What makes a quiz question actually test understanding

A good quiz does more than check whether you recognise a fact. Decades of research on the testing effect (Roediger & Karpicke, 2006) show that the act of retrieving an answer from memory strengthens it far more than re-reading ever could. But that only works when the question forces real retrieval rather than recognition. The difference between a quiz that helps and a quiz that flatters you is in how the questions are written.

The four question types worth generating

• Recall — "State Newton's second law." Tests whether the fact is retrievable at all. Necessary, but the weakest on its own.
• Application — "A 2 kg block accelerates at 3 m/s². What net force acts on it?" Forces you to use the concept, not just name it.
• Comparison — "How does an exothermic reaction differ from an endothermic one in terms of enthalpy?" Reveals whether two ideas are genuinely distinguished or just memorised as labels.
• Why/mechanism — "Why does increasing temperature speed up a reaction?" The hardest to fake; it exposes shallow memorisation immediately.

Common mistakes that make a quiz useless

• Give-away distractors. If three of four options are obviously wrong, you are testing reading speed, not knowledge.
• Verbatim lifting. A question copied word-for-word from the source tests whether you spotted that sentence, not whether you understood it.
• All-recall, no-application. A quiz that never asks you touse a concept leaves you fluent in definitions and helpless in the exam.

NoteTube's quiz generator is tuned for this: it pulls testable claims from your material, varies question type across recall, application, comparison and mechanism, writes plausible distractors, and explains every answer so a wrong response becomes a learning moment instead of a dead end.