Hamburgers
Peter, Fraser and Nick discuss the threat posed by rare hamburgers, and ways of assessing risk in the absence of data.
For more information on Aleph Insights visit our website https://alephinsights.com or to get in touch about our podcast email podcast@alephinsights.com
Transcript
Hello and welcome to the Cognitive Engineering Podcast with Aleph Insights where we take a look at interesting topics, we discuss what we think they tell us about analysis and decision making. I'm Fraser McGruer and I'm here with Nick Hare and Peter Coghill of Aleph Insights and this week we're discussing undercooked hamburgers and the perception of risk. Okay so as you both know undercooked hamburgers are a bit of a bugbear of mine. Gourmet hamburgers are an extremely popular thing in London at the moment and probably elsewhere but part of that popularity or fashion is to have them cooked rare or medium so they're still a bit bloody and it's my understanding from good sources like documentaries on TV and Panorama I believe and things like that that this is not good because with minced meat, unlike a steak for example, with minced meat bacteria gets mixed into the centre of the hamburger and needs to be cooked thoroughly to kill it off which doesn't happen if you're having a hamburger. But I think you both, I know Nick certainly disagrees with me on this, perhaps Peter as well, so tell me why I'm wrong to not want to die of food poisoning from an undercooked hamburger. Well it reminds me of
Speaker B:when I was at a NATO conference a few years ago and it was in Europe somewhere and there was a US representative there who was who was based in NATO and he said that when he first got there, first got to Europe he was disgusted to find that in pubs, in bars, they didn't sort of steam clean their glasses they would just give them a quick squirt under the tap and then fill them up and give them to another customer and then he said but you know the thing is we're all still here and that's really what I'm saying. Now the interesting thing about trying to estimate risks, how likely are you to die of food poisoning through eating an undercooked burger? Well the question is have we got any data on that? Well we probably don't have access to a really good data set for the country as a whole but we personally simply through virtue of knowing lots of people and through the assumption that dying of food poisoning through undercooked burgers is something you'd hear about if it happened to one of your friends. We actually do have quite a lot of data we know we probably know between 100 and 200 people and each of them knows between 100 and 200 people. So my question to you really would be who do you know who it's happened to? Well I'm
Speaker A:not going to answer that question because I don't think it's a valid question because let's say the risks one in thousand for example I only know about 100 people let's say and even though I know there's 100 people whenever we have chats about stuff our first topic of conversation is not the the daily disasters that we've managed to avoid. If one of your friends died of eating an undercooked
Speaker B:burger you would definitely know about it. Now you've got 100 friends and you've known them for let's say an average of 10 years so that is already the one in a thousand more or less you would if it's one in a thousand you would have expected it to have happened to one of your friends by now. I think one in one thousand is probably too low. I think and bear in mind obviously we are only talking about the population of people who are potentially going to be eating undercooked burgers. So it's probably lower than one in a thousand let's start there and I also like to think that probably if one of my friends' friends died of eating an undercooked burger they'd be on the phone to tell you about it. Not necessarily but I may have heard of it I mean I might have heard about it that would be surprising. Okay well hold on does anyone here
Speaker A:know anyone not just from hamburgers but because people always talk about oh I had a bit of food poisoning you today yesterday and I always think that no you didn't you know how do you know and and so does anyone here know someone who's definitely had food poisoning of some kind? Oh come on died of it that's what we're talking about. No no no no let's no let's start low does anyone here know anyone who's had food poisoning it definitely was food poisoning?
