I suspect that the difference between messaging behavior and the minimum age setting is related to the fact that those settings are publicly available. That adds a signaling component to the game, and for 48-year-old straight males I wouldn’t be surprised if it turned out to be the dominant one: I don’t have any actual data here, but it seems likely that a middle-aged man setting the age of consent as his threshold sends a clear “dirty old man” signal that a 32yo threshold wouldn’t. Not a signal that a hypothetical dirty old man wants to send, I think; meanwhile, you can send messages to whoever you like, and large-scale messaging preferences are opaque to everyone but password holders. Actually, messaging someone below your nominal age limit might send a weak positive signal: “I like you enough to make an exception”.
The smaller of the effects discussed is probably genuine, though.
I suspect that the difference between messaging behavior and the minimum age setting is related to the fact that those settings are publicly available. That adds a signaling component to the game, and for 48-year-old straight males I wouldn’t be surprised if it turned out to be the dominant one: I don’t have any actual data here, but it seems likely that a middle-aged man setting the age of consent as his threshold sends a clear “dirty old man” signal that a 32yo threshold wouldn’t. Not a signal that a hypothetical dirty old man wants to send, I think; meanwhile, you can send messages to whoever you like, and large-scale messaging preferences are opaque to everyone but password holders. Actually, messaging someone below your nominal age limit might send a weak positive signal: “I like you enough to make an exception”.
The smaller of the effects discussed is probably genuine, though.
okCupid data. Of interest is the third graph.