Survivorship Bias in Take Ecology

One of the staples of the social-media ecosystem of Takes is the “How I Broke Out of [Social Evil] by Learning [Critical Idea].” These tend to follow a common pattern of being miserably unhappy for reasons that couldn’t be articulated, then discovering [Critical Idea] through academic study or therapy or some linear combination of the two. At that point, eyes are opened, realizations are had, lives are transformed, and then thinkpieces are written. The clear implication is that if only everybody else would follow the same course, they, too would embrace [Critical Idea] and be transformed.

There’s a moderate amount of discussion about how these pieces are affected by social factors and social-media design– click-chasing and “echo chamber” effects, and the like. A thing I haven’t seen discussed as much, but have started to wonder about, is the degree to which these things are affected by a sort of survivorship bias.

That is, the academic/therapeutic activities that lead to the transformative insights in these kinds of stories seem like the kind of thing that naturally work best for a certain kind of person. That personality type shares a lot of traits with the sort of person who would be predisposed to write thinkpieces about their experiences. On the other hand, the kind of personality that would incline someone to find the relevant academic study or therapy an unconvincing waste of time seems likely to also make that person see the writing of “this was a waste of time” thinkpieces as a waste of time.

Of course, that sort of thing is more or less by definition an intractable problem– how do you learn the stories of people who by nature aren’t interested in sharing their stories? I’m not sure how you’d even figure out who those people are, let alone get them to talk.

In the end, I guess this lands in the umbrella category of “Reasons I’m Glad I Don’t Work in Social Sciences.” It’s a thing I wonder about every time one of these “Everybody MUST read this!” stories crosses my feeds, though.