You're right that the headline is cautious enough not to misrepresent the content of the study. I think controlling for parent and child educational attainment, which likely subsumes much of the possible confounders, is sufficient to claim a plausible causative effect of the presence of books.