Research on Rewards w/ Humans

CSGnet Archive CSG1996
1

FROM CHUCK TUCKER 960102 20:00 EST

       I have just completed a more detailed reading of a book that
       I would recomend for everyone's reading: Alfie Kohn. 1993.
       PUNISHED BY REWARDS: the Trouble with Gold Stars, Incentive
       Plans, A's, Praise, and other Bribes. New York: Houghton
       Mifflin. BF 505.R48 K65 ISBN 0-395-65028-3. It is a
       detailed and documented critique of the incorporation of
       Skinner's ideas of reinforcement (called "Popular
       Behaviorism" by Kohn) into three aspects of American life:
       work, education and home. He demonstrates (to my satisfaction)
       that not only do rewards not accomplish their purpose but that
       it is detrimental to working, learning and family life. But
       he qualifies his argument by noting: "But when we talk about
       reward, the goal is very much at issue. If what you want is
       to get a child, a student, or a worker to do what you say,
       the answer to the question "What's the alternative to
       rewards?" is that there probably is no alternative (with
       the possible [179] exception of punishment). To induce
       short-term compliance, behavioral manipulation is the best
       we've got. If, however, your goal is to tap your employee's
       intrinsic interest in doing quality work, or to encourage
       your students to become lifelong, self-directed learners,
       or to help your child grow into a caring, responsible, decent
       person, then it makes no sense to ask "What's the alternative
       to rewards? because rewards never moved us one millimeter
       toward THOSE objectives. In fact, rewards actively interfere
       with our attempts to reach them [180]." Then he writes about
       what changes should be made in what we do in the workplace,
       school and home to accomplish these objectives without
       rewards or praise.

       For those who question his analysis he offers a piece of advice
       that should be considered by everyone on this net when we
       continue our research: "Very few disputes in the social sciences
       can be settled once and for all to the satisfaction of everyone.
       Even with more that a hundred studies already published, there
       continue to be disagreements about how key words are defined
       and what would constitute a decisive proof. Not many
       behaviorists will read this book, or any other that challenges
       the use of rewards (and the paradigm that leads people to use
       them), and announce that it is time they found a new line of
       work. Nor will most teachers, parents, or managers accustomed
       to saying "Do this and you'll get that" suddenly resolve to
       change their ways. My hope, though, is that even the reader
       who doesn't accept every argument offered here, or who has
       doubts about the import of one or another piece of research,
       may nonetheless decide that there is enough evidence to
       justify calling into question as set of beliefs and
       practices that affects us deeply every day of our lives [284]."

           Regards, Chuck

2

[FROM: Dennis Delprato (960103)]

CHUCK TUCKER 960102 20:00 EST

      I have just completed a more detailed reading of a book that
      I would recomend for everyone's reading: Alfie Kohn. 1993.
      PUNISHED BY REWARDS: the Trouble with Gold Stars, Incentive
      Plans, A's, Praise, and other Bribes. New York: Houghton
      Mifflin. BF 505.R48 K65 ISBN 0-395-65028-3. It is a
      detailed and documented critique of the incorporation of
      Skinner's ideas of reinforcement (called "Popular
      Behaviorism" by Kohn) into three aspects of American life:
      work, education and home. He demonstrates (to my satisfaction)
      that not only do rewards not accomplish their purpose but that
      it is detrimental to working, learning and family life. ...

I thought this came up before on the net. I, too, recommend the
book (& probably have "sold" several copies already by plugging
it in various contexts). However, there is a major shortcoming/
flaw/oversight/ in the book that is highly misleading. The critical
variables surrounding the oversight make for interesting speculation--
given, e.g., that the author appears to be a psychologist-turned-
professional-commercial writer. Nonetheless, Kohn omits a key
finding in the research documenting response-decrementing effects
of "reward" (relative to non-rewarded control conditions) once the
reward conditions are terminated. This is that it appears that
a critical condition required for the effect is that the reward be
superfluous. The major researcher in this area (Lepper) is very
clear on this--if anyone asks, I can supply reference from a file
somewhere; I believe, Kohn even cites one or more of Lepper's
articles/chapters in which Lepper make this very clear.

