but there was this moment where people really started to realize how important statistics were and one of the points that i've been trying to make to people is you think about the scramble that statisticians have made to gather all all of this data to try to figure out what's going on um actually that happens all the time it may not be you know it's frenetic it may not feel like an emergency but statisticians and people working for national statistical agencies and other places they're always trying to figure out what's happening what's happening to the planet what's happening to the economy uh what's happening demographically they're trying to figure this out and we just sit here and talk about lies damn lies and statistics we take it all for granted we just assume the numbers will always be there yeah and the early 2020 taught us what happens when the numbers aren't there we can't take them for granted i love you you've got your chapter called don't take bedrock for granted which is about the importance of official statistics and it's lovely to see that i find that very moving because normally you know this enormous amount of work that's going on the whole time grossly undervalued nobody takes any notice of it at all um and as you said this this crisis has brought this enormous attention to it the office for national statistics has been releasing its weekly provisional death registrations you know for decades if not centuries and nobody's ever taken the blindest bit of notice of it except now and tuesday morning at 9 30 everyone's waiting for it to pop out and people are phoning out asking for media interviews and things from the from this document and um i you know to start a decision of course this is this is this is great and and as you said it's brought about you know interest in claims based on statistics um i i i've got to ask you you know what's it been like doing more or less throughout this pandemic uh it's it's been it's been quite intense uh the and of course a lot of people have have had to deal with all kinds of uh trauma loss of job loss of livelihood loneliness uh illness uh bereavement so you know i'm not the only person who's who's had to deal with this but um a a dear a friend and mentor of mine peter sinclair the man who taught me economics the man who persuaded me to become an economist he died of covid in march march 2020 so very early on and so right from the start i had this and i could look at the data i could look at the statistics but i had this personal experience of of what this thing could do and and that was always weighing on my mind uh and then we had this strange situation of like many people i was working from home i was under a duvet and ordered a bunch of acoustic tiles from the internet and the bbc sent me a better microphone and we're trying to figure out how to how to do this how to coordinate all of this activity remotely my producer was was in broadcasting house and the studio engineer was in broadcasting house but most of the rest of the team including me we were all working from home um simultaneously the bbc had by pure coincidence moved more or less to a better slot and you know 9am so it was getting more listeners anyway and then we started chasing down various stories um the way the government was reporting hitting its own testing targets for example um there was a lot of attention something i would i would i mean you've had the same experience david i would i would tweet some stuff and then suddenly it'd be national news tim harford tim harford said somebody just sent him a copy test through the post and it was rather difficult to do you know the guardian would be it was it was weird this sort of weird kind of you know b grade c grade uh celebrity but most importantly i think it was it was wonderful to be working with really dedicated people smart people trying to do something that felt important and and you felt that you were making a contribution and i know a lot of people for the last year have felt that they they haven't been able to to make a contribution they haven't been able to do the work they wanted to do because the pandemics got in the way so i'm very grateful for that yeah i mean i suppose i've had a slightly similar experience and you know your status who cares about statisticians and yet i found myself that just a sort of rather casual utterance to a journalist becomes the story and which is not what the situation i want to be in in any way at all you were quoted in parliament by both sides i know a real you know at this point at that point i thought i got time to keep the head down this is not the sort of role role i want but i i would like to explore that as well a bit um about the role of statistics and more or less in in what has become a polarized debate um i i'm