"Don't Feed the Trolls" Is Bad Science

You’ve seen it. You see it when you point out antisocial behavior, more often when that antisocial behavior involves racism or sexism or something else people are uncomfortable talking about:

“Don’t feed the trolls. Attention is what they want. Ignore them and they’ll go away.”

It’s nearly as strong a silencing tactic as “Oh, they’re just anonymous adolescent boys [which they’re frequently not] so you can’t expect any different.” It is also just as wrong, something that started to come into the general consciousness of the internet with the #mencallmethings campaign last year. “Ignoring” the trolls, dealing with them on your own without social support, doesn’t make them stop. It makes you stop.

It’s not surprising the advice about trolls is wrong if you look at where it came from. The idea is based in the principles of operant conditioning: reward and punishment having an effect on the frequency of behavior, rats pressing levers for sugar, pigeons pecking at buttons for food, etc. The advice itself is essentially what was once given to parents of small children who act out. Don’t give them the attention, and they’ll find better ways to get it.

Do something you know is calming for yourself. Many experts advise you to ignore your child. You child’s behavior is to obtain a response from you. The calmer you are, the easier to thwart them.

"Tantrum Over a Movie Selection" by Bethany L King

The problems with applying that to trolls are at least three-fold. The first two problems come from the fact that we are not trolls’ parents, and these trolls are not small children.

This means we don’t have control over their environments. These trolls have buddies egging them on. If they’re on a Twitter hashtag, they see other people doing the same thing (reward). If they’re on a forum, they’ve driven away everyone who isn’t going to tell them how hilarious their ability to swear at someone is (reward). If they’re on 4chan or some sub-Reddits or other sites that self-selects for proud anarchy, well… (reward).

Nor is this limited to the internet. People aren’t stupid. They notice when someone is an “acceptable” target of vitriol in the real world. We’re a social species with a long adolescence devoted to figuring this sort of thing out, and despite knee-jerk denialist commenters on every post and article dealing with the topic, this stuff isn’t terribly subtle. Even offline, this sort of behavior–if perhaps not always this degree of behavior–is rewarded.

When trolls see no reaction from the rest of the world, there is literally nothing in this equation but reward. Who’s going to go away under those circumstances? Who will extinguish bad behavior when that behavior means they’re winning?

Not having control of their environments also means we usually don’t have opportunities to interact with the trolls when they’re exhibiting prosocial behavior. We don’t see them. We don’t reward them. The only chances we have to affect their behavior come during the period in which they engage with us. If we pass on that opportunity, we have no impact at all, except for the tacit acceptance that silence provides.

Yes, silence is a reaction, which is the third problem with the feeding metaphor. Not only is it behavior, but it is highly rewarding behavior to trolls. The important thing to remember about trolls is their purpose. It isn’t to get attention for themselves per se. It is to control the conversation. Sure, some trolls show up out of the blue to say, “Derp, I’m an asshole”, and wait for the show to start, but that’s very rare.

The majority of internet trolls show up when something contentious is being discussed or when someone contentious (like a non-white, non-male, non-heterosexual, non-capitalist…I think you get the idea) is doing the writing. If nothing happens when they attack the writer, except that their targets get a little quieter, once again, they are only getting what they want.

Combine these three factors, and unless we speak up and say, “This is not the behavior I want anywhere near me,” all the trolls get is reward for their behavior. That doesn’t mean there isn’t a discussion to be had about how to deal with trolls without sacrificing the conversation. It does, however, mean that it’s very much time to stop telling us to ignore the trolls.

{advertisement}
"Don't Feed the Trolls" Is Bad Science
{advertisement}

70 thoughts on “"Don't Feed the Trolls" Is Bad Science

  1. 1

    Going back to your (counter-)analogous example: is “ignore them” good advice when the “troll” is in fact a small child misbehaving within range of his/her parent?

    I sort of need to know this soon….

  2. 2

    … Huh. Never thought of it that way. Actually, somehow I never thought of persistent online bigots as trolls. I find that my working mental definition of “troll” is basically “ze doesn’t really believe that, ze’s just saying it to make you mad.” Which doesn’t quite seem to match the usage here. There’s something funky going on with definitions, that I’ll have to think about.

  3. 4

    Huh? “Operant conditioning” is just a name for it. It says nothing about whether the practice itself is bullshit. (“The advice itself is essentially what was once given to parents of small children who act out” sort of implies it might be.)

  4. 6

    Rieux, like a lot having to do with operant conditioning, things are quite a bit more complicated these days. I’m soooo not an expert, but from what I understand, the idea is to acknowledge that the child has a concern while acknowledging the tantrum itself as little as possible. So it’s a balancing act of ignoring the outburst without ignoring the child.

    Of course, that’s advice that depends on the child being at least somewhat verbal, so you have a little time yet. Up until then, you’ll need to rely on mind reading, I’m afraid. Or get your advice from experts.

  5. 7

    Rieux, this is a digression, so I’ll cease after this.

    If a small child misbehaves within operative range, doing nothing about it is hardly likely to help; the question is obviously what to do.

