Jonathan Haidt: evaluation

Transcript context

…Old Americans also seem grumpier to me. Maybe that’s cable TV, but it’s not that they’re on their phones all the time. And you know all these studies. If you try to assess what percentage of the variation in happiness of young people is caused by smartphone usage — Sabine Hossenfelder had a recent video on this — those numbers are very, very, very small. That’s another measurement that seems to discriminate in favor of my theory, exogenous mood shifts, rather than your theory. Why not? First of all, we have to talk about correlation coefficients here. I’ve been involved in the debate. Jean Twenge and I are on one side. There are a few researchers — Andrew Przybylski, Amy Orben, Candice Odgers, there are a few others — on the other side. It’s been a very productive debate. It’s all conducted within bounds. There are three different kinds of studies: the correlational studies, the longitudinal studies, and then there are the true experiments. There are three different kinds. Zach and I and Jean have put them all — we have a Google doc. If you go to JonathanHaidt.com/reviews, we have all these Google docs. You can see abstracts to all the studies. Most of the discussion has been over the correlational studies. Some studies have said, oh, the correlation between time on screens, or time even on social media and time on screens, and some measure of bad mental health is like 0.04. Now, if that was true, I would say, wow, that’s pretty much nothing, but here’s the thing. Every time I dig into the data — whatever study is published to try to debunk this claim, you dig into the data — what do you find? If they’re looking at screen time overall, then the correlation is low, but whenever you can zoom in on social media, it’s bigger. Whenever you can zoom in on girls, it’s bigger, and an amazing thing has happened. The skeptics and me and Jean — we’ve all reached the same conclusion about the size of the correlation. They say the size of the correlation is between R is 0.1 and 0.15, which is small, but it’s about the same size as many other public health effects. But that’s always for boys and girls combined. What Jean and I have found is that when you look at the girls, now you’re in the realm of 0.15 to 0.2. Because girls are more affected by it, we actually all agree the correlation coefficient is around 0.17. That is not trivial. You would never let your kid do something four hours a day that was correlated 0. 17 with depression, anxiety, and self-harm. So, the effect is actually much bigger than many people realize in the scientific community because you have to zoom in on girls and social media. Why no talk in your book about the benefits of social media? They seem to me extremely large. I know many talented —…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.