How Not to Talk to Your Kids.

[Geoff]

Praise is important, of course, and I think it's good to be generally positive. But criticism is important as well, as long as it's constructive. I'm reminded of my school years in which anything written on a piece of paper, absolutely anything, would get me an automatic "A" in English class. I imagine my shock and dismay when I got into a class where the teacher actually gave us negative feedback and was a tough grader I didn't like it very much. And yet I never learned so much in school as I did in that one class. If I hadn't been so lazy and unmotivated, I imagine I could have learned quite a bit more.

Very much so, tat, and even more so when the students get out into the real world, and find that employers are a lot less generous with their praise than their teachers were (to put it mildly!).

[gib]

Mr Samsa thank you for your reply.

Please be aware that i'm not arguing that Kohn is right - tbh common sense would indicate that he is quite wrong.

Nevertheless you are arguing that science has shown that Kohn is wrong - i am still waiting for you to show me peer reviewed papers. That is how you bring science in to support an argument.

[hadespussercats]

A topic that I'm very interested in ... I've posted this a couple of times in other threads/topics, but it is directly related to what you've raised here, hades: Five Reasons to Stop Saying "Good Job!", by Alfie Kohn.

This is hard, though--

In fact, boy-o is only 8 days old, and I already say "good job" when he lets out a good burp.

Are we doomed?

[Robert_S]

I remember not caring as much about how smart I was or whether I was a good boy. I wanted to know how to do things and how far I had to go to get my skills at whatever up to some level of usefulness.

[hadespussercats]

I remember not caring as much about how smart I was or whether I was a good boy. I wanted to know how to do things and how far I had to go to get my skills at whatever up to some level of usefulness.

I was very approval-oriented (was! heh.) What saved me was that in my family, the role of perfect kid was already taken by my sister, who was "naturally beautiful, and naturally smart." So I had to work hard to play catch-up. I used to get teased for rehearsing-- if you're good, you should be able to do everything excellently on the fly. And smart people don't have to study, or work hard. In fact, studying is proof you're not that smart.

That last bit took a looong time to unlearn.

(Sorry if that sounds bitter. The "smart kids" articles really struck a chord.)

[irretating]

Bookmarking

Edit line.


... Really interested in your experiences and thoughts, irre.

My thoughts are that a mantra of 'good job' (which seems to be something that American mums seem to say a lot) would simply lose it's effectiveness. The child would probably zone out and not even register it as anything more than a smile or similar. Unless a mother/teacher/authority figure were constantly saying it in totally inappropriate circumstances, I can't see any psychological harm coming to the child. I hasten to add these are just my thoughts, not based on any studies.

Re telling a child they are smart, I think the problem there is that the focus in on something that the child has no control over (their intelligence) rather than on something the child does have control over (e.g. how much effort and attention they have or will put into whatever learning task is at hand) and that leaves the child with no useful feedback or encouragement. If a child has low self-esteem, it might help them to hear that they are smart (who doesn't want to hear that) but the key would be to encourage and assist with the practice that will really make the difference to the child's level of mastery.

[Mr.Samsa]

Mr Samsa thank you for your reply.

No problem

Please be aware that i'm not arguing that Kohn is right - tbh common sense would indicate that he is quite wrong.

Fair enough, but of course if a claim has observable effects then we can measure it with science instead of relying on common sense.

Nevertheless you are arguing that science has shown that Kohn is wrong - i am still waiting for you to show me peer reviewed papers. That is how you bring science in to support an argument.

I'm not sure what you mean by "still waiting" since this is the first time you've asked for evidence that science has shown Kohn to be wrong. Understandably, I wasn't going to jump into a discussion on a board like Rationalia with a 20,000 word essay and four pages of references. I prefer to just make my claims and if anyone wants more info then I can provide it for them.

One of the problems with Kohn's claims, however, is that they are as vague and 'not even wrong' as any of that made by a creationist, so some claims require a fair amount of background understanding in order to accurately refute them. But I'll try to provide the best links I can think of.