    (Ignoring the child (in the sense of withdrawing attentiveness) is not doing nothing; that advice is clearly predicated on the behaviour being attention seeking and consequently said ignoring being a negative reinforcement, and therefore fails due to inapplicability if the misbehaviour ain’t due to that motivation)

  6. 8

    Robert, the “not believing what they’re saying” definition is widely used. It is also inconsistent with the practical definition used by people who tell us to ignore the persistent online bigots. I agree that there’s something funky going on with definitions, but it didn’t originate with this post.

  7. 10

    Ignoring trolls rarely works. I generally don’t like an emotional swear laden attack either, that also tends to reward the trolls as well. A well balanced, rational response that may or may not have swears in it may not stop the troll but you may get enough people to join in to drown the troll out.

  8. 12

    There are more ways to troll than to just say contrarian things to stir up controversy. It’s also trolling to derail threads to harp on one’s own agenda.

    But to the point in the main article, maybe one of the key questions here is how to engage trolls and not whether to do so. It’s different to let a troll drag you into a war of attrition about what the troll wants to talk about. It’s another thing to addressing the trolling itself and not debate their arguments’ merits when they’re not proving to be people who argue in good faith.

  9. 13

    I think it’s pretty self-evident that one does not have to be misrepresenting their position to be a troll. Most trolls I run into believe their beliefs quite passionately.

  10. 14

    I’m not understanding something here:

    (1) It’s an awesomely catchy title (it made me click, right?), but I don’t see any science here to back up the contention that “not feeding the trolls is bad science.”

    (2) I’m not understanding how ignoring them is letting them “win.” My goal on any forum is to engage in a conversation with whoever is really trying to communicate. Utterly ignoring them and continuing the convo with the “real” participants seems most sensible to me. As soon as you react to trolls as if they merit a reply, they’ve hijacked the convo with their silly drama, and the real convo dies.

    Please explain, “When trolls see no reaction from the rest of the world, there is literally nothing in this equation but reward.” When the conversation proceeds as it was, as if the silly / immature / overdramatic person hadn’t spoken, I don’t see that as a “win” for the troll, and it’s hard to imagine the troll experiencing it as a win.

    In other words, when it’s practiced, my experience has been that “not feeding the trolls” actually works. I’m not seeing any data in this article that suggest otherwise.

    (3) Please explain how this is true: “The first two problems come from the fact that we are not trolls’ parents, and these trolls are not small children.” Operant conditioning is not some outdated parenting technique — these patterns of behavior were first noticed completely outside of any family dynamics (and replicated with all sorts of animals, including humans, in all sorts of non-family situations, with all sorts of rewards, both tangible (e.g., food, money, drugs…) and non-tangible (eye contact, smiling, attention…).

    Yeah, it works on rats, pigeons, and humans of all sizes, including small kids who attempt to get what they want through disruptive behavior — and although it’s a non-scientific sample size of two, I’ll add that my kids were never whiners / tantrum-throwers / picky eaters when they were younger 😉 — and I don’t see any evidence presented that operant conditioning wouldn’t work to extinguish other disruptive behaviors (such as trolling) either.

    Of course, once anyone takes the troll’s bait (“I can’t believe you would say that! My stars and garters! X! Y! and Z!”), everyone else’s efforts to extinguish the behavior fail, because someone has put the troll on a variable interval schedule of reinforcement (as if you gave the shopping-cart-candy-whiner what he wanted “just once”), making the behavior very hard to extinguish. The kid (or the troll) learns that these strategies can sometimes be effective tools — tools to be wielded obnoxiously and often.

  11. 15

    With a small child’s misbehavior, the thing to do is immediately remove him/her from the center of attention (restaurant, store, whatever) and wait with her/him while stressing that he/she will not rejoin society until that behavior has stopped.

    As for trolls, remember that intermittent reward is STRONGER than consistent reward. So banishment/punishment must be consistent and unhesitating.

  12. 17

    (1) It is based on a poor understanding of operant conditioning.

    (2) I suggest some background reading on #mencallmethings and the isolation experienced by the people trolled.

    (3) You may have missed the part where I explained that we do not have control over the environment of the trolls.

  13. 18

    Randomfactor, do you have an example of a real-life open internet forum where the control to provide “consistent and unhesitating” punishment is maintained and the conversation is still healthy?

  14. 20

    @ 8 Stephanie Zvan said:

    I agree that there’s something funky going on with definitions, but it didn’t originate with this post.

    I’m sorry if I implied that it did, that wasn’t what I was thinking.

    I think maybe the word “troll” is based on some broad categories of behavior (saying things that make people angry) but that there can be a lot of different motives that lead to that kind of behavior. If that’s right, it would mean that any psychological strategy designed for one kind of troll is likely to fail against another kind. Trolls who want to control the argument or silence certain people are not going to respond the same way as trolls who just want to see the pretty lights as the flame wars ignite, who in turn are going to be different than the ones who are doing Poe-ish parody / performance art. And then, psychology being what it is, you’d inevitably get trolls with multiple motives…

    Of course, none of that invalidates real-world experience with what works and what doesn’t against a specific kind of attack, which is what you’re talking about here. I’m on a tangent, really; you sparked my curiosity in an odd direction.