constantly uh you know being asked you know which side am i on almost you know what policies do i support and if i say something criticizing one side it's interpreted as supporting the other side or something and i have to spend my whole time saying i'm not pro lock down i'm not anti-lock down i'm just interested in you know you know trying to work out what's going on and more or less must be have you found yourself accused of being on one side or another oh yeah i mean we're always we're always accused of being on one side or another we're always accused of being biased against whatever the person who emails us believes and you know and of course we try to be uh we try to be fact-based it's going to say balanced or neutral or but we try try to be fair and we try to stick to the facts um of course that's an impossible dream but that's what we do i mean i'm in the slightly strange position of also having a column in the financial times and you know i actually work for the financial times not for the bbc and in this column in the ft kind of expected to say what i think about certain things it was a bit tricky to be a newspaper economist and never express an opinion on everything so so i do have an outlet if i want if i want to um to be opinionated and to say that the you know the government's doing brilliantly all the government's messing things up uh so that's kind of nice but i have to say that the that sort of slightly detached view that the bbc tries to take of sticking to the facts and being neutral it's kind of addictive once you get into that habit it's quite hard to start expressing uh strong views and and i i quite i quite like it um it is it is interesting and i think a little worrying but maybe not surprising that uh people trying to make particular arguments so to say well we need to go for zero coverage you know we need to be much more aggressive in the lockdowns we need to get down to you know one case in a million or something like that which is something that at one point independent sage was saying or only the lockdown skeptics it's just it's like flu or the infection fatality rates we know one in a thousand or the majority of cases are asymptomatic um people it's worrying when i see people let the the the what position they want to reach drive what facts they look at i mean sometimes they are facts just highly selectively reported facts sometimes they're not facts at all um but i suppose you shouldn't be surprised that's what that's kind of how argument works how political argument works it's a shame but that's the way we are well i mean it's what you highlight in your your book all the time about you know the selectivity of evidence that if we're seeing a claim about thinking about what we're being told thinking what the back story is is not just the headlines but i i think you're right that in this over the covert it has been interesting just how many of the polarized claims and what i believe is at some extremes just misinformation has been statistics based now no there's always a number that can be brought out to make any argument i say so i'm going to push you a bit okay what do you think what's your favorite bit of statistical misinformation that you think has been bandied around and not maybe not just about covid but recently what's one of your crack i mean i've spread a little bit of statistical misinformation myself accidentally i'm happy to talk about that but um the um yeah it does happen it does happen i think what one thing that was was quite striking and i saw you engaging in the discussion on twitter um i think it was possibly back in november or or october uh so toby young who's a sort of a lockdown skeptic you know he thinks that lockdown's a bad idea and i should say i think there's a very principled argument against lockdowns you know there is an argument you can make you just have to make it using you know the right numbers um and he he basically said look uh i've done the maths and the infection fatality rate is is 0.1 so effectively you need a thousand people to get coronavirus for one person to die and various people pointed out that he just got it wrong he was out by a factor of 10. the infection fatality rate was in fact one percent so if you have in fact a thousand people you can expect 10 to die which is sort of in the ballpark of what um people had been estimating for a very long time for at least four uh countries you know wealthy countries with with quite a lot of older people in developing countries with a different demographic structure it is it is lower um but what struck me so he tweeted this and he said oh i got it wrong he and he he withdrew the tweet but his confidence in his position doesn't seem to as far as i can work out didn't seem to be shaken by by an error of a factor of 10.

