dave@cylixd.UUCP (Dave Kirby) (12/26/85)
In article <283@h.cs.cmu.edu> rfb@h.cs.cmu.edu (Rick Busdiecker) writes: > ...Positive reinforcement is *always* better >than punishment and it will produce other good qualities in your kids. Just an open question here to all who are familiar with developmental psychology. Once the positive reinforcement is taken away (certainly you aren't going to lavish praise on your child EVERY time he doesn't defacate on the carpet?), doesn't the behaviour diminish? It would seem to me that positive reinforcement would teach the child to do a certain thing EXPECTING to be praised for it. Once the praise disappears, the child would then have no reason to continue behaving that way. ----------------------------------------------------------------- Dave Kirby ( ...!ihnp4!akgub!cylixd!dave)
rfb@h.cs.cmu.edu (Rick Busdiecker) (12/29/85)
> Just an open question here to all who are familiar with developmental > psychology. Once the positive reinforcement is taken away (certainly > you aren't going to lavish praise on your child EVERY time he doesn't > defacate on the carpet?), doesn't the behaviour diminish? It would seem > to me that positive reinforcement would teach the child to do a > certain thing EXPECTING to be praised for it. Once the praise disappears, > the child would then have no reason to continue behaving that way. You're right. There are similar problems with any form of conditioning. The conditioned behavior is extinguished. Rather than suddenly stopping the praise, you switch from continuous reinforcement to intermittant reinforcement. I forget which conditioning schedule is the most effective, but I think it's a random variable interval. Basically, you continuously praise the child until you get the conditioned behavior (not defecating on the carpet) and then switch to occasional random-interval praise. Research, not just peoples gut feelings, has shown this to be a very effective method to modify behavior. Research also indicates that punishment is rather ineffective at extinguishing an undesireable behavior. It does suppress the behavior, but not eliminate it. For example, if two kids find that arguing over toys when Mom or Dad is around gets their toys thrown away, they will probably decrease the amount of arguing of toys they do ... when Mom or Dad is around. My wife tells me that the particular example of throwing away a child's belongings does in fact teach a long term habit, but probably not one which you would like to establish in your childre. The habit is stealing.
krantz@csd2.UUCP (Michaelntz) (12/29/85)
Dave Kirby writes: > In article <283@h.cs.cmu.edu> rfb@h.cs.cmu.edu (Rick Busdiecker) writes: >> ...Positive reinforcement is *always* better >>than punishment and it will produce other good qualities in your kids. > Just an open question here to all who are familiar with developmental > psychology. Once the positive reinforcement is taken away (certainly > you aren't going to lavish praise on your child EVERY time he doesn't > defacate on the carpet?), doesn't the behaviour diminish? It would seem > to me that positive reinforcement would teach the child to do a > certain thing EXPECTING to be praised for it. Once the praise disappears, > the child would then have no reason to continue behaving that way. Actually, reinforcement is most effective when it is `partial'. Partial reinforcement is an irregular pattern of reinforcing the behavior sometimes; when the reward is sufficiently desirable, the subject will respond correctly with the greatest frequency, since (s)he doesn't know which response will bring the reward, but does know that eventually the reward will arrive. Constant reinforcement, as Dave suggested, is not the best way to go; the behavior does become dependent on the reward. With partial the subject gets used to performing the behavior without getting a reward, at first in the expectation of eventually getting it; then just as a matter of habit. - Michael "I was a Psych major once" Krantz - - - - - "The text reveals the process of its own production."
paula@bigburd.UUCP (Paula Matuszek) (12/29/85)
eat, eat, eat...
I can't resist commenting on this any longer.
1. Doesn't the behavior diminish once you remove the
positive reinforcement?
Of course it does. There are several ways to minimize this, including
making the reinforcement schedule irregular, making the reinforcement
strong enough to elicit a LOT of good behavior for a single reward,
and gradually substituting "easy" or "acceptable" reinforcements
like praise and Mommy's happiness for treats, money, etc.