  15. 21

    Operant conditioning depends on providing an appropriate motivation or disincentive to alter behaviour. That is what some people are missing, and that is the thrust of what Stephanie is trying to say.
    In many cases, ignoring a troll in an effort to provide a disincentive to continue the behaviour is as pointless as offering a Rubik’s Cube as a reward to a button pushing pigeon.
    If engagement is not what they are looking for, then no amount of avoiding them is going to serve as a deterrent.
    How I handle trolls is to find their motivation. Are they trying to get the last word? Are they trying to showcase their brilliance?
    Those kinds of trolls are not going to care if you ignore them. In some cases that is what they are looking for. So you either engage them or ban them. I’ve actually told a troll that I have all the time in the world to respond to his every comment- and he left two comments later.

    The same goes for kids. Ignoring poor behaviour only works if the motivation is attention. In many cases those who suggest you do that forget to mention that you are likely working against your own previous conditioning of rewarding bad behaviour. I rarely ignore a behaviour with my own kids, there is always some action involved- and if I do isolate them I do it after explaining the reasons.

  16. 22

    I spent a dozen years as a behavior analyst using operant conditioning to train humans. I’m about as much of an expert at using the accepted techniques of applied behavior analysis as one can be without an advanced degree in it.

    There are several problems with the ignore the trolls formulation as others have pointed out. “Attention” isn’t always the reinforcer for trollish behavior. So unless you’ve done a functional analysis of the posting behavior of the troll in question simply ignoring the person’s posts may or may not reduce the frequency of the trollish posting behavior.

    Ignoring behavior is generally known as putting the behavior in question on an extinction schedule of reinforcement. Behavior with a history of reinforcement when initially put on an extinction schedule typically exhibits an “extinction burst” where the behavior actually significantly increases in frequency for a short period before “extinguishing.”

    A generally sure way to tell if you have correctly identified what is reinforcing a behavior is to stop doing the thing you think is the reinforcer. If you get an extinction burst then yeah, you’ve probably got a good handle on what is reinforcing the behavior and can continue to not reinforce. If the behavior continues at the same rate then you probably don’t have the primary reinforcer identified.

    The post makes a good point about controlling the environment. Unless you are able to make sure there are no reinforcers delivered EVER once the decision is made to put a behavior on an extinction schedule then you are pretty much screwed due to the way schedules of reinforcement work. In general, when you are shaping a new behavior you reinforce every instance of the behavior. When the behavior is well established, you put it on a variable schedule – typically a variable ratio sort of a thing where every (N + y) instance of the behavior (N = a specific number and y = a chosen range to vary so N could equal 10 and y could be a randomly selected number in the range of 1 to 100) gets reinforced. Variable schedules are known as “maintenance” schedules because they keep behavior going.

    So, unless you can be sure that all comments on a blog will “ignore the troll.” you are pretty much guaranteed that the trollish behavior is on a maintenance schedule.

    As a side note, failure to have control over environments, rather than the actual science of behavior describing those environments, is generally what makes people say that “Operant conditioning doesn’t work!”

    Sorry for the tl;dr post.

  17. 23

    Stephanie, while you’re right that ignoring trolls isn’t going to simply make them go away, I think you’re missing the broader picture as it applies to blogs. Troll-feeding results in massive derailment and hundreds of posts (mostly not by the troll), often spawning sub-arguments amongst other posters, instead of what would otherwise have been a normal discussion with a few idiotic non sequitur comments tossed in. It’s why I’ve pretty much given up reading Pharyngula – there’s a core group of about 10 people with a lot of time on their hands who feel the need to respond overwhelmingly to every stupid bigot who crawls out of the woodwork (or for that matter any commenter who makes a somewhat inappropriate remark).

  18. 24

    Going back to your (counter-)analogous example: is “ignore them” good advice when the “troll” is in fact a small child misbehaving within range of his/her parent?

    I sort of need to know this soon….

    Rieux, all I got is anecdote, personal experience, but for what it’s worth:

    My mother was a school teacher. In the 60’s, when I was a small child and this sort of thing was fairly new, she used positive reinforcement on me; rather blindly praising behavior she approved and ignoring behavior she disapproved of.

    I’m 53 years old. I still have nightmares. Not kidding.

    Seriously. Don’t assume you know why your kid is acting out, and don’t make the kid feel like hir feelings are unimportant. An honest, empathic, “I understand you feel [whatever], but this is what I think is best right now” is better for all concerned.

    Ignoring shit does NOT make it go away. And ignoring small children’s suffering is never a good idea (don’t assume they’re “just doing it for the attention”).

  19. 26

    It’s why I’ve pretty much given up reading Pharyngula – there’s a core group of about 10 people with a lot of time on their hands who feel the need to respond overwhelmingly to every stupid bigot who crawls out of the woodwork

    Well, I disagree. It’s one of the reasons I read Pharyngula, including the comment section–I know that people will engage the trolls, and they will do so with wit and (usually) reason.