you would think that when somebody looked at the days oh my goodness i made a terrible mistake and it it completely changes my conclusions but instead he said oh my goodness i made a terrible mistake i fixed the mistake i have the same conclusions i guess well he's not the only one we're all like that one way or another but i try not to be that guy but i mean i does i think that's a fantastic example of where a number was being used and he was trying to use it as ammunition for his argument and when when it was pointed out to him that it not only was wrong but it actually completely destroyed the argument he was making as it didn't affect him at all because that number was only being used as a selective weapon and if it didn't work or you throw it away and you find something else that will work yeah yeah it it yeah you're right it really was a a wonderful example of this um you know weaponizing of of numbers in a casual way and not actually caring too much whether they're right or wrong yeah i mean some of this stuff is genuinely hard so i've seen um people i respect say for example oh although most cases are asymptomatic um so the idea is that you know this is quite a mild disease and we're probably uh probably overestimating how dangerous it is because we know that actually most people don't even know they have it um that's just based on an error people even link to various media reports and if you read the media reports carefully you realize that it's based on an error and the error is the definition of asymptomatic uh most people who test positive don't have symptoms on the day they test positive but usually because they're post-symptomatic or they're pre-symptomatic most people aren't asymptomatic maybe 20 we've got quite good studies of this but it's a persistent error uh that people make and it picks up a point i try to make in the book that often the the mistake is not the number the number's correct the the error is the label that you've attached to the number you don't unders you know you've used this word asymptomatic but actually you don't really know what it means until you've left to the wrong conclusion happens a lot yeah i mean i saw that just a recent report i mean from said oh 80 percent ray symptomatic but you know it actually meant at the time of the test that included pre-symptomatic and post-synthetic so yeah it was hopeless so that this is your lovely chapter on premature enumeration you know yeah i suffer from it sometimes just um yeah exactly and you know rather than leaping to analyzing the numbers actually trying to work out well what do they actually mean and amazing throughout covet so much of the the wrong arguments that have been to do with just lack of definition of of what it meant i mean and one of the you know ones that uh karl hennigan and his colleagues spotted was the fact that for some time in the in england anyway covert deaths were being counted as anyone who'd ever had a positive test and then died of anything at any time afterwards this is hopeless obviously so what okay what about some other examples of where you know in more or less or elsewhere where you've checked you know just by merely finding out what something actually meant what people were really counting um has it sort of just you've just been so surprised and shocked at that it's just the the misleading definition i suppose yeah yeah well one that really uh struck me um it's a very sad example and and i think it's it's quite subtle it's not nobody made any any gross error but we were looking at infant mortality and the the um a couple years ago the infant mortality rates in the uk started to possibly in england and wales started to rise people were worried about this and trying to figure out why this was and it turns out that um it looks as though it was actually just a change in the definition of what it means to be an infant when you think about it you realize what the the definition between a fee the distinction between a fetus and a baby is one of them you know the most contentious in uh you know moral philosophy is one of the most contentious in uh is certainly in american politics and what was happening was there's a small number of women thankfully this doesn't happen very often but a small number of women who were delivering very premature infants and then the question was was that a late miscarriage or was that a premature birth and the and and then the baby died uh and very sad but there's a there's a gray area as to what you call that and and in recent years i think uh medical professionals have shifted towards saying you had a baby the baby died because i think parents feel that that reflects their grief more um but fundamentally the same thing has happened but when you do that the infant mortality rate goes up and that's that's a a really important case where you can you can run around trying to figure out what is happening what a doctor is doing wrong what a nurse is doing wrong uh you know is it a public health thing is it a nhs funding thing and you realize actually it's probably not happening it's a statistical uh artifact because people have changed the definition for i think understandable reasons yeah yeah i i think you know a lot of stories are more or less reflect the fact that if people just realized what was being counted it might not represent what they what they thought it was um i think you did something on violent crime where you know it shows that a substantial proportion of so-called violent crime doesn't involve any physical contact between people or people at all and and if you think of covert deaths which i can think of at least four definitions for covert debts i mean there's still a huge argument going on which i get you know emails every day about this fact oh these are just deaths with covid they would have died anyway it's not a code of death there's a whole business of code deaths of covet or with kobe still going on and on yeah it's a hard one to resolve because