The biggest I've seen in a published source in the wild is an 80-fold error in a reported distance, which I think came from a series of at least three unit conversions and area/length misinterpretations.