    And, all things considered, I think Pharyngula has fewer trolls than might reasonably be expected, considering PZ’s subject matter–precisely because trolls are dealt with so firmly and mercilessly there.

  20. 27

    There is a huge difference between ignoring a person and ignoring the behavior of a person. It is quite possible to interact with a person while completely ignoring aspects of they way they are behaving at the time. It isn’t easy. It takes some training. A parent who completely ignores a kid who is having a “tantrum.” is doing the kid, and everyone else, a disservice. It is quite possible to continue to interact with a child who is “behaving badly” and not reinforce the bad behavior, while at the same time getting the kid back under some measure of operant control.

    Some people are just good at it at a sort of intuitive level. In general they make pretty good parents and teachers because they can cope with the stresses of dealing with bad behavior without escalating the situation.

  21. 28

    [Semi-OT]

    Midnight Rambler:

    It’s why I’ve pretty much given up reading Pharyngula – there’s a core group of about 10 people with a lot of time on their hands who feel the need to respond overwhelmingly to every stupid bigot who crawls out of the woodwork (or for that matter any commenter who makes a somewhat inappropriate remark).

    Which merely shows you didn’t belong there — troll-bashing is Pharyngula‘s signature hobby.

    (You were an OK commenter, IIRC, FWIW)

  22. 29

    @ Stacy,
    Yeah, parents have hugely significant control over the environment in which their children live. A truism of behavior analysis is “control the environment and you have control of the people in it.” If you do a shitty job controlling the environment, the people in it are in for a shitty time.

    Another thing is that coercive, punishing environments generally produce a significant amount of counter control – the people being controlled will do pretty much anything in their power to make it difficult for the people in charge.

  23. 30

    bybelknap, there’s a big difference between being a parent and being a teacher. Maybe what works well in the classroom could be abusive if applied in the home?

    I don’t think “controlling the environment” and “controlling the people in it” is a laudable goal. Shit, it sounds totalitarian to me.

    And if behaving in a way which is disruptive or inconvenient (to the parent) is always interpreted as “bad behavior” in the kid’s own home–and ignored–that’s a seriously controlling environment.

  24. 31

    Which merely shows you didn’t belong there — troll-bashing is Pharyngula‘s signature hobby.

    Mmm…it seems like it’s gotten a lot worse in the last year or so. It went from more arguing with goddists to simply hurling insults back and forth with reactionary morons, and often between each other. To use an analogy from the Usenet days, it was like going from talk.origins to alt.tasteless.

  25. 32

    Stacy, If it is abusive in the home it is abusive in the classroom. Abuse is abuse is abuse. I am not advocating a controlling environment. I am describing the control that parents have over the environment in which their children live. Parents do, in fact, control their children by controlling the environment in which they grow. One hopes that they use the enormous control they have for the good of their children and society in general.

    In my case, despite a fairly shitty job by my parents, I turned out not to be a serial killer or rapist or evil-doer of consequence. I think my kids will probably be fairly well-adjusted with only a few years of therapy needed to undo all the damage my wife and I have inflicted.

    So yeah, when you have the power of a parent having a huge amount of empathy is pretty important. If a kid is throwing a tantrum in the supermarket because they want a candy bar in the checkout line you can ignore the crying and whining and foot stomping, while addressing the want, and explaining why that want will not be fulfilled at the moment – “I understand you want a Snickers. We will be having dinner in 30 minutes, and a Snickers will make you full and not want dinner.” Very often there isn’t any need to make mention of the emotional aspects of the child’s outburst. They are not the core issue. The core issue is the nutritional health of the child.

    I hope this helps to illustrate what I mean by ignoring a behavior. Hell, in this case, since the antecedents are most likely hunger and previous success at getting sweets in the checkout line the behavior probably isn’t “for attention.” and can be addressed – “Crying and foot stomping will not get you a Snickers. At this point there isn’t a Snickers in your immediate future so I suggest you knock off the outburst.”

    “Ignoring” behavior is only useful when the behavior in question is “for attention.” In general “ignoring” is a really shitty thing to do to someone – and note that here I am saying “someone” and not “a behavior” because unless one has some fairly decent training what happens is the person in charge hears “ignore the behavior” and they just ignore the person or do a half-arsed job of ignoring the behavior.

    I would recommend reading Skinner’s “Beyond Freedom and Dignity” to get some insight into the whole control of the environment thing. Agency and Free Will are illusions. We are under the control of our environments. Who is pulling the strings? Who watches the watchers?

  26. 33

    Silencing happens to readers too. I don’t read comments on certain blogs because I know the trolls are sure to be there in force. I feel I have to read all comments in order to avoid being repetitious or missing explanations that would affect my comment. But reading through hours of flame wars takes too long, so I just go away wishing I could have joined in. Anyone remember the endless demands for explanation of just why women feel uncomfortable with being harrassed in elevators? (Yeah, I know. Let’s not start that up again.)