i mean of course we will all die anyway in the end and a lot of the people who die of covid are very old um so it's not an absurd point to make but i think my understanding is when you start looking at the death certificates um most of the people who who die following a positive coveted test most of them you know covid was a a leading cause in the opinion of the of the doctors yeah i mean i think it's quite clear 90 of what if covid's on the death certificate in other words it involves covert in ninety percent of the time it's the it's the main underlying immediate cause of death but there's always pretty pretty well always other things listed in other words people who die of covet have got other chronic conditions that they would have died of possibly eventually but as you said every death is a death brought forward and the crucial thing is by how much and that i mean that's and that's what obviously been a lot of the discussion you know are these people who were who were at death's door and they were going to die anyway and this happened that coveted happened to have been happened to do to um be their the immediate cause which is why it's top of the death certificate but i mean the data doesn't support that that argument has gone a little bit quieter now because if that had been the case for the 40 45 000 deaths in the first wave 50 000 in the first wave then we would have expected much more of a dip in in deaths after in the summer and that didn't really happen very well yeah i mean there was a there was a sort of testable prediction uh which was made by by some people they said i i think that you know we'll we'll have negative excess deaths effectively fewer deaths than expected there was some but not that as many as you might as you might think so essentially yeah every death is a death brought forward but actually it's not just a few months so these are these are quite subtle arguments but again they don't really involve you know mathematical statistical discussions or whatever they're all ones that can i think on the whole be brought down to um yeah as you should mention before almost common sense yeah well i mean a lot of it for me it comes down to curiosity actually because i i don't i don't think it is always a straightforward common sense question so with the infant mortality question i don't think it's common sense um that there is this very subtle distinction that it's changing over time um i think you do need at that point to to rely on an expert somebody who really understands the numbers um but the question for the ordinary person in the street is are you are you looking for those explanations are you curious are you open-minded are you asking around or are you just seizing on something because it it justified what you thought in the first place that you thought that the lockdowns were overrated people are panicking you thought that the you know the tories had slashed the nhs budget and that's why the tiny babies were dying you did you want to seize on the on the data or were you were you open to further discussion and curious about what was going on behind the numbers i think this is the most crucial point is that i i and i know that you if you told a number and a story about it usually very difficult to critique it if that's the only information you've got because you're never being told the whole business and the main thing is where did that number come from how was it collected and so on so yeah always it's not what you see in front of you it's not what is not the headline not even the story sometimes that will enable you to spot the problem it's by searching around and knowing more about the context and you talk about this a lot about you know looking for the back story um how i usually say is you know what you're uh the first question is what am i not being told almost the dog that didn't bark at the night what's the information in what i'm not hearing and then quite often to do with you know cherry-picking evidence and so on because you can't tell from the evidence you've got whether it's been cherry-picked or not i don't know yeah if you've got a you know when you're when you hear things where was your how did your mind work how do you try to how do you sniff something to work out you know what's going on here what am i what do you know what is what's going on behind this yeah i mean apart from you know email you and ask you what's going on um which of course most people don't quite have the privilege of being able to do that but they all do sure they have the privilege of sending the email but they i don't know if you have time to answer everybody but um if you do you're even more of a hero than i than i thought um i think the first thing that i do when i when i'm confronted with any number is that i do i think i have now got into the habit of taking my own advice and noticing my emotional reaction immediately am i am i on the defensive about this number do i feel challenged by this number or shocked or do i feel yay of course yes as i thought all along uh this is just more proof that i have sound judgments and see the world clearly um i just try to notice that and uh and sit with that for three seconds uh and then i get on with the rest of it uh but yes the the the the other questions the most important for me are if let's say this number is true let's just say it's true then what does it actually mean in terms of the definitions behind the number and second how important is it how significant is it not statistically significant but just practically significant if i hear that um the chance of the exchequer is running up a deficit of 300 billion pounds it doesn't really mean anything unless i know because i'm an economist and a nerd i do know that that happens to be a lot of money but for an ordinary person it doesn't mean anything 300 or 400 billion pounds means nothing until i can think of that in terms of what what does that compare with previous deficits or probably more