Why shouldn't the behavior diminish? How long would you keep working
if your paychecks stopped? Or if no one ever told you you were doing
a decent job? Kids are no different.
So how do you use positive reinforcement to control behavior? You
DO keep praising the behaviors you want to see. You can diminish the
rate for well-established behaviors, and you can, as mentioned earlier,
substitute social reinforcements like praise and smiles if you have
started with more tangible rewards. But you don't really WANT to stop
praising your child. (This tends to be the biggest failure of
"behavior modification plans"--did you follow the plan? well, yes.
did the behavior change? well yes, but once I stopped the rewards the
nasty brat went right back to his nasty behavior--this ISN'T a
surprise).
The other aspect of how you keep the behavior is that if you're
doing it right, the result of the behavior SHOULD ITSELF be rewarding to
the child. Clean room, more fun with siblings, good grades in school,
etc--children do like these things, and if their behavior is contributing
to them that in itself is reinforcing. Now if what you're using the
reinforcement for is something like trying to make a six-year-old sit
quietly for an hour, you're going to have to work a lot harder at it
(and I'm not going to help you any).
2. Positive reinforcement is always better than punishment.
This is a bit naive. Different kinds of reinforcement have different
effects. Punishment IS effective at temporarily suppressing behavior.
If it's strong enough it can be permanently effective (ever read the
experiments about one trial learning in animals?) Why might you want
to do this? Mild punishment to suppress one behavior (eg, fighting
with siblings) long enough to allow a competing positive behavior
to be built up. Somewhat stronger punishment for behaviors that to YOU
it is really critical that your child not engage in. Some such
behaviors carry their own punishment--getting burnt on a hot stove,
for instance. Sometimes you would rather that your child not do
something because the consequences COULD BE fatal--I have punished
mine for things like running into the street. The risk of the latter
is that if the connection between act and punishment is arbitrary,
with you as the arbiter, it will only prevent it to the extent that you
are likely to know about it.
3. Modeling.
Since people aren't really rats, there is more to human learning than
reinforcement (brace for flames from Skinnerians). Children, as
all of you who have them know, are great little models. When they are
faced with a situation, their first approach will be one they have
seen tried successfully, if they know of one. You CAN teach children
appropriate concepts and phrases like "taking turns" and "you do this
part, I'll do that part" and "let's trade". And "I'm sorry" and
"Let's make friends again" and all the other little niceties that
make them more pleasant to be around. Your own behavior is, of course,
one very important model. Books, playmates, schools and TV provide
others. Appropriate behavior includes not only the motivation
for appropriate behavior (which you can affect through reinforcement),
but also the knowledge of appropriate responses and behaviors--and for
humans, modeling, including just plain TELLING a child ways to deal with
something, is the fastest way to provide this knowledge.
4. My "background" for this--PhD in school psych, 12 years professional
experience working with children, and four children of my own, all
of whom I enjoy (most of the time!).
5. Most important thing I've learned about kids: they're surprisingly
resilient--don't let all the jargon and worries and theories keep
you from enjoying them!