    By way of contrast, I read all comments on certian other posts because of the wit and reason demonstrated by commenters. On blogs like these, there is fierce debate but only a few trolls (who go on for pages. Why is it that troll comments are so long-winded – sorta like this one?)

    In other words, trolls chase me away, so I appreciate a strict banning policy. On the other hand, I think all blog readers should support and defend bloggers who are being harrassed – unless we wish to be considered part of the problem. Conundrum.

  27. 34

    Crap, have I derailed this thing? I honestly didn’t mean to. Just to get back on track, “ignore the trolls.” is a bad idea because it pretty much can’t work in an open blog environment. It is difficult if not impossible to identify the reinforcers for particular trollish behavior. Absent identification of the reinforcers it is at best a crap shoot to know if ignoring the troll will have any effect on the trollish behavior. Just posting a comment may be inherently reinforcing. No response required.

  28. 35

    belknap, thank you for the thoughtful response. And Stephanie, I apologize for derailing, too.

    It is difficult if not impossible to identify the reinforcers for particular trollish behavior.

    This. It may work on some of the commenters lumped together under the rubric “troll”, but it won’t work on all. Bullies and fanatics can’t be counted on to slip quietly away because nobody responds to them.

  29. 36

    Stephanie, first time commenting on your blog here at FTB, but have enjoyed your postings since you joined the group.

    I don’t know about whether not feeding the trolls is bad science, not being a scientist myself, but it has never felt right. I often respond to egregious trolls with what I hope is polite but sarcastic and rational replies. They rarely get it, but it often prompts them to respond to me as opposed to the rest of the thread, and I could care less what they might think of me. They are, after all trolls, and, in a sense, so am I.

    I do think that we need to call “BS” when we see it, and to attack fiercely those who threaten other respected, even if sometimes disagreed with, members of our community. I remember being appalled at the responses RW got over the “elavatorgate” incident. It more than adequately emphasized that sexism was a HUGE problem in the atheist/humanist/skeptical community, and one that needed to be addressed. That sort of behavior had been ignored FOREVER and it has not lessened its offensiveness, nor its ability to do harm.

    It is also one of the reasons that I use my real name when commenting, or connect my real name to any “usernames” I adopted before realizing that I’m 60 years old and proud to stand by any of my thoughts, beliefs, and/or criticisms that I’m willing to send out to the internet.

  30. 37

    I don’t think the problem is so much with the “don’t feed the trolls” argument per se, but rather that people are too quick to label all anonymous internet doucebags as “trolls”. MRAs are not trolls. Creationists are not trolls. White supremacists are not trolls. Trolls are people who do not believe in a single stupid thing they say, they just laugh at the righteous indignation they inspire in others. They are people that get off on getting under people’s skin without a second thought to the content of what they are saying.

    I agree that true trolls are few and far between, so another apporach might be better at the cost of engaging a couple of the “real” trolls.

  31. 38

    Y’know, this is why I like blogs like Pharyngula, for example. In there, trolls are given the Internet equivalent of tar and feathers, and then banned, with their stupid behavior forever inscribed in a “dungeon” for all to see.

  32. 39

    Rieux, I seriously recommend Alyson Schaefer’s “Honey, I wrecked the kids.
    The thing about not giving attention is that you have to be sure that attention is the thing they want.
    If your kid empties the glass of milk on the floor and plays in it because they actually want to play in a puddle of milk, ignoring isn’t the right thing to do.
    If they spill the milk on the floor one drop after another so you can react, they want your attention (remove glass without saying a word, hand them a sponge and tell them that you can read y book together once they finished).

    That’s probably the mistake people make with trolls, too: They think they want attention when actually they want to shut you up.
    If you ignore the bully in meat space who’s picking on the small kid, the bully gets away with it. They don’t want attention, they want to pick on small kids.

  33. 40

    Bybelknap–excellent contributions. Just for the sake of those who might want to follow up on your excellent advice, essentially what you are lobbying for is a “functional analysis” of behavior. Just as Stacy has described, the too-simple “ignore the bad stuff; attend to the good stuff” is only useful if attention (particularly, *your* attention) is the reinforcer. Often, of course, it is not, but the only way to know for certain is to (as you have said) manipulate the contingent attention to see if you get a change (often, but not always, the extinction burst you describe–it is worth noting that without an understanding of the underlying principles, many parents and teachers misread the extinction burst as a sign they did the wrong thing).

    “Don’t feed the trolls” is not *necessarily* bad science; the advice is there presumably because it has occasionally worked. If indeed *your* attention is the reinforcer (or “a” reinforcer), withholding it contingent on trolling could work.

    But.

    As per Stacy’s testimony, a simplistic attempt at control ignores the multiple contingencies in a real-world environment. With internet trolls (and how reinforcing it can be to respond to them!), we are unlikely to be able to eliminate their feeding, even if we know that it is attention they are after. In a worst-case scenario, we simply put them on a leaner schedule of reinforcement, making them A) work harder for their payments, and B) far more resistant to having that behavior extinguished. We’re essentially putting them on the schedule that built Las Vegas.

    So.

