From this thread: http://boards.straightdope.com/sdmb/showthread.php?s=&threadid=147958.
quote:
Originally posted by Triskadecamus
The concept we are searching for here is called “extinction” in behavior modification. Behavior has consequences. There is an informational channel in both directions, and one can modify the behavior of any subject by intelligently modifying ones own responses.
In the case of a behavior you want to have repeated more often, then it is most important that the response be very quick, and at least moderately reliable. That is called positive reinforcement. Even punishments immediately following behavior are positive reinforcement, because it is positive in schedule. The behavior happens, the response follows. In most cases, unless the punishment is severe, unavoidable, and distinctly undesirable to the original subject, positive schedule reinforcement is most effective in increasing the frequency of behavior.
Negative schedule reinforcement works like this. The stimulus is present at all times, until the behavior occurs, and it then stops, immediately, and reliably. Aversion stimuli are possible in a negative schedule program, and can be startlingly effective. It isn’t possible in our system, though, because a troll can avoid negative stimulus at will, simply by not reading the aversive sources. The negative schedule reward stimulus, in our case, is the existence of a large community of very intelligent and well-spoken people. The possibility of negative schedule reward is implicit in the function of the board, and looking at post counts of members will show you that that one is working just fine, thank you. We continue to post, and read, because the board provides us with our reward as long as we do so. When we log off, the reward stops. So, we log on again.
Then there is extinction. One must eliminate any cause effect between the stimulus and the behavior. That doesn’t mean you ignore anyone you think is a troll. It doesn’t mean you refuse to answer any question because it might be a troll. It means you don’t change your behavior because of the behavior of the other poster. If you find the question, or point useful, or interesting, you reply to the point. If you find it trivial, or stupid, you move on to something else. If it is a troll, you don’t change that. You don’t get into a fight, unless what you want to do is encourage fights. You don’t tell them that they are trolls or anything at all, unless you want to encourage trolls. If you answer an inflammatory question with a factual and rational reply, that is not changing your behavior and still falls into the definition of extinction. (Well, this assumes that factual and rational replies are usual for you.)
No one really just posts to post, without any thought of the reply. And certainly no one posts repeatedly to a forum that they don’t find some response from. The absence of stimulus cannot cause an increase in behavior, unless the stimulus itself is present under natural situations. In that case, removal of the stimulus is a response. No response at all causes extinction of behavior. (Remember, in this analysis, a response is any change in behavior by the general membership of the board.) Now, the phenomenon of the drive by troll is not going to go away. But, repeat visits will go down if the trolls find that nothing they do changes us at all.
So, answer the post, even if the poster is a troll, if, and only if, the post itself is something you wish to see repeated. (And that doesn’t mean answer only posts you agree with, unless you want a message board of clones who all agree with you.) If the post is one you think the board would be better without, then don’t reply. If the thread has other posts you find worth your time and effort, simply answer those posts, and continue your discussion. When posters you admire are posting less, or more than usual, post your approval of their posts, and they will likely post more often, in either of the cases. If you don’t admire the poster, then don’t post about them at all. Use the board’s ignore function, if your own ignore function is not operating correctly.
It works. Now some trolls are pretty stupid. But this method works on lab rats, invertebrates, and even graduate students.
Tris