The reason Lepper's finding that superfluous rewards are necessary
to obtain the crucial effect underlying Kohn's argument is that
this means that, despite Kohn's claims--or apparent claims, all this
has next to nothing to do with authentic Skinnerian theory and
practice actually based in it. From the standpoint of Skinner's
reinforcement perspective, if the person is emitting [sorry for
this awful term here] a response, reinforcers are already
following response occurrences, and all that would be accomplished
by introducing extrinsic reinforcers (rewards) is that alternative
responses are likely to be reinforced. The key is that if a reward
is superfluous, it is not FUNCTIONALLY a reinforcer for a
particular operant, so Skinnerian theory is NOT being applied to
the operant in question (e.g., a child's coloring of pictures).
If the child is coloring pictures at a given rate and we introduce
M & Ms contingent on coloring, the contingent M & Ms are not
positive reinforcers. Thus, "punished (response decrements produced)
by rewards" is important but not because it is informative about
Skinnerian theory and practice. "Pop behaviorism" is not the
faithful application of Skinnerian behavioral theory in the
world. "Pop behaviorism" refers to practices that very much
antedate modern behavioral science (repent and earn thy reward,
do this for me and I'll repay in kind) although they may be
verbally related to one technical principle or another that has
evolved in behavioral science.

Related to what I believe is Kohn's basic thesis--don't undermine
intrinsic motivation with extrinsic and artificial rewards, people
accomplish much more of lasting value when they are not under
external control--is that Skinner made much more of intrinsic
reinforcement than he is given credit for. I suspect that many
of his followers are not particularly comfortable with what he
said about intrinsic reinforcement in that this construct can
seem rather close to the dreaded mentalism. But in discussions
of education, e.g., Skinner was very clear that they only way
for education to be effective is for it to take the person
to "come under the control of the many intrinsic reinforcers
of learning" (paraphrase).

Skinnerian theory may be flawed (I think it is) but Kohn's
Punished by Reward does not administer a blow to it, and
may strengthen it. I find Kohn's book useful as a compendium
of flawed popular attempts to externally control the behavior
of others by artificial and often disingenuous means. OK,
who is the CSGNETer of the month? Oh, __ is, sorry all the
rest are losers. Better luck next time LOSERS.

3

RE: Dennis Delprato 960103

     You may be correct that Kohn does not administer a severe
     blow to some technical aspect of Skinner's Theory (if there
     is anything worth that label that Skinner has written) BUT
     he does demonstrate that the application of rewards, grades,
     prizes, praise (such as "You're a Guru!") as is being done
     everyday in our schools, work, and homes not only does not
     accomplish the purposes of "motivating" people but is very
     destructive. His suggestions for changes in how we interact
     and "behave" with regard to one another will be quite disturbing
     to most people. Most are consistent with the practices that
     Ed Ford and his associates recommend be used in schools and I
     would recommend that those who use PCT take his recommendations
     seriously (I plan to change some of my teaching practices that
     I now realize are very coersive and controlling). He also
     provides and extensive bibliography which one can use if
     there is an interest supporting the idea that "rewards
     don't work."

     I don't think (as Kohn believes [note the last statement
     on p. 284 of his book that I cite]) that any theory in the
     so-called behavioral science can be disproved to the
     satisfaction of those who believe it. No theory I know of
     has been disproved; theories (to cite an idea of Morse
     Peckham) are just abandon. It is extremely unusual to find
     anyone who has a theory of human behavior to even imply
     what Bill Powers did when he stated something like "we
     may have to go back to the drawing board if these experiments
     don't work out" [sorry, I don't have the exact quote, maybe
     Brian {who was very impressed with it} has it]. I don't mean
     to appear to be sloppy about research (who would ever accuse
     Mr. Instructions of such a thing!) but few people will
     admit their theory is flawed. I would recommend Clark
     McPhail's book THE MYTH ... as a careful examination of
     the major theories of human behavior (including SR/RT)
     where he provides clear evidence none are supportable
     by adequate evidence and several are simply fabricated
     from gossip.

