Moral language may help a social media post catch fire — but only up to a point. Mohammad Atari from UMass Amherst, in collaboration with researchers from Northwestern University and Chile’s Universidad del Desarrollo, analyzed more than 1.6 million posts about 13 political and social topics on Twitter, Reddit, and 8chan, finding that messages framed around right and wrong generally drew more replies or reposts, while engagement dropped when such language became too concentrated.
The study, published in Nature Human Behaviour, identified an average “sweet spot” near 30% on the researchers’ moral-density scale; simulations predicted substantially less engagement for posts on either side of that optimum, with the steepest penalty for overmoralized content. The pattern held across topics ranging from climate change and gun control to the #MeToo movement and Brexit. Still, the researchers cautioned that the observational study shows association, not cause, and does not account for factors such as a user’s follower count or posting history.