1) Manipulating children: This claim isn't even consistent and is more of a moral issue than a scientific one.

2) Creating praise junkies: This literally requires us to ignore everything we know about classical and operant conditioning. Basically, following a behavior with reinforcement will always increase the probability of that behavior occurring again in the future. If it is reinforced intermittently (as all informal training is done anyway, since teachers and parents will often miss instances of behavior) then the behavior becomes almost impossible to extinguish using extinction alone (i.e. removing reinforcement). This means that Kohn's "praise junkies" would continue to work at high levels even without praise, simply because they've fallen into what is called a "behavior trap". This just means that the behavior that was initially created and maintained by reinforcement (like praise), get trapped in a 'natural' reinforcement contingency; that is, the act itself becomes reinforcing.

3) Stealing a child's pleasure: Studies utilising Premack's principle tells us that reinforcement increases the probability of a subject engaging in an activity when given a free-choice. If an activity doesn't receive 'external rewards' but the child still freely chooses to engage in it, then what else could we attribute this to but some kind of intrinsic motivation?

4) Losing interest: The problem here, as I mentioned here, is not the use of praise, but the excessive use of praise. I think all teachers and parents agree that indiscriminate praising of children for anything they do will result in problems.

5) Reducing achievement: This summary of the research on praise by the APA is quite good.

[Warren Dew]

1) Manipulation: "The reason praise can work in the short run is that young children are hungry for our approval. But we have a responsibility not to exploit that dependence for our own convenience. A "Good job!" to reinforce something that makes our lives a little easier can be an example of taking advantage of children’s dependence. Kids may also come to feel manipulated by this, even if they can’t quite explain why."

He suggests that instead of doing something "to" children to get them to do something, it's better to do something "with" children to get them to do something. In essence, it's still manipulation and he can't escape this fact, but of course 'manipulation' isn't a bad thing. If I teach a kid not to run in front of speeding cars using candy then the fact that I've "manipulated" him into not killing himself is irrelevant. Kohn obviously views teaching children important societal norms and rules as simply being "convenience" or for our own good, but this is a bit silly.

While I disagree with you on the rest of the points, I agree with you on this one. Yes, when I say "good job" to a toddler for drinking out of his glass without spilling it all on the floor, it is for my convenience, since I'm the one who would have to clean up the spill. However, the kid is constantly going to have to be doing things for other peoples' convenience throughout life - for example doing a productive job to earn income - so it's not as if the kid is learning a behavior pattern that will hurt later. Additionally, the time spent not cleaning up spills is time that can be spent "with" the kid.

2) Creating praise junkies: This literally requires us to ignore everything we know about classical and operant conditioning. Basically, following a behavior with reinforcement will always increase the probability of that behavior occurring again in the future. If it is reinforced intermittently (as all informal training is done anyway, since teachers and parents will often miss instances of behavior) then the behavior becomes almost impossible to extinguish using extinction alone (i.e. removing reinforcement). This means that Kohn's "praise junkies" would continue to work at high levels even without praise, simply because they've fallen into what is called a "behavior trap". This just means that the behavior that was initially created and maintained by reinforcement (like praise), get trapped in a 'natural' reinforcement contingency; that is, the act itself becomes reinforcing.

I think you are confused here. When you praise a toddler who responds "four" after you say "what's two plus two", you aren't reinforcing generalized "work[ing] at a high level". You're just reinforcing doing arithmetic on demand, or maybe even just knowing the specific fact "2+2=4". It isn't going to help the kid learn multiplication or calculus or perhaps even double digit addition on his own; indeed, there is some evidence that praising results within the kid's abilities actually deters attempts at challenges at the limits of the kid's abilities, which could hurt future development.

That's the problem with behavioral "science", especially in the case of children: choice of hypotheses results in the scientists failing to distinguish between reinforcing 'working at a high level' and 'working at a four year old level' in a three year old - the former would be good, but the latter catastrophic if it becomes "impossible to extinguish".

