You're right about the initial cause, but I'm talking about the end effect rather than the cause. Some baseline number of teenagers experience mild depression. Most of them probably don't talk about it with anybody. Some of them choose coping mechanisms like alcohol or the internet, which cause their depression to become worse. Some of them don't do that, and either stay the same or get better. Much like teenagers who drink, teenagers who are spending all their time on the internet are possibly already depressed, and it may be constructive to them to spend less time on the internet. In the article's defense, the researchers admit that correlation doesn't equal causation, and that high internet use might be a symptom as much as it is a cause.
On the other hand, it is true that telling anyone to spend less time on the internet, without presenting alternatives, is probably as effective as telling someone to stop drinking without presenting any alternatives. The internet and drinking can both be pretty fun, and if they're the only fun thing in someone's life, they need to be replaced with something else, not just eliminated. That's harder to do, though, which is why you get articles like this that talk about the internet as the problem, but you don't get any articles talking about solutions.
So bad coping mechanisms cause severe depression in people who already had mild depression, but just because the underlying problem was already there, doesn't mean that the bad coping mechanism didn't make it worse.