Natural frequencies make Bayesian reasoning click

Correct: B — Natural frequencies (e.g., '8 out of 10 sick patients test positive; 95 out of 990 healthy ones also test positive')

Gigerenzer and Hoffrage (1995) showed that the notorious difficulty of Bayesian inference problems (e.g., 'given a positive mammogram, what is the probability of breast cancer?') largely vanishes when probabilities are reformatted as natural frequencies — counts of cases out of a reference population. With the same numerical content, accuracy on classic medical and legal Bayesian problems jumped from roughly 16% to 46% in their experiments. They argued evolution adapted human reasoning to count-like ecological data; modern probability formats translate poorly. The result has reshaped medical statistics teaching and risk communication.