[Mr.Samsa]

While I disagree with you on the rest of the points, I agree with you on this one. Yes, when I say "good job" to a toddler for drinking out of his glass without spilling it all on the floor, it is for my convenience, since I'm the one who would have to clean up the spill. However, the kid is constantly going to have to be doing things for other peoples' convenience throughout life - for example doing a productive job to earn income - so it's not as if the kid is learning a behavior pattern that will hurt later. Additionally, the time spent not cleaning up spills is time that can be spent "with" the kid.

Agreed

2) Creating praise junkies: This literally requires us to ignore everything we know about classical and operant conditioning. Basically, following a behavior with reinforcement will always increase the probability of that behavior occurring again in the future. If it is reinforced intermittently (as all informal training is done anyway, since teachers and parents will often miss instances of behavior) then the behavior becomes almost impossible to extinguish using extinction alone (i.e. removing reinforcement). This means that Kohn's "praise junkies" would continue to work at high levels even without praise, simply because they've fallen into what is called a "behavior trap". This just means that the behavior that was initially created and maintained by reinforcement (like praise), get trapped in a 'natural' reinforcement contingency; that is, the act itself becomes reinforcing.

I think you are confused here. When you praise a toddler who responds "four" after you say "what's two plus two", you aren't reinforcing generalized "work[ing] at a high level". You're just reinforcing doing arithmetic on demand, or maybe even just knowing the specific fact "2+2=4". It isn't going to help the kid learn multiplication or calculus or perhaps even double digit addition on his own; indeed, there is some evidence that praising results within the kid's abilities actually deters attempts at challenges at the limits of the kid's abilities, which could hurt future development.

But you are reinforcing their desire to work at a high level if you use praise in the method I suggested (essentially, correctly using homework will result in a high work rate). Praising kids incorrectly will of course lead to unintended results, but I haven't argued that all praise is good. It's similar to training a dog - using praise can be a hugely useful tool to get them to do what you want, but saying "Good doggy!" when he's pissing on the floor is just going to get him to piss on the floor more often.

Reinforcement increases the future probability of whatever behavior preceded it, so if you praise a kid for figuring out what 2 + 2 equals (by using pictures, or blocks to count, etc) then you are necessarily reinforcing their understanding of addition, but if you praise a kid for knowing what 2 + 2 equals then you will likely be only reinforcing their memorisation processes.

I agree with the last sentence though: if you provide the same reinforcement for an easy option over a hard one, then we will find that they will forego the hard option and just stick to the easy one. This is predicted by the basic equation that controls our choice behavior; the matching law.

That's the problem with behavioral "science", especially in the case of children: choice of hypotheses results in the scientists failing to distinguish between reinforcing 'working at a high level' and 'working at a four year old level' in a three year old - the former would be good, but the latter catastrophic if it becomes "impossible to extinguish".

I'm not sure why you've used scare quotes over the word "science" there, and then followed it up by conflating it with educational theories which usually actively reject behavioral science. Most applied behavioral science accepts the distinction you make (and can explain precisely how and why it occurs) so this is more an argument against the educational system, which I agree with. The problem is that teachers seem to be actively opposed to implementing scientific theories in their job and like to rely on outdated theories that "make sense" to them, like the myth that different people are different "types" of learning (like auditory or visual learner, etc).

[Warren Dew]

But you are reinforcing their desire to work at a high level if you use praise in the method I suggested (essentially, correctly using homework will result in a high work rate). Praising kids incorrectly will of course lead to unintended results, but I haven't argued that all praise is good. It's similar to training a dog - using praise can be a hugely useful tool to get them to do what you want, but saying "Good doggy!" when he's pissing on the floor is just going to get him to piss on the floor more often.

I agree it's quite similar to training a dog. And while I think there's more to "correctly using homework" than just patterns of praise, I even agree that correctly using homework can result in a high work rate. However, you then conflate a high work rate with working at a high level. They are two different things. For example, it sounds like hadespussercats' sister was working at a high level without having a high work rate. Educating children goes beyond the kind of training one wants a dog to have.

