Screens are very very likely irrelevant, they did not control at all for content of the text. It's so incredibly logical that reading a full book gives better reading comprehension than reading tweets/reddit/tumblr/4chan to the point that this study is just one of the "duh" studies that may be necessary to say "yeah we already knew that but now it's confirmed"

Or dark mode adoption.