They need to study the difference between familiar and unfamiliar words. They probably note this in the conclusion of the actual paper. I speculate that unambiguous and familiar words are recognized whole (in the same way a chess master recognizes the pattern of the whole board), and kids and adults finding a new or ambiguous word look at each letter and pause there to figure it out (like a chess novice looks at each piece and can only anticipate the next move). I hope this work isn't used as ammunition in the whole word vs phonetics holy war.