Reinforcement increases the future probability of whatever behavior preceded it, so if you praise a kid for figuring out what 2 + 2 equals (by using pictures, or blocks to count, etc) then you are necessarily reinforcing their understanding of addition, but if you praise a kid for knowing what 2 + 2 equals then you will likely be only reinforcing their memorisation processes.

I question whether it's possible to reliably differentiate "figuring out" something from "knowing" something. The known behavior is only that the child answered correctly, perhaps for the first time. You can't reliably know whether that was from just having figured it out, or whether it was something the child learned - whether by figuring it or by being told - a substantial time ago, and just now happened to use.

As for "understanding of addition", that applies, as I said, only to addition - even assuming that behavioral reinforcement can reinforce understanding rather than behavior, something I'm skeptical about.

I'm not sure why you've used scare quotes over the word "science" there, and then followed it up by conflating it with educational theories which usually actively reject behavioral science. Most applied behavioral science accepts the distinction you make (and can explain precisely how and why it occurs) so this is more an argument against the educational system, which I agree with. The problem is that teachers seem to be actively opposed to implementing scientific theories in their job and like to rely on outdated theories that "make sense" to them, like the myth that different people are different "types" of learning (like auditory or visual learner, etc).

I'm referring to theories published in journals and books with "behavioral science" in the title, not educational theories. For example, it was believed for decades and published in behavioral science texts that children didn't develop "theory of mind" - the ability to attribute different states of knowledge to different people - until the age of 3 or 4. It has now been shown, though, that this knowledge develops at least as early as 2, and the problem was apparently that the previous experimenters were conflating lack of knowledge with lack of the ability to communicate the knowledge using language. This kind of error is rampant throughout behavioral science.

Is there some valid behavioral science? Yes, there is, and I agree that the known effects of reinforcement on behavior is one example. However, it seems to be swamped by fuzzy thinking, things like the above mentioned conflation of inability to communicate with lack of knowledge, or conflation of understanding with behavior.

Now, I do definitely agree with you that some educational theory fails to correctly apply even the valid portions of behavioral science, as in the example of using constant praise for ego reinforcement rather than using selective praise for behavior reinforcement. However, I also think the educational system has some goals that are beyond the present limits of valid behavioral science.

[Mr.Samsa]

I agree it's quite similar to training a dog. And while I think there's more to "correctly using homework" than just patterns of praise, I even agree that correctly using homework can result in a high work rate. However, you then conflate a high work rate with working at a high level. They are two different things. For example, it sounds like hadespussercats' sister was working at a high level without having a high work rate. Educating children goes beyond the kind of training one wants a dog to have.

I agree that there's more to correctly using homework than just patterns of praise, and I didn't intend to imply that. However, I don't think I've confused working at a high rate with working at a high level, instead I've just used an ambiguous term to refer to working at a high rate. With that said, working at a high level (in the sense you're using it) is easily done through 'patterns of praise' and is often used in behavioral therapies. (All you do is provide reinforcement contingent on a scale of increasing difficult, either in terms of amount of questions answered or in terms of questions answered from each level, e.g. easy questions, to medium questions, to hard questions).

Reinforcement increases the future probability of whatever behavior preceded it, so if you praise a kid for figuring out what 2 + 2 equals (by using pictures, or blocks to count, etc) then you are necessarily reinforcing their understanding of addition, but if you praise a kid for knowing what 2 + 2 equals then you will likely be only reinforcing their memorisation processes.

I question whether it's possible to reliably differentiate "figuring out" something from "knowing" something. The known behavior is only that the child answered correctly, perhaps for the first time. You can't reliably know whether that was from just having figured it out, or whether it was something the child learned - whether by figuring it or by being told - a substantial time ago, and just now happened to use.

Whilst there can be some overlap, it is usually pretty easy to differentiate between figuring out and knowing something. In maths classes, for example, this is usually done by getting students to 'show their working' and marking them on both their working, and their answer. This means that even if they had memorised the answer and we accidentally reinforce them for memorisation, we also reinforce them for the process of figuring it out.

