It's not a bad paper actually, but I suspect what will eventually come out of it is that it's evidence of a psychological quirk - that the people remembering words are drawn through some psychological/cultural bias to remember/predict the words that the person randomly choosing later will pick.
(I say this, mind you, as a guy who, when once tested with Zener cards, consistently identified not the card being "transmitted" but the *next* card in the pack, through a whole pack... But if that was a sign of reading the future, how come my lottery picks never come up?)