Why testing yourself beats rereading

Two students prepare for the same examination. One rereads the chapter four times. The other reads it once and then tries to recall it four times, badly, with long pauses. The second student will almost certainly do better, and will have felt worse throughout.

What everyone sees

Rereading is smooth. The text is familiar, comprehension feels high, and the sense of mastery grows with each pass. Self-testing is the opposite: halting, uncomfortable, full of gaps. The obvious inference is that the smooth method is working.

What is actually happening

Retrieval is not a read operation. Pulling something out of memory modifies the trace and strengthens the path back to it, which rereading does not do. Roediger and Karpicke ran the comparison directly: students who reread reported higher confidence and performed worse on a delayed test than students who practised recall. The gap widens as the delay lengthens — exactly the period that matters.

Why it stays hidden

Because the feedback signal is inverted. The method that produces the sensation of learning produces less of it, and the method that produces the sensation of struggling produces more. Nobody can feel their own retention curve, so almost everybody optimises for the feeling instead.