Correlations, of course, are also not causation. For example, there is no control in this study which rules out an effect like "some men have less testosterone, and this causes them to both be more glued to their phones (using while charging), more social on their phones (talking for over an hour a day), and to have a higher incidence of abnormal sperm." It's sort of like a headline: "If you're shorter, you may be a better parent: a study found an inverse correlation between height and how much time you spend with your kids." You'd be tempted to disbelieve me until I tell you that on average, women are shorter and also spend more time with their kids. If in this case we know the causes, how much more difficult is it when we don't to isolate the confounding variables?