As with all things psychological, there's a lot of error variance in the data, which is why it's really hard to find correlations between anything in social science research. The fact that we do find it with social media usage makes it stronger in my view. Many other "ills" that we believe to negatively affect people often wash out in the data.