     Thank you Dennis for pointing to this qualification of Kohn's
     work. Kohn uses the work of Deci quite often. His recent
     book is: Edward L. Deci. 1995. WHY WE DO WHAT WE DO. NY:
     Putman. BF 575 .A88 D45. I RECOMMEND IT ALSO

     Regards, Chuck

4

[Martin Taylor 960104 12:00]

Chuck Tucker Wed, 3 Jan 1996 15:49:07

I haven't read Kohn's book, but if I understand you correctly, the person
being "rewarded" is controlling a perception of which the reward forms
some part. In the view of the teacher/rewarder, the actions of the
student approach a desired perception of the student's actions. The
teacher intends that these actions continue after the cessation of reward,
but the reverse happens--cessation of reward makes those actions less
likely than they would have been if there had never been a reward. Is
that the situation?

If the situation is approximately as described above, then one can ask
a few questions. Firstly, do the actions being rewarded sometimes serve the
student in controlling some perception that had previously been hard to
control? If so, then the "reward" serves to enhance both the student's
and the teacher's ability to control their perceptions. If not, then
the loss of control of the perception to which the reward contributes,
that occurs when the reward is withdrawn, probably leads to a greater
error than would have been the case before any reward had been given,
both because of the conflicts induced during control for reward (reward-
getting actions affect the student's ability to control for other
perceptions), and because of a transient in the perceptions to which
getting the reward contributed.

If the reward-getting actions were previously being performed by the
student as part of some useful perceptual control, then the only real effect
of withdrawing the reward would be the transient, because the usefulness
of those actions in controlling other perceptions would already
probably have been developed through the reorganization process.

If the reward-getting actions had not previously been performed (much) by
the student, and they contributed to control of some other useful perceptions
as a side-effect, then the result should be to stop them from being
reorganized away. Using the reward has permitted the student to enhance
the control of perceptions that were not previously being well controlled.

It seems, therefore, that the use of reward should be neither encouraged
nor discouraged as a general principle. If, even for a while, its use
helps both rewarder and rewardee to control their perceptions better than
they otherwise would, then that's a benefit to be weighed against the
possible future reduction in controllability. If its use helps the student
to discover useful actions for control of useful perceptions, then both
teacher and student benefit in the long as well as the short term. The
problem arises when the student is performing the rewarded actions ONLY
for the reward, and doing so conflicts in an important way with the
control of other perceptions. Is this the situation that leads to:

    he does demonstrate that the application of rewards, grades,
    prizes, praise (such as "You're a Guru!") as is being done
    everyday in our schools, work, and homes not only does not
    accomplish the purposes of "motivating" people but is very
    destructive.

?

···

------------------------
The situation seems parallel to that of "affirmative action" programs. The
program may seem initially to benefit people hired under them, but many such
people find that their qualifications are in doubt--as is their own self-image.
And there is conflict with other people who might otherwise have landed the
job in question. But if the person hired proves to be as capable as the
other people would have been, then a subsequent applicant for a related
job from the "designated group" is perceived as better suited than would
otherwise have been the case. Having hired a black/woman/blind... engineer
has proved to allow a supervisor to control perceptions of a job well done;
to hire another might be as fruitful.

I don't know if this relates to what Kohn is saying, but it seems to me
to relate to the effects of reward (and I think a parallel argument can
be made in respect of "punishment" possibly inducing some reorganization
that might aid the victim to control better).

Martin

PS. I've been following with interest the language discussion started by
Shannon, and hope to post a longish message on it some time soon.