    What can be done, though, is to take advantage of more operant conditioning (particularly, the matching law), and systematically reinforce the behaviors you *do* want to see. If 90% of what a troll says is worthless, responding to the 10% appropriately might (see above regarding functional analysis; it is an empirical question whether your responding is a reinforcer) increase the appropriate behavior. We’ve actually seen it work on Pharyngula, of course, where (a small number of) former trolls are valued contributors.

    I’d recommend another very user-friendly book (in addition to your BFD recommendation), Karen Pryor’s “Don’t Shoot the Dog”.

  34. 41

    George W at #21 and bybelknap at various:

    Thanks so much for those posts, those were very informative and helpful to me in thinking about this phenomenon. I grew up under the “don’t feed” model (which I’m thinking now was probably reinforced by my parents, and it’s only now in my life that I’m starting to realize that, though I love my parents, they don’t know much about human behavior studies).

    Is trolling mostly an internet phenomenon? From the descriptions in this thread it sounds like trolling is a form of bullying, or am I misunderstanding?

    I wonder if trolls ever get trolled, and if so, how they react? Given what the thread has discussed, the answer is probably complicated because of the variety of motivations behind trolling.

    Still learning,

    Robert

  35. 42

    @Stacy,

    Seriously. Don’t assume you know why your kid is acting out, and don’t make the kid feel like hir feelings are unimportant. An honest, empathic, “I understand you feel [whatever], but this is what I think is best right now” is better for all concerned.

    Ignoring shit does NOT make it go away. And ignoring small children’s suffering is never a good idea (don’t assume they’re “just doing it for the attention”).

    This. Especially your second paragraph.

    I’ve done a lot of babysitting/nannying in my time and it can be very difficult to determine what a child wants. To make it worse, most young kids aren’t good at expressing their needs (or even knowing what they need) which makes it difficult for the adults caring for them to know what they want/need. So, the kid gets upset and throws a tantrum.

    Sometimes, it’s short lived and easy to deal with, but sometimes it’s not and ignoring it is rarely a good idea. Additionally, it’s almost never done just to get attention. The kid is usually upset because his/her needs aren’t being met or he/she is having expressing him/herself. It’s infuriating when parents and other adults assume a young child is acting out solely for attention or to annoy the adults. Grrr.

  36. 43

    Cuttlefish,
    Thank you for your response and yeah, Karen Pryor is a hell of a lot more readable than Skinner.
    A functional analysis was the essential first step in putting together a case file when we’d get a new kid because we’d have to present the case to a human rights committee within 30 days of intake. We worked with kids who everyone else had given up on. We worked with severe self injury, aggression, pica, and other behavior that had been shaped up by well meaning but inept care providers. We had a pretty good success rate. By success I mean that we got the kids stable in our environment. Very few would I trust out in the real world. Generalization to open environments is a bitch. I only lasted a dozen years.

  37. 45

    Rieux –

    Ignoring is an excellent idea, if, as Gilliel mentions, you are sure that attention seeking is the motivation for the behavior you might ignore. It is also a good idea to consider that ignoring doesn’t mean you have to leave it at that. When a child is throwing a fit, for example, you might ignore the fit and address the issue after they have calmed down. This is something that I have done when one of my kids has thrown a fit in public – forcing me to remove my child from that location. This hasn’t happened more than a handful of times.

    I remove my child – generally putting them in the van and standing outside (I often take that moment to have part of a cigarette). When I am done and the fit is good and over, I will then explain briefly what privilege that child has lost and that the loss of that privilege was due to my having to stop what I was doing. I don’t say more than “you have just lost tee vee and dessert for the day” (or whatever) and “you lost them because I had to stop shopping.” They might renew their fit, ask why, assert they hate living with me, claim I am a mean, terrible person – that I ignore completely, unless feeling snarky/wanting to disarm I ask “who’s the meanest papa ever?” – which I can in turn follow up with “and who’s the best papa ever?” I never do that without having complete silence and really only do so because (due to how I generally use that device) it tends to diffuse any lingering anger – both mine and that of the child in question.

    I would make it clear that ignoring works exceptionally well – even in cases where you absolutely *must* have compliance (ie. running out into the road, playing with outlets and the like – remove and provide consequences after the fact), but you are unlikely to use it perfectly – all the time. Consistency is best, but the only parents who are absolutely consistent have earned themselves a place with other mythical beasts – they *might* exist, but it is exceedingly unlikely. I work very hard to be a good parent. My kids have a lot of fucking bullshit to deal with and I don’t want to contribute to that with crappy parenting. But I still sometimes have minor meltdowns myself – rarely, but sometimes complete with yelling. Instructing my children to walk away and ignore me when that happens has gone a long ways towards helping them understand how this all works.