As for "understanding of addition", that applies, as I said, only to addition - even assuming that behavioral reinforcement can reinforce understanding rather than behavior, something I'm skeptical about.

I'm not sure why you're skeptical of it? Understanding is just another behavior that we do, and there is a wealth of literature supporting the fact that reinforcement principles apply to things like thoughts, emotions, etc. As a practical application, look at CBT.

I'm not sure why you've used scare quotes over the word "science" there, and then followed it up by conflating it with educational theories which usually actively reject behavioral science. Most applied behavioral science accepts the distinction you make (and can explain precisely how and why it occurs) so this is more an argument against the educational system, which I agree with. The problem is that teachers seem to be actively opposed to implementing scientific theories in their job and like to rely on outdated theories that "make sense" to them, like the myth that different people are different "types" of learning (like auditory or visual learner, etc).

I'm referring to theories published in journals and books with "behavioral science" in the title, not educational theories. For example, it was believed for decades and published in behavioral science texts that children didn't develop "theory of mind" - the ability to attribute different states of knowledge to different people - until the age of 3 or 4. It has now been shown, though, that this knowledge develops at least as early as 2, and the problem was apparently that the previous experimenters were conflating lack of knowledge with lack of the ability to communicate the knowledge using language. This kind of error is rampant throughout behavioral science.

That is more an issue within cognitive psychology (which I agree makes a number of fuckups), as well as problems with popularisation of scientific findings. The original findings that you're referring to suggested that there was no evidence that children under that age demonstrated ToM, not that it was impossible for them to do so. Theories were then developed based on the information they had, which were then refined when new evidence came to light.

Theories on ToM in behavioral science rely a lot less on strict notions of 'developmental stages', so behavioral science never had this problem.

Is there some valid behavioral science? Yes, there is, and I agree that the known effects of reinforcement on behavior is one example. However, it seems to be swamped by fuzzy thinking, things like the above mentioned conflation of inability to communicate with lack of knowledge, or conflation of understanding with behavior.

Behavioral science tends to question the validity of verbal reports, so this conflation doesn't tend to happen all too often. The tests which have refined our understanding of ToM have been non-communicative behavioral tests, which stemmed directly from behavioral science (in reply to the mistakes of the cognitive psychologists).

Now, I do definitely agree with you that some educational theory fails to correctly apply even the valid portions of behavioral science, as in the example of using constant praise for ego reinforcement rather than using selective praise for behavior reinforcement. However, I also think the educational system has some goals that are beyond the present limits of valid behavioral science.

Like what?

[hadespussercats]

For example, it sounds like hadespussercats' sister was working at a high level without having a high work rate.

I need to point out here that while this assessment is true, up to a certain point, when my sister started to encounter subjects that didn't come easily right off the bat, she had huge meltdowns of confidence, and then would go out of her way to avoid those subjects-- early on, the multiplication table was like this-- later algebra and physics. She quit theatre because she didn't know what to do when she kept forgetting lines on stage (because a quick perusual was no longer enough for memorization), and ultimately dropped out of college because she thought she was a fraud.

In other words, I think she suffered from the very problems described in the article I posted. Thinking she was supposed to be able to just "be excellent" did a real number on her. And I felt those effects, too-- but circumstances were different For me, and I learned other ways of coping.

[Warren Dew]

That is more an issue within cognitive psychology (which I agree makes a number of fuckups), as well as problems with popularisation of scientific findings. The original findings that you're referring to suggested that there was no evidence that children under that age demonstrated ToM, not that it was impossible for them to do so. Theories were then developed based on the information they had, which were then refined when new evidence came to light.

Theories on ToM in behavioral science rely a lot less on strict notions of 'developmental stages', so behavioral science never had this problem.

Ah, the "no true Scotsman" argument - A behavioral science publication isn't really behavioral science if it publishes anything later proved wrong, eh?

[Warren Dew]

For example, it sounds like hadespussercats' sister was working at a high level without having a high work rate.

