Re: Programmed instruction

In a message dated 04/17/2002 6:43:58 PM Eastern Daylight Time,
[email protected] writes to B2K list:

<< Here's a post from a highly respected old hand at the behavioral and
 academic games: >>

If the 'old hand' is highly respected, can't you identify who it is,
and provide a reference citation?

<< "The main reason programmed learning ala Skinner failed is
 that, as was often the case, he had preconceived ideas about
 how things should be done. >>

Were Skinner's ideas about instruction and programmed learning
pre-conceived or post-conceived?

<< In this case, he was convinced that errorless learning was
 the ideal and so when a program was tested and errors
 occurred, smaller intermediate steps were added so as to
 preclude errors by future students.  The result
 was, indeed, deadly dull programs for most students. >>

As you may know, I am not a proponent of so-called "errorless
learning."  In Precision Teaching I tend to favor "errorful
learning," or at least contingency arrangements where learners
may make errors, setting aside for a while what exactly we
mean behaviorally by the word "error."  But if an errorless
program can be developed efficiently, and if it produces
the same kinds of REAPS* outcomes as an errorful program,
I wouldn't reject it out of hand.

That said, either way, and in all instruction, a goal is to produce
certain responses given certain sD's.  Those certain responses
are called "correct" if they align with what a verbal community
denotes as such.  We call the responses "errors" if they deviate
from what the verbal community specifies.  Errorful and errorless
learning are two ways to achieve the same overall result.

By the way, verbal communities evolve and drift with respect to
their cultural practices, including what the members thereof say
about things.  In schools today there has arisen practices
such as "invented spelling," where student errors are not only
not corrected, it is deemed somehow harmful to their fragile
psyches and to their self-esteem TO correct them. I can't
wait until this age cohort reaches college.  They'll wipe out
the last of the profs who dare to hold to any standards.  The
tragedy that is now higher education will have come full
circle by that point, and perhaps then things can be set right.

There was a scholar back in the 80s who wrote a book titled
something like "Illiterates with Doctorates" (or something close
to that; both author and title are absent from my verbal repertoire
at the moment).  It seems as though that scholar's observations
will prove prophetic.

But, getting back to PI, just so you know, I am both an admirer of
it and a critic of it, and someone who has developed it, and someone
who has taken it to the next steps by way of computer-based
instruction, too.  Been there, done that, best credentials of all.

I would no more advocate using the same kind of PI programs of
the Skinner era today than designers of commercial or military
aircraft would suggest that we go back to open-air cockpits and
bi-plane designs of aircraft of the 1910's and 1920s. The computer
has meant that the technology has advanced way beyond what
Skinner could do with paper and pencil systems and then later
with mechanical teaching machine systems.  Still, criticizing
Skinner for his pioneering work would be like an aviation expert
criticizing the Wright Brothers for their dogmatism about aero
planes!

<< Much as I admire Skinner for many of this creative ideas, he was
 probably the most opinionated person I ever knew.  Actually, errors
 are beneficial to learning when appropriate correction is allowed.
 Sometimes we learn more from our mistakes than otherwise.  What was
 needed was "branching programs," ones that moved along rapidly but
 branched off to teach something when an error occurred. >>

Branching, as Don Cook* pointed out, is something of a false issue.
Cook pointed out that branching can realistically go only 2,
maybe 3, branches deep before the size of your branching program
grows too large at a geometric rate of progression.  Two branches
may not be sufficient to correct an "error" (once we define behaviorally
what an "error " is, of course). Three branches may, but adding that
third level of branching increases the production cost tremendously.
A four-deep branching system would be cost-prohibitive.

There are other solutions besides branching, but that's a topic
for another time.

<< Another way of saying all of this is that Skinner did not have a
 good THEORY of learning to bring to bear on the method; he had his
 dogmatic truths.  It's a pity, because the methodology has so much
 to offer."  >>

Skinner asked a question, are theories of learning necessary? If
they are not necessary, then it stands to reason that there is
no need to have a theory, good or bad.

Characterizing his pioneering work, which ran against common and
accepted practices, and which adopted a totally new world view,
as "dogmatic" seems very uncharitable. It's too bad people say
these things, because if they examined their premises about
what is and what isn't dogmatism, Skinner hardly qualifies as
dogmatic. Consistent, yes.  I guess to some people consistency
reduces to evidence of dogmatism. Dogmatism, as I understand
the word, refers to offering arrogant assertion of opinion as truth.
Whatever else one may wish to say about Skinner, his entire
science and application of it rested on an experimental base.
I would think that those philosophers, scholars, and writers
who did not have such an experimental basis upon which they
built their theories would be more likely candidates for the
dismissal as "dogmatic," wouldn't you agree?

Oh well, there seem to be many critics of Skinner out there. I guess
the deal is, how well will Skinner's contributions stand the test of
time? Moreover, what have the critics contributed?  Very few of
them have pioneered and generated their own branch of science. -- JE

----------------------------------
* REAPS stands for Retention, Endurance, Application, Performance
standards, and Stability. It refers to the effects of fluent mastery.

**Cook pointed out this issue of branching in an article titled
"CBT's Feet of Clay."  I can't find my copy of it at the moment.