Paula Matuszek
colonel@sunybcs.UUCP (Col. G. L. Sicherman) (12/30/85)
> Just an open question here to all who are familiar with developmental > psychology. Once the positive reinforcement is taken away (certainly > you aren't going to lavish praise on your child EVERY time he doesn't > defacate on the carpet?), doesn't the behaviour diminish? It would seem > to me that positive reinforcement would teach the child to do a > certain thing EXPECTING to be praised for it. Once the praise disappears, > the child would then have no reason to continue behaving that way. If you're genuinely pleased and show it, that in itself is an incentive that won't go away so long as the behavior continues to please you. Of course this depends on the child's loving you. In any case, artificial incentives suffer from the weakness you describe. The point is that acquired habits are hard to break; by the time that the incentive is withdrawn, the child ceases to realize that he has a choice. -- Col. G. L. Sicherman UU: ...{rocksvax|decvax}!sunybcs!colonel CS: colonel@buffalo-cs BI: csdsicher@sunyabva
cushner@ttidcb.UUCP (Jeffrey Cushner) (12/30/85)
>Just an open question here to all who are familiar with developmental >psychology. Once the positive reinforcement is taken away (certainly >you aren't going to lavish praise on your child EVERY time he doesn't >defacate on the carpet?), doesn't the behaviour diminish? It would seem >to me that positive reinforcement would teach the child to do a >certain thing EXPECTING to be praised for it. Once the praise disappears, >the child would then have no reason to continue behaving that way. > >----------------------------------------------------------------- >Dave Kirby ( ...!ihnp4!akgub!cylixd!dave) Actually, the point is to _reduce_ the amount of praise after the behavior has changed to the positive. This will encourage the behavior to improve and the child will work harder to achieve it. This principle of diminishing rewards is very familiar both to skinner freaks and gambling establishments. If it didn't work, not many people would spend alot of money on slot machines and lotteries. -- ============================================================================== Jeff Cushner @ Citicorp-TTI Santa Monica CA 90405 (213) 450-9111 x2273 {randvax,trwrb,vortex,philabs}!ttidca!ttidcb!cushner ********************************************************************* ** The above comments do not necessarily reflect the opinions of ** ** Citicorp-TTI and if the corporation wants them to, they'll have ** ** to pay through the nose for the rights! ** *********************************************************************
martinl@molihp.UUCP (Martin M Lacey) (12/31/85)
In article <625@cylixd.UUCP> dave@cylixd.UUCP (Dave Kirby) writes: >In article <283@h.cs.cmu.edu> rfb@h.cs.cmu.edu (Rick Busdiecker) writes: >> ...Positive reinforcement is *always* better >>than punishment and it will produce other good qualities in your kids. > > >Just an open question here to all who are familiar with developmental >psychology. Once the positive reinforcement is taken away (certainly >you aren't going to lavish praise on your child EVERY time he doesn't >defacate on the carpet?), doesn't the behaviour diminish? It would seem >to me that positive reinforcement would teach the child to do a >certain thing EXPECTING to be praised for it. Once the praise disappears, >the child would then have no reason to continue behaving that way. > >----------------------------------------------------------------- >Dave Kirby ( ...!ihnp4!akgub!cylixd!dave) Dave, Its been a while since I had taken a psychology course, but I seem to remember several experiments on types reinforcement and their regularity. It seems, if memory serves correctly, that the positive reinforcement makes a greater effect when administered randomly. That is to say, don't give praise on every occations or every other occation, but space the praise in regard to the last TIME a reward was given; making the time interval roughly equivilent. I *believe* this is correct, but I am more than happy to be corrected. Martin the Magician. <DISCLAIMER: Opinions or ideas expressed are all mine, mine!... Dispite fears, I don't read others thoughts and display them - without permission (usually). Rest easy associates. >
waycott@mot.UUCP (John Waycott) (01/01/86)
> > to me that positive reinforcement would teach the child to do a > > certain thing EXPECTING to be praised for it. Once the praise disappears, > > the child would then have no reason to continue behaving that way. > > You're right. There are similar problems with any form of conditioning. > The conditioned behavior is extinguished. Rather than suddenly stopping the > praise, you switch from continuous reinforcement to intermittant > reinforcement. I forget which conditioning schedule is the most effective, Instead of switching to interrmitent behavior, what I do is gradually tone down the praise. For example, when my daughter first learned how to put dirty glasses in the sink, I would hug and kiss her and tell her what a nice thing it was for her to do so. Once that behavior was established I stopped the hugging and kissing but told her how nice it was that she put a dirty glass in the sink. Now she gets a smile and a simple "thankyou". Does any one know how effective this is compared to switching to intermittent praise? It seems to work for us. -- John Waycott, Motorola Microsystems, Tempe, AZ, (602) 438-3164 {seismo!terak, trwrb!flkvax, utzoo!mnetor, ihnp4, attunix}!mot!waycott oakhill!mot!waycott@ut-sally.ARPA
