The second problem looked like the first, so people ran the first method and never noticed the shorter arithmetic sitting in the same numbers. They were not careless. The earlier problems had taught them a procedure, and the procedure still produced a correct answer. Correct was enough. Short never got a look.
Marcel Binz and Eric Schulz, at the Max Planck Institute for Biological Cybernetics, reconstructed the tables from Abraham Luchins's 1942 doctoral thesis in Reconstructing the Einstellung effect. Luchins asked people to measure a target amount of water using jars of stated sizes. A first problem uses two jars, 29 and 3, to get 20. Five problems after that all yield to the same longer method: the middle jar, minus the first, minus twice the third. The problems that follow can still be solved that way, and they can also be solved by a shorter one.
The long way still works
One of those later problems offers jars of 23, 49 and 3, and asks for 20. The practised method runs 49 minus 23 minus two times 3. The short method is 23 minus 3. Both are right. Luchins called the long one the Einstellung solution, a mechanised state of mind, and the short one the direct solution. People who had worked the five practice problems kept reaching for the long method. In his first comparison, that group used it on the first two critical problems between 77 and 82 per cent of the time. The control group, who never saw the five practice problems, never used the long method at all.
A later problem cannot be solved by the long method, only by a short one. The pull was stronger on the earlier critical problems than on the later ones. A workshop that ends on a correct answer has stopped where the long method still looks like skill.
A warning is a weak solvent
Luchins also told some people, once the practice problems were done, to write the words "Don't be blind" before they continued. Binz and Schulz, transcribing the tables, found that 87 of 153 people in that condition still used the long method on the first two critical problems. That is 57 per cent, against 83 per cent in the plain condition they compared it with. The note helped. It left most of the practised method in place. A slide that says to stay sharp is that note. It is not a case where the short method is available and someone scores who used it.
Speed made the set stronger. In the tables they transcribed, 95 of 98 people under time pressure used the long method on those first two critical problems, against 126 of 153 without that pressure. A day that rewards the fastest correct answer pays people to stay inside the procedure they just learned.
What the agenda gives away
A weak proposal teaches one procedure on several similar cases and then congratulates the room for applying it. There is no item the procedure solves the long way when a short way is sitting there. There is no item the procedure cannot solve. The team leaves fluent and blind to the direct step, which is the result Luchins was measuring. If the cases are toys with one intended method, you have bought the practice problems and skipped the test.
Collective Campus has done capability building for teams since 2015. The critical thinking workshop is for a team that can bring a procedure it already trusts and a case that procedure makes too long. Whether a general label travels to a later call is in how to judge the day for transfer. Why the check matters once a model is in the draft is in why critical thinking beat AI literacy. What happens when the first fluent page is kept is in the first draft a team keeps.
What to bring
Bring the procedure the team already runs well, and one live case where that procedure is correct and also too long. If every example has only the long way, the day will rehearse a set and call it judgment.
