Only if the question is designed to be a “gotcha”.
You still have the fact that an unexplained an highly atypical ratio of major:minor accidents is subsequently presented without explanation front and center. Only then is the ambiguous question asked. And you are supposed to ignore the second prominent thing and go back to check the exact wording of the first thing to find out which of two ambiguous interpretations of the question is correct?
As the question is designed to highlight a specific error in human perception it necessarily retains some ambiguity and imprecision and of course exists in the realms of the abstract.
However, I wouldn’t call it a “gotcha”, I’ve seen real-world problems and scenarios that very closely mirror exactly this sort of imprecise framing and have seen people misinterpret, misremember, assume and avoid clarification in exactly the way that many have in this very thread.
I do agree though that were I given this scenario in the real world I would apply the data “sniff test” and question the likelihood that we’d see major:minor injuries in the ratios described.
As this is a hypothetical though, if we aren’t going to fight it then we go with the data as shown and other than that the scenario is valid enough for me.
I can see exactly what kind of cognitive error the writer is trying to illustrate, and I think the scenario is very poorly constructed to do so.
If you’re going to try to illustrate a specific type of cognitive error, you are fair game for criticism for the fact that the scenario you came up with is unrealistic for that purpose, and that you yourself made a cognitive error in failing to recognize that the word “equivalent” is ambiguous and undermines the premise.
But one absolutely valid answer to the question would be "It can’t be answered unless you more precisely define “equivalent” ".
But the purpose of the example is to show that once people grasp on to the rather obvious “correct” answer then they tend to cling on to it and not ask those more fundemental questions or challenge their own preconceptions and check their understanding.
The ambiguities in the question are an important part of the question.
Is the type of cognitive error he’s trying to illustrate just a generic failure to ask for more information when wording is ambiguous? I don’t think so. I think he was just careless in his wording of the question.
I’m curious how you would describe the kind of cognitive error you think it’s designed to illustrate, and what scenario you think would be better at doing so.
Okay, I have an idea but I’m not exactly sure. You’re determined to make me watch the video, aren’t you? 
It isn’t just that no, but by removing all ambiguity in the wording (in a way that real-world scenarios seldom do) and being completely precise in all aspects of the question it would loose the ability to highlight that common human failing.
It would be as if a magician walked you through a card trick and purposefully pointed out the misdirections as they did it. You would end up un-fooled.
It’s true that the question, as posed, does not have a single definite correct answer. That’s not a failure of the math or of the set-up, though, because it’s a property of the real world: The real world does, in fact, contain problems which do not have a single definite right answer. But that doesn’t mean that math is completely useless, because we do know some things about the real world. For instance, we know that a major accident is definitely worse than a minor accident (even though we might disagree on how much worse). And so we know, with absolute mathematical certainty, that the answer to the question must be something greater than 1016, and likewise that any answer less than 1016 must be wrong.
If one person says they think that the answer would be about 2000, and another says it would be about 10,000, those two people are both engaging with the actual problem, and can come to an agreement within the common framework they’re dealing with. Maybe they’ll go looking for statistics on how bad, on average, the minor vs. major accidents are. But if one person answers about 10,000, and another answers 8, only one of them is working on the problem they’re supposed to be working on.
Another problem is that it is a false dichotomy, accidents are either major or minor, no other options. In reality, injuries range from papercut to slow, agonizing fatality. For any useful mathematical model you would have to have several levels of seriousness, I think at least 5, and the more the better.
TL;DR. I don’t see that the question is answerable without establishing some ratio of “n minor accidents are the equivalent of 1 major accident”, which isn’t given.
The issue I have with the question is that I don’t feel like it’s always clear what “make equivalent” means. I could see people thinking it means “make proportional.” And that isn’t the same cognitive error as the other examples.
That said, I’m not entirely sure if the original question asked by the psychologist was worded that way.
True, but if you carefully listened to and remembered the first part of the question and understood that they were looking to evaluate which junction to build, then “make this number proportional” makes much less sense than “how many minor accidents are needed in model B to make it the worse option?”
In fact these problems are faced all the time. And people deal with them irrationally all the time.
How many young adults having two to three day admissions for vaccine associated myocarditis offset preventing how many deaths from Covid?
So on.
The immediate response to the problem of the OP is a recognition that “major” traffic accident implies some frequency of death and disability in addition to more easily quantified costs, and that minor implies some costs of repairs and possibly missed work. Really “minor accidents” can have monetary value placed on them without much difficulty.
What is the frequency of death and of disability in the major accidents? How much are you willing to spend to prevent each of each of those? Is every death or disabled individual worth the same amount? Or is a healthy fit 20 year old worth more than a person who is 90 with dementia (the model of quality adjusted life years, QALY, saved)?
They are the real world issues. What is a life saved, or a QALY saved worth?
Are you sure? “Equivalent” can mean many different things–mathematical, financial, time lost, cost to the taxpayers, etc.
So it illustrates that when a problem is posed in a manner that is unrealistic, contains ambiguous wording, and is presented in context that perhaps misleads people to expect that there is supposed to be a “correct” answer given the information provided, that people sometimes don’t reason it out well, and don’t always conclude that the rational response is to question the unexplained strange aspect of the hypothetical and request clarification of the ambiguity?
Count me underwhelmed. This has demonstrated nothing surprising about human cognitive ability.
The interesting part to me is not that they come up with a wrong answer. It’s that a particular wrong answer is so common. Surely that shows some sort of cognitive pattern, a desire to make things proportional.
Well, yes. That’s sort of the point. It is an artificial example to highlight a real-world phenomenon.
When you state it like that it sounds ridiculous that people don’t balk at the strange set-up or seek clarification on the ambiguous wording, but they simply accept it and look to make sense of it (and often fail).
Read back what you wrote. Is that not a great description of most classic magic tricks or illusions? How easy are most people to fool? how well do we fool ourselves?
Not surprising to me either but many people aren’t aware that the same cognitive failings that lead us to be fooled by magic etc. are also at work in rather more mundane settings such as this (superficially) simple “maths” question.
This. Exactly this.
If a professional question-asker mis-frames a question in a way that leads amateur question answerers who’re expecting a game / quiz to deliver not only a bogus answer, but a specific predictable bogus answer, all that’s been proven is that framing works. No shit, fellow Sherlocks. People will answer (the simplest version of) the question you appeared to ask.
If the question had been instead: “Design a decision framework for trading off major vs. minor accidents versus various designs of roadway intersections”, now we’d be starting to ask the question the vidiot usefully could have wanted to ask amateurs and be able to get useful feedback.
It’s now late in the thread but I have to give the minor threadshit that I really dislike threads where the title asks one question, the OP body discusses something significantly different, and the actual question to be discussed can’t be determined at all without watching a friggin’ video.
Please, please put the relevant content and context in the post, not in a cite. And especially not in a cite to a video.
and not just a wrong answer, a massively wrong answer.