I've been assessed during a similar group assessment (we had to work together to produce a layout for a shopping centre), and also been an assesser for one of these excercises. All you are really trying to identify are:
1) The people that won't help others, and will intentionally try to throw other candidates under the bus (so you make sure you never hire them).
and
2) The people that do genuinely try to help the other candidates who are struggling (so you can hire them).
It's not exactly some sort of highly-complex psychological evaluation, you are just seeing how people work in a group task.
1. "You're all being mired in nonsense. Give me the board, hand me the packets, and I'll show you how it's done."
2. "Okay, let's organize this somehow. I'll make a list of related widgets and gadgets. Becky, can you see if we all have the same dooh-dad structure?"
I'd say 1 is an asshole, and 2 is a Team Player™
We are seeing how people can work in a team to complete a simple task to help a hiring decision, not replicating the Stanford Prison Experiment.
Do we need professional actors and casting directors to judge the role-play excercises? We should probably get journalists to do the interviews too.
I mean, if you had to work in a team with someone.. this sounds like a good way to see how they'd fit in actually working in team under a pressured situation.
You're not labelling them as this or that, or necessarily checking for 'hidden indications' that they could have some condition you know nothing about, it's just finding people that fit into the culture that you're looking to fill.
I (probably awful) analogy could be, if I go to car boot (garage) sale and find things that are good value (and/or haggle).. I'm not telling the person how much the item's worth, but what price it's worth to me/what I can afford.. I'm not claiming to be a antique evaluator.
(Update: Well, I guess I don't _actually_ know they're doing/not doing, but it's my interpretation)
Do you know if a better plan which scales to companies with hundos of thousands of employees?
I doubt a similar baking mistake would get a similar response than a wrong-order mistake in a restaurant. Maybe it does, or maybe it doesn't, but making a prediction from one to the other based on such little evidence... well, the person who gets tested really is the one doing such a test, and it does not look good. Okay, there's an A for effort, at least he tried it at all.
I'm sure the very-negative case works, if he freaks out or becomes too obsessed over such a mistake that speaks against the person. But the cases you described sound pretty tame, if he merely asks for the mistake to be corrected that in itself isn't noteworthy. I certainly don't see it as being useful for determining what he would do if there was a similar mistake in the bank business. The context is totally different.
Honestly, I'm the guy who would just "roll with it" and eat whatever was brought to me because I honestly can't bring the energy to arguing with the waiter.
But having hired people and other professionals, I almost always hire the person who would assert themselves and demand the right thing be done.
Why?
Well, I recognize that I don't like to be "that guy". So I hire people who can do it for me. It has always worked out quite well in the past, where I've been able to form a team of people who work well together, but can be assertive to external groups. (And I, of course, only need to get out of their way.)
This is a false dichotomy, not to mention plenty of people who _are_ abrasive and rude to service workers would probably have the smarts to not do so in that situation... Kindly mentioning that the order was wrong (especially in the case of an allergy/health issue) shouldn't be seen as assertive or demanding and I hope that's not what they were implying!
It's knowing P(A|B) and trying to obtain a P(A|C) where B and C are mostly disjoint, making the info not too useful.
Depending on what was brought I might not even be confident that they messed up the order-- e.g. if what they brought was also something that I credibly could have ordered.
I hate dark patterns in dates, that was a date. Learn to communicate and choose that.
Author says they were shadowing, not an expert.