    I would also note that there are often extenuating circumstances that need to be dealt with. In my family it is related to the mother of my children leaving and some of the lingering problems that fosters. I just snuggled my four year old who was in the midst of a crying fit. Not a fit over any one thing – he was upset because of a number of things that combined to make him genuinely sad. But rather than throwing a tantrum, he was just sitting, crying for momma – not crying for “momma” that is me, but for the momma he no longer has. When he is crying for me momma, I ignore and offer attention when he asks appropriately. When he is crying for momma momma, I generally will give him a snuggle – but only because he very rarely does so and because when he really wants his actual momma, he is dealing with a very different level of hurt – one that is probably very confusing. He hasn’t seen his momma in almost eighteen months and doesn’t recognize pictures of her anymore.

    It isn’t easy to work out what is actually going on sometimes, so it’s really important to recognize that you are going to screw it up sometimes and doing so isn’t the end of the world.

  38. 46

    I try (and regularly fail) to ignore the posters who are obviously going for the off-topic derail. To include tone trolls. But I think its a good idea to respond to the more on-topic trolls. Doing the latter can help you clarify your argument. Some of the best explanations (or expansions of a person’s point) that I’ve read have come from such troll responses.

    For the tone trolls and derailers, let their posts sink into ignominity. Possibly the best response to them is to post in response to someone else’s more on-topic post. You want to prevent us from discussing the topic at hand? Then I will frakking fill the blog with posts on the topic at hand.

  39. 47

    I work in child care and ignoring negative attention-seeking behaviour followed with attention given for positive behaviour works like magic when the implementation is perfect. The problem is that it is extremely difficult to implement perfectly. Every adult has to react the same way and, to a lesser extent, so do all of the other children. Even a glance at the child or frustrated sigh puts you back at square one with an even further reinforced behaviour. At home it is much easier (as mentioned earlier, home is a more controlled environment).

    With people on the internet, I feel it’s a lost cause to try to ignore; even if the behaviour *is* attention-seeking. If one person addresses the troll, they have their attention. And, as mentioned by others, there are many other troll/bully motivations besides attention. (Note: I’m excluding stupid troll behaviour like a rallying call to attack China as payback for Pearl Harbor; that crap is easy to ignore and I don’t think anyone in China is concerned)

    There is also an extra benefit to addressing trolls/bullies that I don’t think any other commenters mentioned, it is a way to support the people they are attacking. If the only people speaking up are the a-holes, everyone else feels a lot more isolated.

  40. 48

    I always understood the “don’t feed the trolls” advice to apply to the people who would post on Batman forums to say that comics are stupid and no one likes nerds, or roleplaying forums to call everyone with that particular hobby delusional and immature. That kind of behavior is pointlessly attention seeking, no doubt. I guess I can see the projection of the idea into muddier waters; calling for the silencing of anyone with two X-chromosomes on a feminist blog can superficially seem like the same kind of rabble rousing for the sake of it. But the goal and the scale are both significantly different. Batman fanciers have no real reason to engage with Batman bashers and, frankly, cultural forces have been drifting their way, but women have to live in the same world with men who call them things. And women who call them things in the hopes of winning male approval. And men who defend women wind up ostracized, too, and queer and transgender folks get the short end of both sticks, and so on. I’m going to go ahead and keep ignoring whatever brilliant individual thinks posting impassioned dismissals of Joss Whedon when Firefly’s being discussed, but I’m perfectly capable of grasping the difference.

  41. 49

    Hmmm, interesting post but I don’t think I agree.

    The thing is this is not simply about someone either speaking to a troll or being silent – it’s whether the public conversation revolves around those who want to discuss the issues or the whims of the troll.

    There’s nothing more frustrating that reading an interesting post only to discover the comments are a bunch of good people dancing to the tune of trolls.

    It isn’t about what trains the troll to behave better but whether those who comment engage in a constructive discussion or waste their time in knee jerk responses to people who aren’t arguing honestly in the first place.

  42. 50

    Actually, Jim, for many of us it is about what gets people behaving better–in general and not just trolls. You’re welcome to only read places where someone is doing a bunch of screening for you so you don’t have to do any of the work, but you ought to at least acknowledge that someone is still doing that work. That means both that someone else is having to deal with all the trolls you want to just have magically go away and that the places where it happens are less open to conversation in one way or another.

  43. 54

    On a parallel rail. How do I deal with a situation where I’m taken for a troll just because I said something that offended someone that I don’t know and who I couldn’t have predicted that they would be offended?

    There are a few blogs that I enjoy but where I have stopped commenting because of the community there.

    I used to have an Internet name but now I use my real name, just for this reason, hoping that I will be taken seriously and not just blown off or ignored.

  44. 55

    I’m disappointed in this article because I expected it to actually cite scientific research. Instead I find amateur sociological theorizing backed up by links to eHow and opinion articles on Skepchick which similarly provides zero citations to back up the claim — just anecdotes of her own experiences since elevatorgate and a reliance on a false dichotomy between engaging with the trolls and being silenced.

  45. 57

    Thank you.

    It is bleeding obvious that treating douche-bag adults like attention-seeking kids is A Lot Different. Yet so few people think so.

    It is not as if you don’t pay attention to trolls, no, you read what they say. Then you stay silent because you do not want to feed them.

    The troll accomplishes his purpose.

    You want to stop feeding the troll? Lock him out!

