Could someone link to the research in the first chart? I couldn't find a citation, and I'm not sure I trust simulations of children who can't read.