Friday, January 4, 2019

Expert/Layman Triage Fallacy

I love ferreting out common cognitive failings that are otherwise unnoticed. This one is huge - you can spot it everywhere - but I've never seen anyone point it out. It’s so unilluminated that it’s nearly impossible to make someone see it when they’re in the midst of it. In fact, it's so hard to explain that I can only do so anecdotally. Two examples:

Example #1

Computer Programmer: "Please stop adding tasks to this job. I am full up. Consider the request line closed unless you're ready to pay me a ton more money."

Me: "Ah, ok, sure. It's a pity, as I was really hoping you could make the thing do the thing when they do the thing."

Programmer: "Oh, well, that’s trivial. I can do that, no problem."

Me: "Oh, great! Also, it would be nice if it did the opposite thing if you don't make the thing do the thing when they do the thing."

Programmer: "That’s just a simple toggle. You’ll get that, absolutely."

Me: "Wow, fantastic! Hey, one last thing. Those last two turned out to be trivial, and there’s another, though I will not - repeat: will not! - add it to your task list if it's time-consuming. If it's okay, I'll ask just on the odd chance that, like those two previous things, it turns out to be super trivial. Can’t hurt to ask, right? Hahaha! Would it maybe be super easy to make the blue thing act like the red thing when they press the yellow button? Again, if this is a pain, just ignore this."

Programmer: "Jim. My friend." [Speaking through clenched teeth] "I enjoy working with you, and I wish you every success with your project. But I have already informed you that the task list is closed. What you ask would take hours of work. Kindly stop burdening me, effective immediately."

Me: "But...but...but...you just said...."

Example #2

Patient: "I'm not sure if this is meaningful, but my handwriting keeps getting smaller and smaller."

Doctor: "That’s very meaningful. It's a classic warning sign of Parkinson’s. Hold on while I muster every possible medical resource to stave off this debilitating illness."

Patient (a year later): "I have this strange numbness radiating down my left arm. It's not really bothering me, but it's weird. And I can't quite catch my breath."

Doctor: "Thank goodness you told me. You are quite possibly having a heart attack."

Patient (a year later): "I know this is crazy, and I doubt it's of any importance, but sometimes minor stuff's significant, so here goes. My thumbnail is growing more quickly than usual."

Doctor: "It means Spring's coming late this year? Maybe you're part lobster?" [suppresses eye-roll, mentally noting this classic example of petty patient alarmism]


Patient: "But...but...but...you just said...."

Followup posting here.

8 comments:

Richard Stanford said...

Very well observed.

As a software consultant, one of the hardest things to teach a client to do is to ask, as simply as they know how, for what they actually want. People are pretty bad at guessing what things are easy and which are difficult, and quite frankly we should allow them to focus on what's important rather than teaching them the nuance that often takes seasoned experts years to master.

ObXKCD: https://xkcd.com/1425/

Where I find this reasonable difficulty comes to be a serious problem is exactly when lay-people attempt to make things "easier" by asking for what they think would be a simpler version of what they want. Many times I've gone down the rabbit hole trying to figure out something close to impossible (or at least very expensive), when it later came out what the customer actually wanted seemed far harder but would have been trivial to implement. That's why you should never make someone feel silly for asking the questions.

I'm guessing that if doctors presented a list of "interesting trivialities" to see if people experienced ("Handwriting changes," etc) that they'd get far too many false positives for it to work. Unfortunately many warning signs do fall into that category and serve more to make us (and the experts) feel bad for not seeing something sooner, when in reality it wasn't reasonable to expect us to - and we should treat them more as confirmation of a diagnosis.

Jim Leff said...

All true, and all good points. I particularly liked that XKCD comic, which I hadn't seen in a while. But that's not quite the issue I'm hoping to highlight (not to say your comment wasn't welcome....I don't need comments to stay squarely on-topic!).

I'll strain to state it flat out (a good exercise; thanks!): Expert frustration/bafflement with laymen gingerly inquiring about triage patterns the expert fails to realize s/he's previously established, resulting in disconnect for experts (bafflement at best, anger at worst) and confusion for laymen. That's clunky, but it's the best I've got for now.

It's probably a subtle variation on the "forest-for-the-trees" dilemma.

Richard Stanford said...

In the vein of missing the mark, that was (part of) the point I was attempting to make - that professionals expecting laymen to understand those nuances is doomed to failure. Similar to the Dunning–Kruger effect, we (professionals in general) tend to minimize the necessity to have years of experience in order to make heads or tails of those points, and learning how to accept the layman's point of view is both completely worthwhile and thoroughly difficult.

Richard Stanford said...

You're right though - expecting frustration is probably a very good place to begin. I had jumped into the idea that the expectation of non-frustration was flawed, but your stated point is even more fundamental.

Jim Leff said...

So you're suggesting that all buckets should remain open (i.e. let experts triage, not laymen, because laymen are inherently ill-equipped to do so). The problem is that customers/patients truly ARE annoying when freely encouraged to inject random noise into the workflow of busy professionals. Doctors can't encourage full reportage of every bodily oddity, nor can programmers spend hours assessing mammoth Trello task queues. In fact, that's how we got here! Experts can't spend infinite time on triage, which is one reason they affect that typical briskly aloof manner...as if to say "I'll listen to you, but have limited time/patience for nonsense".

But I'm describing more of an edge case - when customers/patients do hesitate to overwhelm experts with triage tasks - but still have it blow up in their faces because experts....well, I'm not sure what exactly happens. They get baffled or angry....even if you preface your inquiry with the many self-aware disclaimers I used in my examples. They cannot recognize the pattern they themselves have established.

It's an expert failure, not a layman failure, but the solution can't be to expect experts to engage every byte of nonsense flying their way.

Richard Stanford said...

My poorly communicated thought with, " Unfortunately many warning signs do fall into that category and serve more to make us (and the experts) feel bad for not seeing something sooner, when in reality it wasn't reasonable to expect us to - and we should treat them more as confirmation of a diagnosis." was that seeing it as a failure is fallacious and that it may well be that expecting anyone (laymen to notice it or experts to sort it from the chaff) to use those factors as anything other than confirmation is unrealistic.

Interestingly this is an area where AI may start to help, since it can both observe and sort through astonishing amounts of chaff, and may (may) be able to figure out what things to bubble up as significant.

Jim Leff said...

You seem bonded to the accurate and interesting observation that laymen see a blur of minutiae both salient and (mostly) not and that it's difficult to find a viable way to bridge that with the experts with whom they interact. But you'd be closer to the point I'm struggling to make if you'd consider BOTH examples I gave, and, particularly, the social/emotional outcome (me sputtering and the expert exasperated).

I don't want my doctor to spend unnecessary time on fast-growing thumbnail, or my programmer to feel burdened by an additional difficult task. I'd just like them to be able to recognize the (really fairly clear) pattern of triage they themselves established with me, or, if not that, at very least, take seriously my ample disclaimers and pre-apologies rather than disgustedly assume I'm doing the thing I've just exhaustively disavowed doing.

Display Name said...

Jim I see that last comment as a familiar pattern in my life. I often choose my words carefully to avoid misunderstanding to no avail. You can be the best communicator in the world but if the receiver is faulty/turned off not able to hear it's worse than useless.

Blog Archive