wiebe@ut-ngp.UUCP (Anne Hill Wiebe) (01/02/86)
Positive reinforcement consists of following some action with a stimulus that will make that action occur more frequently in the future. This means that anything that happens after your action could be a reinforcer to your action, if you tend afterwards to do that action more frequently. THIS means that many things can be reinforcing, although they may not seem to be that wonderful or positive to an outside observer. So! what does that mean? Well, the child who is originally reinforced for positive behavior by means of a smile together with something like a cookie comes to associate the cookie (primary reinforcer) with the smile or hug (secondary reinforcer). Then the smile becomes reinforcing because it has been associated with good things like cookies; this is why it's called a secondary reinforcer, because it only reinforces after such association. (A newborn baby does not "know" that a smile is good, does not automatically seek smiles, but DOES automatically seek food. So a smile-as-reinforcer has to be taught, food-as-reinforcer doesn't have to be taught.) Okay. Now we can associate the smile (or praise, or hugs, etc. etc.) with good behavior, and it reinforces the good behavior. The combination of smile+good behavior also produces other effects on the child. Notably, s/he feels happy when this combination occurs. That means that a happy feeling (and maybe some self-praise on the part of the child) begins to be associated with the good behavior. This begins the development of a moral sense, when the child learns in this way to reinforce his/her own good behavior. People who learn "behavior modification" are just making deliberate use of this everyday occurrence or principle. So -- anyway, the child DOESN'T have to be rewarded or reinforced by the parent forever and ever. It has to happen a great deal at first, then occasionally, to keep up the association. Eventually, as the child grows up, "occasionally" can become VERY occasionally, and can be a very broad statement: the grownup child is strongly moved when the parent says "I'm proud of you." This is that same reinforcement, broadened and happening on a very occasional basis; but it sure does make the grownup child feel good (assuming that the parent-child relationship has remained good enough for the parent's opinion to still be respected by the child). Anyhow, this is why reinforcement works; my credentials are a bachelor's degree in psychology and teaching assistant work at an extremely behaviorally-oriented university (Wichita State, Kansas) and three years in a Ph.D. program in psychology at the Univ. of Texas at Austin. The above summary is not precisely the teaching or theory of any one behaviorist, but would probably be agreeable with most of those who call themselves behaviorist psychologists. Happy New Year, Anne Wiebe
colonel@sunybcs.UUCP (Col. G. L. Sicherman) (01/04/86)
[Negative reinforcement for the bug] > Well, the child who is originally reinforced for positive behavior by > means of a smile together with something like a cookie comes to > associate the cookie (primary reinforcer) with the smile or hug > (secondary reinforcer). Then the smile becomes reinforcing > because it has been associated with good things like cookies; this is > why it's called a secondary reinforcer, because it only reinforces > after such association. (A newborn baby does not "know" that a smile > is good, does not automatically seek smiles, but DOES automatically > seek food. So a smile-as-reinforcer has to be taught, > food-as-reinforcer doesn't have to be taught.) I think ill of developmental psychology in general, and of this instance of it in particular. My objections: 1. Food is good in its own right only when you're hungry. At other times it's an artificial incentive. 2. Children learn very young that smiles are good. 3. When the child behaves well, the smile comes naturally; the cooky does not. It's typical of behavioral "science" to regard the child as part of the experiment and the parent as standing outside it. The child knows better. _._ "On behalf of Bogopolis, I am honored to accept this ... Wait a minute! This is not the World's Largest Diamond!" "Of course not, Captain BUFFOON! It's a big chunk of VOOPTONITE!" "Auggggh! Urggggh! Gasp! ... And you're not Commander Pinkney, the Brazil- ian explorer! You're ... you're ... you're ..." "Ha HO ho haw! HAW HAW yak yak yak!" -- Col. G. L. Sicherman UU: ...{rocksvax|decvax}!sunybcs!colonel CS: colonel@buffalo-cs BI: csdsicher@sunyabva