Speaker B:Yes yeah okay right all right yeah I've had it loads of times so so we know that the probability of getting some sort of food poisoning is probably I it probably happens to me about you know once every uh few years so it's you know with a normal kind of diet you're probably looking at one in five one in ten chance per year of getting food poisoning. Okay but let's look at this in another
Speaker A:way I've got some statistics in in front of me here which I'm not going to go into them just yet but let's sort of turn it around slightly and say well look if there is and actually Peter perhaps you can answer this if there is some sort of risk of some kind shouldn't we just just just avoid that okay and look at the trade-off here to have hey no I want my nice juicy bloody burger and well I mean it sort of it comes down to really what do you want to get out of life I mean if you want
Speaker C:a really boring life where you take absolutely no risk and do nothing at all and stay at home that's fine you'll you'll totally maximize your uh your sort of lifespan uh saving you know ignoring the fact that you might you might suffer sort of psychological trauma of being locked away from from everybody but you'll you'll you'll live you'll live to 90 100 easily um but you'll have had a very boring existence you know you you you won't have had uh those juicy burgers they serve that you won't have had the thrill of uh of eating wherever you wanted whatever you wanted um uh and for me the risk is kind of worth it you know I haven't done the maths but I know uh through the sort of friends of friends of friends uh uh kind of methodology that you know I don't know anyone who's died of eating a dodgy hamburger not yet um not yet I'm not saying it's impossible but I've not I don't know of anyone so my my sort of perception of that risk is quite low uh it's certainly lower than the enjoyment I think I'll get from having it having that burger
Speaker B:okay okay whereas I do know for example I know people who have not been killed necessarily but have been seriously injured uh in doing sort of sporting activities so so you know that is a sort of risk I know people who've who've been severely injured uh cycling um probably I know the same number of cyclists as I know really committed burger eaters so I know that the probability of getting of getting food poisoning from an undercooked burger is is probably considerably lower than than of being injured while while cycling well in defense of your
Speaker A: at year this was I think it's: Speaker B:is that and I can't remember the figure offhand for murders in the US yeah but it's I think it's the tens of thousands certainly I mean in the UK it's about 500 a year I think but in the in the US it's because it's it's tens of thousands I think um but but you know contrast that with the deaths from E. coli which you just admitted would were 31 you should be a lawyer it's that is that is uh it's how many orders of magnitude well they're about 300 million I think yeah but but that I mean it's it's not quite it's it's in a different it's in a different league
Speaker A:I like I like the way you I like your language there that you have admitted with and we can
Speaker C:improve this model slightly okay we can probably discount a large number of those deaths because they they will be eating in the sorts of places that we would not choose to eat right so places with are clearly of a lowest food hygiene well actually incidentally part of the same study
Speaker A:showed that um that uh where most of this happened actually was uh in the home that's the sort of worst place to do this well yeah just poor hygiene practices okay but then yes maybe
Speaker C:I wouldn't cook an undercooked burger at home I don't mainly because I don't eat burgers at home but I know that any burger I get out the freezer or I get from the butcher hasn't been freshly ground you know hasn't been ground from from steak meat very recently okay whereas you know I put my trust in restaurateurs when they say this is a steak ground beef this is steak ground beef then yeah that's probably and that's good for you um okay look I just want to move it on
Speaker A:slightly so um can I can I actually unmove it on sure it's just it goes back to what you were
Speaker B:asking earlier which is about about the question of the extent to which we should be trying to um trade risk off against benefit if you don't like undercooked burgers then not eating them definitely strictly dominates eating them if you just simply rather have a well-cooked burger because you like the taste um now the question is that that trade-off is really fundamental the trade-off between a an out trying to do something which has a higher benefit but also potentially a higher risk is the essence of decision making under uncertainty it's a um it's and the interesting it's interesting well obviously when you now the thing is it's relatively straightforward when the probabilities are in about the kind of yeah 20 30 percent that sort of order of magnitude and the difference between the outcomes you know maybe it's twice as good or or whatever the um it becomes much harder to manage certainly cognitively and and statistically as well um when when the probabilities get very very small and the outcomes are extremely um extremely large uh one side or the other so when you're you know that's why that's why we have a there's there's a real a statistical um field really looking at these very extreme events of which we have very uh few statistical examples so things like what's the risk of there being a sort of catastrophic um meteor strike or a catastrophic um nuclear war or something where where very small changes in how bad we think those things are or very small changes in what we think the probability are uh have a very dramatic impact on what our optimal decision is okay i think i got most of that um
