As noted the eternal September effect is primarily caused by new-member with new-member interaction. Instead of taking cultural cues from established members, new-members take cultural queues from other new-members and learn incorrect cultural lessons. Mechanisms to prevent eternal September are to assimilate new members more rapidly and to dissuade new-members from posting as much until they have been assimilated (and especially dissuade them from influencing other new members). Filtering is only useful in that it retards the acquisition of new recruits slow enough to allow the old recruits to assimilate.
Assuming we’re in danger of an eternal September, the correct question to ask is not, “How do we filter better?” but “How do we convince new members to lurk until they’re assimilated?”
the eternal September effect is primarily caused by new-member with new-member interaction.
An obvious solution : Make the site appear, to new members, as if only (some desired fraction) of members are new.
Distinguish between “new” and “experienced” members. Let new members turn into experienced members when they meet some criteria, possibly post count, karma, or even votes by experienced members. Systematically prevent new members from interacting with too many other new members by simply not showing them the posts made by these other new members.
I’m not actually sure if I think this is a good idea, but it might be worth mentioning anyway.
Discussion threads would truncate for new users from new user comments (experienced user comments on new user comments would be invisible to new users). Our caching gets more complicated. Many candidate tests for “experienced” seem obvious, but some might be very easy to game (funny comments on HPMOR posts qualify you).
This does nothing to increase the capacity of older members to tolerate newbies—and that’s important, too. You’d be giving all the older members … how many times as many messages? I’m new, and I can’t keep up with my messages. I can’t imagine what it would do if I was an old member, and all of these new people were responding to me. If I were an old member in that situation, I would try to ignore the new users, and also, I would become increasingly annoyed with them demanding so much of my attention. That would lower the value of using the forum, and it may cause old members to quit.
It would also frustrate old members when new members weren’t aware of each other’s comments. That would be confusing.
Mechanisms to prevent eternal September are to assimilate new members more rapidly
There’s a limit to how fast this can be done. That’s, essentially, why something additional is needed.
dissuade new-members from posting as much until they have been assimilated
Deterring them from posting will ward off good people because they’ll lose momentum or be annoyed, and will increase the proportion of thick-skinned and / or persistent types who can deal with the annoyance. Not all thick-skinned / persistent people are bad, some are leaders or are gifted with those abilities, but creating gauntlets of annoyance will increase the proportion of undesirable thick-skinned / persistent types like trolls, newbie debaters, etc.
Essentially, dissuasion IS filtering, so if you’re going to filter, you may as well be conscious of it and use a method that is likely to attract the type that you want. My questionnaire would filter for people who like learning or don’t mind looking things up. The karma system currently in place filters for trolls and debaters. Dissuading people from posting will exacerbate the effect of the karma system if it remains as-is. The combination of the two would may result in a hideous unintentional filter.
Filtering is only useful in that it retards the acquisition of new recruits slow enough to allow the old recruits to assimilate.
If done well, it would also encourage a higher proportion of people that are the right type, discouraging mainstream people who aren’t genuinely interested in the culture from creating a new majority and taking over. Which is why i suggested the questionnaire that I did. That would select for people genuinely interested in rationality, most others won’t take the time to fill out such a questionnaire.
“How do we convince new members to lurk until they’re assimilated?”
I disagree, but it might be “How do we convince members to lurk until their assimilated without scaring any of them off
Filtering is not the answer.
As noted the eternal September effect is primarily caused by new-member with new-member interaction. Instead of taking cultural cues from established members, new-members take cultural queues from other new-members and learn incorrect cultural lessons. Mechanisms to prevent eternal September are to assimilate new members more rapidly and to dissuade new-members from posting as much until they have been assimilated (and especially dissuade them from influencing other new members). Filtering is only useful in that it retards the acquisition of new recruits slow enough to allow the old recruits to assimilate.
Assuming we’re in danger of an eternal September, the correct question to ask is not, “How do we filter better?” but “How do we convince new members to lurk until they’re assimilated?”
An obvious solution : Make the site appear, to new members, as if only (some desired fraction) of members are new.
Distinguish between “new” and “experienced” members. Let new members turn into experienced members when they meet some criteria, possibly post count, karma, or even votes by experienced members. Systematically prevent new members from interacting with too many other new members by simply not showing them the posts made by these other new members.
I’m not actually sure if I think this is a good idea, but it might be worth mentioning anyway.
This seems absurdly hard to implement
It seems not hard to implement naively.
Discussion threads would truncate for new users from new user comments (experienced user comments on new user comments would be invisible to new users).
Our caching gets more complicated.
Many candidate tests for “experienced” seem obvious, but some might be very easy to game (funny comments on HPMOR posts qualify you).
If this is done, posts upvoted past a threshold should also be visible to everyone.
This does nothing to increase the capacity of older members to tolerate newbies—and that’s important, too. You’d be giving all the older members … how many times as many messages? I’m new, and I can’t keep up with my messages. I can’t imagine what it would do if I was an old member, and all of these new people were responding to me. If I were an old member in that situation, I would try to ignore the new users, and also, I would become increasingly annoyed with them demanding so much of my attention. That would lower the value of using the forum, and it may cause old members to quit.
It would also frustrate old members when new members weren’t aware of each other’s comments. That would be confusing.
Do you see a way to resolve these issues?
There’s a limit to how fast this can be done. That’s, essentially, why something additional is needed.
Deterring them from posting will ward off good people because they’ll lose momentum or be annoyed, and will increase the proportion of thick-skinned and / or persistent types who can deal with the annoyance. Not all thick-skinned / persistent people are bad, some are leaders or are gifted with those abilities, but creating gauntlets of annoyance will increase the proportion of undesirable thick-skinned / persistent types like trolls, newbie debaters, etc.
Essentially, dissuasion IS filtering, so if you’re going to filter, you may as well be conscious of it and use a method that is likely to attract the type that you want. My questionnaire would filter for people who like learning or don’t mind looking things up. The karma system currently in place filters for trolls and debaters. Dissuading people from posting will exacerbate the effect of the karma system if it remains as-is. The combination of the two would may result in a hideous unintentional filter.
If done well, it would also encourage a higher proportion of people that are the right type, discouraging mainstream people who aren’t genuinely interested in the culture from creating a new majority and taking over. Which is why i suggested the questionnaire that I did. That would select for people genuinely interested in rationality, most others won’t take the time to fill out such a questionnaire.
I disagree, but it might be “How do we convince members to lurk until their assimilated without scaring any of them off