  46. 58

    I agree with at least Dolly, Stephanie Zvan, jonathanray, Jim Jepps who responded earlier.

    I also have a feeling that this article, with “science” in the title, but no science anywhere to be found, and in addition giving bad advice, IS at least a little bit of… TROLLING.

    One does not need science to see that “don’t feed the trolls” actually is the only way to deal with them: trolling is orders of magnitude easier than being reasonable, thoughtful, etc. Just a few random words for the one, and LOTS of effort for the other – you have not the trace of a chance to ever win against a troll! Like this article, claiming “science” (cheap!) but not providing it (doing ones research is expensive).

  47. 66

    Hi, a few problems with your post.

    First of all, I agree that “don’t feed the trolls” doesn’t work. I wouldn’t call it bad science though, because it’s not science.

    “It’s not surprising the advice about trolls is wrong if you look at where it came from. The idea is based in the principles of operant conditioning: reward and punishment having an effect on the frequency of behavior…”

    (1) This idea might be BASED on operant conditioning, but it is NOT operant conditioning.
    (2) There is a distinct difference between “reward” as you say and “reinforcement” (a procedure) and “reinforcers” (discrete events that increase the rate of a target behavior).

    A reward is something that the speaker *thinks* the performer of behavior enjoys. A reinforcer is an event that has been proven to increase the rate of a given behavior.

    ***If it does not increase the rate of a behavior, it is NOT a reinforcer. Reward and reinforcement/reinforcers CANNOT be used interchangeably.

    “When trolls see no reaction from the rest of the world, there is literally nothing in this equation but reward. Who’s going to go away under those circumstances? Who will extinguish bad behavior when that behavior means they’re winning?”

    Again, reward is NOT the same as reinforcement. And more importantly, “Don’t feed the trolls. Attention is what they want. Ignore them and they’ll go away,” is problematic from a behavior analytic perspective, because it is only assumed that what trolls are after is attention. What is likely much more reinforcing to trolls isn’t attention, but lulz, what people (trolls) do things for.

    However, that is a hypothesis that would require a functional analysis and single-subject designs to determine if indeed lulz are the reinforcer, and what behaviors by the trolled are a source of lulz for the troll.

  48. 68

    There were definitely more than one problem listed, all of them undermining your general thesis and misapplication of principles of behavior analysis, but I can list more.

    “The majority of internet trolls show up when something contentious is being discussed or when someone contentious (like a non-white, non-male, non-heterosexual, non-capitalist…I think you get the idea) is doing the writing. If nothing happens when they attack the writer, except that their targets get a little quieter, once again, they are only getting what they want.”

    –There are many examples that can demonstrate that these assertions don’t square with reality. (1) While trolling certainly does occur around contentious topics and contentious people, it also occurs in relatively non-contentious contexts, like for instance video games and forums. (2) The targets of trolling getting quieter in response to trolling may or may not be reinforcing to trolling behavior. However, I suggest that silence is much less a source of lulz than the irrational, emotional responses that trolling evokes. (3) I don’t think it’s accurate to assert that silence is what trolls want, or that silence is reinforcing of trolling behavior. To say so with any degree of certainty would require functional and experimental analysis of trolling behavior that could convincingly rule out individual differences and time coincidences.

    “Combine these three factors, and unless we speak up and say, “This is not the behavior I want anywhere near me,” all the trolls get is reward for their behavior. That doesn’t mean there isn’t a discussion to be had about how to deal with trolls without sacrificing the conversation. It does, however, mean that it’s very much time to stop telling us to ignore the trolls.”

    I think the biggest problem with “don’t feed the trolls” is that in a public online social environment, almost without exception, there’s always someone who takes the troll bait. All it takes is one, and there’s almost always at least one. The last sentence in your post is a like waving a giant flag that says “TROLL FOOD HERE LOL.”

    It is only when you realize the futility in caring about trolls and trying to stop trolling that you can truly be free.

  49. 69

    You didn’t actually demonstrate that I’ve misapplied anything. You just whined at pedantic length about my using lay terminology for a lay audience and said a study would be needed to prove what I said. Also, that’s still just one problem. Unproven by science is not a valid critique of a post that says, “Hey, people? Those principles you’re relying on in telling people to shut up don’t work the way they think you do.”

    Now you’re complaining about my “misapplication” of behavioral analysis in a post that pointed out that “Don’t feed the trolls” was a misapplication of behavioral analysis. You’re also doing it exceedingly badly.

    (1) When game developers and writer receive death and rape threats, video games are one of the least supportive examples you could have chosen for a “non-contentious context”. (2) This ignores any third possibility and wrongly equates irrational with emotional. (3) Hahahahahahahaha! Or you start by not acting in ignorance of what trolls say. Just like you can find evidence that they’re organized.

    Nice of you to put forth your own unevidenced theory right after demanding a study (twice) for the behaviors I already have evidence for, though. That was great for a laugh this morning. So was the idea that I’m trying to be “truly free” from trolling rather than getting people to stop giving their targets pointless, victim-blaming advice.

Comments are closed.