One huge reason why effect sizes are small is that you're comparing the well-being of users vs. non-users, when we know that social media harms the well-being o

One huge reason why effect sizes are small is that you're comparing the well-being of users vs. non-users, when we know that social media harms the well-being of both groups. It's like estimating the health effects of automobile pollution by comparing the health of people with driver licenses against people who don't have driver licenses. Since automobile pollution harms both groups, you might not see large differences between the groups, even though pollution is very harmful to everyone's health.