Reliability of extinction risk assessment: revisiting the (0,1) CI debate