While wrong by a statistically significant amount, Millikan's initial value was still much less than 1% off. It's really not a good example of the scientific consensus being completely wrong, because there was no consensus until Millikan's number had been shown to be wrong.