I need to point out here that while this assessment is true, up to a certain point, when my sister started to encounter subjects that didn't come easily right off the bat, she had huge meltdowns of confidence, and then would go out of her way to avoid those subjects-- early on, the multiplication table was like this-- later algebra and physics. She quit theatre because she didn't know what to do when she kept forgetting lines on stage (because a quick perusual was no longer enough for memorization), and ultimately dropped out of college because she thought she was a fraud.

I certainly didn't mean to imply that working at a high level was the only thing that was important. A high work rate is useful, too - as well as good habits, like looking before crossing a street, that are highly trainable using a variety of techniques.

However, I will say that both the multiplication table and lines for theatre involve memorization, which I see as distinct from working at a high level, although that might not have been clear to her. It's easy for smart kids - and even to some extent smart adults - to assume that their natural intelligence is the only ingredient to all forms of success, but that assumption is not correct.

In other words, I think she suffered from the very problems described in the article I posted. Thinking she was supposed to be able to just "be excellent" did a real number on her. And I felt those effects, too-- but circumstances were different For me, and I learned other ways of coping.

If her ability to "be excellent", even to the limited extent that it was true, was due to uncritical praise, that might be an argument in favor of such praise, at least to some extent. Plenty of kids do study hard but despite that still don't excel. Ideally we'd like both talent and good habits, and while constant uncritical praise might not be a path to that, occasional uncritical praise might be an element.

[floppit]

Large chunks of this thread send shivers up my spine. I'm not writing this to win an academic debate, I don't feel any need or have any wish to debate the functionality of applying behavioural analysis in the home or the use of abstracts to inform parental rewards, and, this debate so far has reeked of abstract surfing without the depth of learning by reading the methods and discussions. In behaviourism, especially Applied Behavioural Analysis (ABA) that's where you need to look - and shit, you want to try this out with kids I can't stress the word NEED enough.

Reward schedules are complex in their own right, it's correct that a randomised schedule is the most powerful in increasing RATE of behaviour, but in schooling a child speed and quantity of response is not always (in academics, rarely) the priority. Reward schedules are chosen for much more subtle reasons, to support rigidity of behaviour, or flexibility, to enhance long term goals or target short term ones, to produce conditioned, paired, responses or encourage a more operant approach.

Then there's the limits of behaviourism. Behaviourism relies on the observable and the ability to measure and count behaviour. But there's an elephant in the room, behaviourists from both sides of the camp know it. The elephant is our internal world, the unseen stimuli. I might jump up and suddenly run out of the room, a behaviourist can note the previous or coinciding environment, they can record all that happened and when, BUT if I ran out the room because I remembered I left my gas on the stimuli is unseen and has nothing to do with the physical environment I left. The behaviourist can observe the number of times I match a word card written in chinese to the correct object, they can measure the percentage I get correct but they cannot observe the means by which it's done - is it because I'm learning to read chinese or because I have latched onto the creases in well thumbed word cards?

I sat face to face with children for 4 years using ABA, I worked with Norwegian consultants lucky enough to live in a country where it's challenges and usefulness are recognised and further research supported by the state, people who had both academic learning and experience reaching over decades - and they were STILL learning. Applying behaviourism in it's purest form is a bit like trying to guide a hippo with a dog lead! It's frigging powerful, each mistake can not only render the learning goal ineffective but can reroute learning, creating counter productive beasts that are inordinately hard to extinguish.

The behaviourism that would most help parents isn't the trial based fire water, it's an understanding that in behavioural terms, behaviour that works survives. In terms of reward schedules the most important to consider is the child's age, the younger the more immediate rewards need to be, with older children it begins to expand and token economies (basic pocket money rewards or sticker charts etc) start to work really well. The most important issue with rewarding children is variety and novelty. I specialised in ABA in my last year at uni and then went on to work in the field for real, before I went to uni I was an animal trainer - I know my shit but I would NOT EVER use ABA in it's pure form with a normally developing child because the screw up factors far outweigh benefit. If a child has learning difficulties that restrict other alternatives then I think it is justifiable and useful to use ABA, however, I still wouldn't touch it with a bargepole without support, without a fecking brilliant consultant available.