Overflow in science and its implications for trust