Speaker A:but i mean it often seems to be the case when we talk about stuff that it boils down to um hey everything looks like it's like this but actually when you have a closer look really it's like this um and in this particular case i mean we've we've we've kind of talked about it here a little bit but when we first discussed this over a hamburger actually um you um yeah you said have you sort of you know my friends model that sort of thing so you know do you know anyone who this happened to and do they know anyone who this has happened to in this particular case it was hamburgers now so is this is this have you have you patented this model no and in fact to be fair
Speaker B:uh i it was inspired by something uh steven pinker said in his book um the better angels of our nature which looks at the decline of violence in um across the globe in all sorts of different categories and um he is using it as a way of putting into context um murder frequency frequency statistics which are always expressed for some reason as um is numbers per hundred thousand and um the question is you know if the idea being that if you lived in somewhere like columbia which has a murder rate where i think that figure is in the tens per hundreds of thousands yeah um you are likely to know someone who has been murdered but if but if but if uh if it's lower than that if it's sort of in the ones then you will know someone who knows someone who's been and if it's in if it's considerably lower than that as it is in the uk yeah even that is very most most people won't know someone who knows someone who's been murdered so and that that sort of inspired me to have a bit more of a think about that and and and then really just realize that we we all have access to our own personal quite sizable data set um which we could use in principle just just at least for a starting estimate of the risks of things can you give me two or three other examples where you might use it um well i i've got some real life stats which we could uh which we could use in fact we could start here right as long as you're between being the victim of a burglary okay so what who do you fraser do you know yeah have you been the victim of my family was when we're right okay so that's a friend uh peter been the victim of a burglary personally no okay like i have so we've got um i you know it's happened to me once and and uh or twice so i reckon the order of magnitude is talking we're looking about sort of one one in 50 100 a year and in fact that is in the uk about right uh you using your you know has it happened to a friend um would would come up with a um uh would come up with a lower estimate but perhaps burglary is less frequent in the lovely places you live others other well indeed other stuff
Speaker A:anything else uh divorce okay been divorced uh no not personally not personally no i'm not
Speaker B:personally either early yet so but i know has do you know other friends who've been divorced you have friends um yes so uh you know we're talking about one in a hundred to one and one in a
Speaker A:thousand sort of estimate but hold on isn't that sorry was that similar or not per year so i'm talking about per year not ever in your life is that similar to the numbers for burglary yeah it's not far off so you're as likely to get divorced each year as you are no no you're more
Speaker B:likely to be burgled um it's about one in 300 people per year get divorced not one in 300 married people obviously that's the thing if you're actually married that your risk is considerably higher than if you're not married it's very unlikely so the key to not getting divorced is clearly to not marry i i exactly so i mean and again this goes back to our our issue about are we you know we've got to trade off the slightly higher benefit perhaps slightly higher that's a bit mean the the higher benefits of getting married with the potential perhaps risks of
Speaker C:distress if you get divorced and we've not begun to discuss the um the conditional probabilities of what likelihood of getting divorced having been burgled but that's a whole nother topic
Speaker A:right exactly well that would be interesting i bet you'd be more likely to uh to get burglary
Speaker B:we i mean i think this is a topic for another uh podcast is why we why we the extent to which we should use expected outcome as a guide to to what we should do um it's not it's not actually as straightforward as as it seems very good well that seems like an appropriate note to wrap up well i just want to add though that the menace is is not um undercooked burgers it's brioche buns and i'm not even i'm not even against brioche buns particularly but i am against not having a choice
Speaker A:okay and um but but do you like because where we go i mean we've just just to highlight we've not been recording these uh for that long but we've certainly fallen into a routine of after recording a podcast we often go off and have a hamburger and i believe at that place they do brioche yeah there's no choice right and but given the choice would you rather not have a brioche bun
Speaker B:with you sometimes i think i'd like to have a toasted sourdough bun okay i don't think that
Speaker A:makes me a criminal well yeah i i think you know that i agree with what you're saying all that sounds fine and we're in favor of choice um okay well look we'll wrap peter anything to add well have we talked to you around frazier how are you having a burger not really no no no no no so no absolutely not so today i will not be having a bloody burger uh because no you've not talked me around at all because i think um i think it is potentially dangerous you can get e.coli and die horribly and i'm actually a not i'm not a risk averse person i will take risks you know i will do certain things but um i just know it's just there is a danger still you know and i yeah i do like my steaks uh sort of medium etc and i probably would i do probably like the taste of a burger that's medium it doesn't dry out etc but but now i've seen these devastating statistics um i'm i'm sorry i'm just not convinced by your argument determinedly irrational and on that note we'll wrap up so thank you to everyone as always for listening to the cognitive engineering podcast with aleph insights um i'm fraser mcgrew and we've been here with peter cogill nick hair thank you very much until next time bye
