A repeatable procedure for the two types most students lose marks on.
Draw the smallest, most cautious diagram the statements allow — overlap sets only where the statements force it. Then test each conclusion against the diagram: it follows only if it must be true in every diagram the statements permit. If you can sketch one valid arrangement that breaks the conclusion, it does not follow.
Default to No unless the diagram forces you to say yes, and read the wording like a lawyer — all X are Y is a different world from some X are Y. The killer mistake is assuming: a conclusion that probably follows does not follow. This is logic, not likelihood.
A quick way to test: flip the statement and hunt for a counterexample. All doctors are intelligent; Sarah is intelligent; therefore Sarah is a doctor? No — all doctors being intelligent says nothing about all intelligent people being doctors. One counterexample ends the conclusion.
And always draw. Your brain is for thinking, not for remembering — circles, boxes, any sketch beats holding three statements in your head.
Convert the clues into a small table or an ordering line on your noteboard and apply them one at a time, starting with the most restrictive clue — the one that pins something to a fixed position. Test possibilities systematically and eliminate anything that violates a constraint. Do not hold the puzzle in your head: working memory is exactly what the question is designed to overload.
Expect the real exam's puzzles to be more abstractly worded and need more steps than practice versions. That is not a reason to panic — it is the reason the prepared students pull ahead here.
Decision Making gives you a generous ~63 seconds per question, so the discipline is the opposite of VR: slow down enough to draw, because accuracy is everything. A careful diagram in 70 seconds beats a confident misread in 40.