uh useful what is that per person and actually a bit of mental math says with 60 70 million people 400 billion pounds is uh is an awful lot i mean it's nearly um nearly 10 000 pounds a person uh borrowed in a single year my maths are right so these are these are sort of the questions that i would ask to try to put the number into context and very often a little bit of mental maths a little bit of general knowledge uh or just a quick search on google and a quick you know a couple of numbers on a calculator and you're immediately in a position to start processing this claim and figuring out whether it's something you should be worried about or not yeah it is amazing how you know just basically is this a big number you know it does take you a long way in these things my favorite example about that one sorry is um the summer of 2020 matt hancock who's obviously looking for something to talk about other than how he was kind of rigging the testing statistics said oh um if everyone in the country who was overweight they all lost five pounds the nhs would save a hundred million pounds over five years and loads of people emailed me and they said how does he know this you know what studies were done what's the you know can he be confident of the causal connection just calm down hang on first of all before we do any of that there's nearly 70 million people in the country two-thirds of 100 million people in the country 100 million pounds that's one pound 50 each okay my my nine-year-old son can do that maths one pound fifty each and it's over five years so that's 30 pence each okay so who cares who cares if you want to lose weight because you it'll give you more energy or make you feel more comf confident great knock yourself out don't expect a a thank you letter from the head of the nhs because nhs is not going to notice well sending the letter would cost more than their charge than then saved absolutely so absolutely with the price of stamps these days you're right exactly now and you do finish off by this you know keeping being curious which i think is a lovely way to summary of what what one should do but you also got a chat about keeping an open mind you know essentially be willing to change your mind so i'm going to challenge you what have you had to change your mind about recently through um actually finding actually finding out that you were wrong yeah um so i think the the the two main things that i changed my mind about during the uh during the the whole pandemic first of all i i really hoped that the infection fatality rate would end up being something more like 0.1 percent one in a thousand than than one percent uh and we still don't know and i mean it's hard to define exactly what an infection fatality rate is um i really there were these stories that maybe lots more people had been infected there were lots of asymptomatic cases and they all seemed quite plausible to me um i just think more and more information came in from antibody tests from particular circumstances where you had a community where everyone had been tested many times and it just looked like i wish it was true uh but it looked like the infection fatality rate in which countries probably was closer to one percent falling over time because we got better uh you know better treatments and so on but um you know i really hoped that that you know i wouldn't have to embrace that conclusion but in the end i had to the other thing i've changed my mind about uh and i still don't know whether i'm right about this then i changed my mind about mask wearing but early on we looked at the evidence for mask wearing and it didn't look that strong really um and but more and more places started embracing mask wearing more evidence did come in and more epidemiologists were saying oh this is this is useful this is important and so now i would all you know would always sort of have gone into um indoor space wearing a mask you know and wear a mask wherever it felt appropriate um i'm not sure that i really ever saw the evidence change my mind changed and as i reflect on that i wonder whether that was just a case of of just going with the consensus and not wanting to be disagreeable uh i don't know whether i was right to be skeptical in the first place or i was right to change my mind but i did change my mind what about you david oh god i always wrong about everything i mean i'm right at the beginning of this yeah i i kind of just couldn't believe that we would go for these measures of locking everyone down and wearing masks around i thought oh you know that's not the kind of thing that happens to us that happens to other people who get things like mers and sars and stuff like that so i think i was eternally optimistic not in terms of the infection fatality right i you know believe that right from the beginning and those that's been has been absolutely justified you know by further by further data but i was generally generally been optimistic about um the consequences in terms of deaths and so on um so i've i've been proven wrong on a number of times yeah yeah because the world the world is an interesting place and new data always comes in and it is it is important uh as we've been encouraging to see certain people on twitter just actually make up almost a performance of it to say here's a list of things that i was wrong about and here's why i've changed my mind and i realized there's there is a little bit of performance about that but i think they are setting a very good example i think i mean this this humility of admitting uncertainty but of course the lesson should be not that oh i can be wrong but that in future when you have do make a claim just be a bit more cautious and acknowledge the uncertainty and also the strength of evidence behind your claim much as you might want to be confident and certain everything my thing is whenever i've been confident uncertain about something i've always it's always been wrong so i i'm trying to bring that into my into my