[Squeakland] Panel discussion: Can the American Mind be Opened?

Alan Kay alan.kay at vpri.org
Fri Nov 23 15:56:43 PST 2007

Here's what we did for City Building, Playground, and then Etoys 
(I've written about this before, but I don't think on this thread). 
"City Building" is a wonderful (at that time non-computer) curriculum 
designed by Doreen Nelson that is very rich and has been used 
successfully for many age ranges - in our case we implemented it with 
Doreen's help for 3rd graders - which was the youngest group tried up 
to that point. Google Doreen and "City Building" for a wealth of info 
on this terrific curriculum design.

Playground was a different way to do Etoys (similar graphics model 
and a different programming model). This was implemented in a grade 
4-5 classroom (the school didn't have grades by age, but "clusters" 
by developmental level - which works a lot better).

Doreen helped in every step of introducing "City Building" to very 
willing "3rd grade" teachers. Still, it took 3 years before the deep 
quality in the curriculum was manifest in the classroom and in the 
students and what they did and how they did it. Photographs of each 
of the three years would not reveal much visible difference. It was 
what the children were concerned with, how they talked about it, and 
how they went about the processes that changed profoundly. Trying to 
trace all this back into "what happened?" we came to the inescapable 
(and not too surprising) conclusion that the teachers had also 
changed -- they had learned much more about design and systems over 
the three years, and this was manifested in a "well above threshold" 
assessment from Doreen and the rest of us in the 3rd year.

It's worth noting that assessments of fluency do not require control 
groups because what is being judged is not a teaching method or a 
curriculum per se, but results. Were the children doing deep "City 
Building"? No for the first two tries, Yes from the 3rd try onwards. 
Similarly, "are the children doing real math and real science or 
not?" Questions like these are easily answered by people who can tell 
the difference (just as musicians and coaches can assess their 
learners for degrees of fluency).

The City Building experience and our long stay in this school allowed 
us to try the same multiple year assessment for Playground 
programming and its curriculum (with similar results). Basically, 
there are just a lot of things that don't get normalized in single 
trials of even worthwhile curriculum ideas that get smoothed out over 
a few years. The teacher gets more knowledgeable and confident. The 
curriculum is improved from some of the bugs found. The software 
often requires tons of work over the three years before it is above 
threshold, etc.

When we started on Etoys 10 years ago, we had the three year trial in 
mind, and decided that all the initial curriculum would be tested 
over three years before we wrote it up (the substance of Kim's and 
BJ's book "Powerful Ideas in the Classroom" is about a dozen 
projects, each of which was tested over three years).

What we don't know from this methodology is whether there are better 
ways to teach Etoys and the math and science powerful ideas in these 
examples. And we don't know whether the choices of the math and 
science examples are the most appropriate. But what we do know is 
that the processes of their book are highly likely to result in more 
than 90% of a class of children getting fluent in what's in the book, 
and that includes strong elements of differential vector geometry, 
acceleration and Galilean gravity, etc.

This leads to interesting arguments, especially wrt young children, 
of the kind "if you can get 10-11 year olds to do real math and real 
science, then it doesn't much matter what the specific subject matter 
is". And "if the specific subject matter can be strongly related to 
adult uses and thinking about real math and real science, then all the better".

This bypasses the much more difficult problems of taking a given 
theory of subject matter (school maths, etc.) and trying to contrast 
different ways of teaching it. We do not do that at all, and the 
Etoys work was done as part of "science time" in these classrooms (a 
great place to teach real math given the difficulties with the school 
math goals and processes).

The main point here is that above threshold fluency for 90%+ of the 
children is one of the most important benchmarks -- and it can be 
done a little more easily than trying to use specific control groups 
if the subject matter is very different from school theories, yet 
still recognizable by experts.

A side comment. The reactions against "the new" take partial form in 
demands for "super scientific studies", and most of these are simply 
not feasible, if our "three years for a good experiment" is valid. 
But the largest most devastating studies in the US are the "whole 
country" results that show beyond a shadow of a doubt that the 
existing educational process is not resulting in more than a small 
percentage of children getting above acceptable thresholds in 
reading, writing, math and science (and thinking). This is the 
problem they don't want to even discuss. Contrastive studies are not 
interesting unless both are above threshold. If neither are, back to 
the drawing board. If one is, then a more detailed contrast is of little value.




At 05:29 PM 11/22/2007, David Corking wrote:
>Tony Forster wrote:
> > Controlled blind large studies are rarely done. This is because the lab
> > rabbits are real kids and there are real ethical concerns. We are 
> stuck with
> > anecdote and assertion for the large part. We need to critically 
> examine all
> > this, as there is little hard evidence.
>For better or for worse, our society uses real kids for blind (and
>even double blind) trials of medical treatments.
>The ethics of a pendulum swinging from 'new math' to 'new new math' to
>'back to basics' and on, based each time on anecdote, are, to the
>naive observer, as great a cause for concern as giving two matched
>groups of children differing curricula for a couple of years.  Perhaps
>saying that ruins my chances of influencing education, but instead of
>advocating such trials, and dismissing current research methods, my
>next step is to understand how, as a society, we should interpret an
>anecdotal study.
>What are the benchmarks a study must meet to be considered good
>evidence to support making a change (to the learning environment, the
>learning methods, and even the learning objectives, or even just to an
>individual lesson plan?)   Educators like yourself work hard on these
>studies to get them through peer review, or incorporated in government
>policy, and often aim for to be utterly dispassionate.   So, how
>should a concerned parent (or administrator or politician) work with
>teachers in their community to separate the wheat from the chaff.
>Squeakland mailing list
>Squeakland at squeakland.org

More information about the Squeakland mailing list