daily practice and there was this wonderful paper in the in the by michael blaster and others and who on this need for humility and acknowledgement of uncertainty and in the british medical journal at the end of it they listed all the things they've been wrong you know it's a real sackcloth and ashes you know they're they've flagellated themselves in front of their audience so i thought that was that was that was very good and i think that i i kind of do hope that at the end of all this um this need to acknowledge uncertainty um will be possibly a laugh a lasting lesson i don't know what do you think you know the long-term consequences of all this in terms of the statistical interest or even literacy of the public might be well i'm i'm slightly hopeful i've been talking uh to the vaccine experts i've been making a bbc program uh for a while now called how to vaccinate the world and one of the things they tell me is we've learned amazing things about how to make vaccines really quickly we've got all these cool new technologies and this is potentially revolutionary i do hope the same uh is true about statistics that we're learning things uh as a society about the importance of the numbers we're learning how how useful they can be in showing us the world around us and we're learning that it isn't i mean of course yes politicians use numbers to win arguments advertisers use numbers to sell us products that's not going to change but there is something else that we can all do with data uh and it's for all of us uh to understand the world around us and and i think that i maybe i'm naive in this but i i think that that lesson may well be it may well stick for some people at least yeah i mean i suppose i'm quite optimistic that to have so much public attention focused on statistics and epidemiology and evidence and the claims made on evidence you can't think it must be a good thing it must you know and and to realize that scientists can argue with each other in science it's not just you know a single body of facts scientists say boffins do this i mean this is complete nonsense which we've known all the time but i i think that this in a way that the dirty linen has been aired in public um actually i hope will help perceptions in the future i mean i i must say you know we've laughed about the fact that oh everyone's an armchair statistician now and now we both get sent continually people's analyses and claims and arguments and things like that i i do try to look at them if people are polite i do try to look at them because i have learned a lot you know and uh you know i really people have got some interesting insights and i have learned a lot from my regard it's just people i don't know sending me stuff i i you know if anyone's watching this i might say don't send me stuff because i can't cope with it but you know i do try because i do you do learn and you actually if you can you know get i said if they're not polite i don't bother but if people are polite i'll i will make an effort yeah i mean it is that there's a lot of email so it's partly a question of the sheer um sheer time involved to to read all this stuff but certainly or more or less um i do encourage people to send stuff to more or less rather than send it to me because we just there's there are several people some of whom work full time i can actually go through and we learn so much from our listeners so many interesting ideas uh sometimes really good answers often just really good questions but really good questions i think uh that's enough um to to look around at the world and to start asking smart questions is really something that i'm hoping to leave uh readers of my book with i do i do have this final chapter about curiosity and i think it is important because there is this tendency for people to to use numbers as a as a weapon now i i want to win an argument i'm going to prove i'm right in the pub i want my side to win the election whatever and that's what numbers are for and of course that is how they will be used by some people but i think if you have a different view where you say i am interested in the world it's a fascinating place the economy the environment disease risk all of it it's all really interesting and it's something that i can only understand with the help of the numbers if that's the approach that you have genuine open-minded curiosity uh you know i think that life becomes so much more interesting and numbers become not weapons but friends yeah and i'm not sure if we should even think of numbers because what we're really saying is about magnitudes the fact that you know is something big or small you know the magnitudes rather than just things being true or false i mean you know one of my favorite things that more or less did what did you do if you took all the jam in the world and spread it over england how deep would it it would be you know this is a sort of this is a sort of thing that we that we need a serious government grants to investigate you know well no robert robert eastway uh the author of maths maths on the back of the envelope he figured it out for us and he reckoned and i don't think he might rigged these numbers but you never know he reckoned there was there was basically enough to make a cornish uh jamming cream tea exactly the size of cornwall there was enough jam in the world to do that once which i think was a very pleasing number okay so with that final image of cornwall completely covered in jam um i think we should wrap up now and it just remains for me to thank on behalf of everyone who will watch this um tim harford for doing a being a good sport and uh and answering all the questions i've actually really wanted to ask and i i hope that some of those are the questions that you would want to ask too so thank you all very much for watching us and thank you very much tim for uh giving us your insights david you've been wonderful thank you so much been great [Music] you

2021-04-01 06:27

Show Video

Other news