Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-L] thoughts on how science is done



Here's an article, What Is Science?, I found online two years ago:

www.fotuva.org/feynman/what_is_science.html



Phys-L@Phys-L.org writes:
Not long ago, in a galaxy not far away ...

1) Suppose you are doing an experiment involving tiny
droplets. Radius = 100 nanometers or thereabouts.
Volume = a few attoliters. Lots of fun physics and
molecular biophysics you can do with such things.

2) Data suggests the pH in the droplets is seriously
messed up. It's off from what you would expect for
pure water, off by several units. Evidently a lot
of the OH- ions are getting trapped on the surface.
Not good. Really not good.

3) You try asking various high-power theory professors
how to solve the problem. They don't immediately know,
and they're not going to drop what they're doing to
figure it out.

4) You try doing the theory yourself. There are a lot
of things that "might" be going on, but it's hard to
figure out which bits are important.

5) You do some numerical simulation and visualization.
It starts out as a simple spreadsheet and some graphs,
then gradually gets fancier.

This tells you the right order of magnitude to worry about.

It also suggests a simple three-layer model: core, mantle,
and crust.

This also provides some scaling results: the thickness of
the mantle scales like the square root of the electrolyte
concentration. It also scales like the square root of the
dielectric constant.

6) Now you go back and do the theory. You grind out the
algebra. In some (but not all) of the parameter space,
it winds up being an inside-out version of Debye screening.

7) You redesign the experiment. You mix in some NaOH and
Na2CO3 (to buffer the pH) and some NaCl (to screen the
stray fields). You take some more data .........

==================================

I mention this because schoolwork in general and textbooks
in particular tend to suggest a wildly misleading picture
of how science is done.

Even in a relatively advanced course, a typical homework
question starts with step 6: "Assume a three-layer model,
and show that the mantle thickness is equal to the Debye
screening length." IMHO that is predicated on an awful
lot of 20/20 hindsight.

A certain amount of 20/20 hindsight is necessary and
normal. The way you /explain/ something you already
know is naturally different from the way you /discover/
something new. However, there are tradeoffs. As a
teacher, you have to choose your priorities:
a) teaching students to know stuff, or
b) teaching them to figure stuff out on their own.

There's nothing wrong with item (a), but it's not
sufficient by itself. IMHO item (b) usually gets
nowhere near enough emphasis.

As a related point: Everybody is always in a hurry,
and having somebody tell you the answer is hundreds
of times quicker than figuring stuff out on your own;
recall step (3) in the story above. On the other hand,
what are you going to do when nobody knows the answer,
or (worse) when somebody is trying to sell you the
wrong answer?

http://www.amazon.com/The-Greatest-Hoax-Conspiracy-Threatens/dp/1936488493
Being able to figure stuff out is an important skill.
It is learnable and teachable.

Another point: Your average undergraduate (or high
school student) has no idea how theoretical physics is
related to experimental physics. A theorist can get
away without doing experiments, although I'm not sure
that is a good idea. In any case, an experimentalist
cannot get away without doing theory.

Yet another point: Computers have been around for
a while now. It would have been insane to skip the
numerical simulation / visualization step in the
story above. Galileo is called the father of modern
science because he unified theory and experiment.
Nowadays there are /three/ ingredients:
-- experiment
-- numerical simulation and visualization
-- algebraic, analytical, theoretical calculations

Yet another point: Science is an /iterative/ process.
You do a preliminary experiment, which leaves you more
confused then when you started. You do some reading,
some simulation, and some algebra, and then redesign
the experiment. And so on, iteratively, getting a
little bit smarter at each stage.

Again I realize that everybody is in a hurry, and in
school it seems more efficient to do the experiment
only once ... but this efficiency comes at the expense
of realism. In the real world, nobody is smart enough
to do the experiment right the first time. Why why
why do we expect students to be smarter than real-world
experts? Why do we require them to do the experiment
right the first time? Even in science-fair experiments,
all too often the student is supposed to predict the
results of the experiment beforehand. Gaaack! That's
not research; that's not even science.

This problem is exacerbated by vanity. In the scientific
literature, in many cases, people don't report the messy
iterative process by which they obtained the result;
instead they report the data from the last, best version
of the experiment. They pretend they understood the
theory from Day One and designed the apparatus accordingly.
I suppose this makes the authors seem smarter. Also it
makes the report easier to read; no journal would
publish a paper that included all the messy false starts
and blind alleys. The problem is, the sanitized version
deceives students; it gives them a false picture of the
/process/ whereby physics is done.

============

How to fix this? I don't really know, but here are
some possibly-helpful thoughts:

I've seen upper-division courses for physics majors
that integrate experiment with theory plus numerical
simulation and visualization. That's nice, but what
about all the non-physics majors?

Besides, why wait until college junior year? This
stuff ought to be taught in third grade if not earlier.
That would require third grade teachers to know how
this works, which is a problem, because they haven't
been trained to do this, and there is virtually no
support for this in terms of books or anything else.

When I was about six years old, my father set up a
system for electrolysis of water. It took him several
tries, over the period of a week or so, to get it
right. I suppose he would have been happy to get it
right the first time ... but in retrospect it was a
better lesson to see the whole iterative process of
figuring stuff out.

Obviously it is not sufficient, by itself, to
watch somebody else figure stuff out ... but it's
better than nothing!

There is a tricky tradeoff here. You want to set up
a situation where the people involved are not overly
knowledgeable; otherwise they just know the answer
and get everything right the first time. At the same
time, though, you want them to be smart enough to
figure out the right answer before too long, so they
don't waste time and/or end up with wrong ideas.

In this context, google is a curse. It is hard to
find things that are figure-outable but not googleable.

Physical Education allows people to do things over and
over so that they gradually improve; why doesn't Physics
Education allow people to do the same experiment more
than once? IMHO it would be a good tradeoff to do fewer
experiments overall, but do at least some of them with
a realistic amount of iterative refinement.

Ditto for music: practicing is allowed. In the public
schools around here, sports and music are extra-curricular
nowadays, but they still exist.

Here's a well-thought-out activity that allows for some
iteration: paper towers:
http://www.nsta.org/publications/news/story.aspx?id=52799

Under favorable conditions, having students work together
helps, because they get to see the other guy's thought
processes. OTOH all too often nobody in the group has
any idea how the process is supposed to work, and things
just go from bad to worse. So (as usual) the rich get
richer and the poor get poorer.

Also here's an amusing video about hunting for sprites:
http://www.youtube.com/watch?v=OXBsnqZt_D8

In the first 3 minutes of the video, a bunch of scientists
load a gazillion dollars of equipment into a couple of
Gulfstream G-IVs, fly around all night, and come back
with nothing to show for it. Nothing! I found that to
be refreshingly honest. That's real science. Most
documentaries would have edited out that part.

After waiting impatiently for a week, they get a chance
to try again. Things go quite a bit better.

As a related point: At various points, e.g. 10:28 in
the video, there are brief glimpses of a bunch of
scientists in a conference room, poring over the data
and trying to figure out what it means. That is some
real science going on. Most people have never seen
or even imagined what that looks like.

Think about the camera they mention about 7 minutes in.
I can imagine a TV camera. I can imagine a low-light
camera. I can imagine a high-def camera. I can imagine
a high-speed camera. But low-light high-def high-speed
all at once? I'm surprised it's even possible; at some
point you just run out of photons. That's an amazing
camera.

That's relevant, because data like this does not fall
into your lap. You don't find it by looking under the
proverbial lamp-post. You've got to really, really want
it and really, really work for it.

This video looks much better in high def. There's a lot
of nifty detail. The youtube version offers 720p. The
original PBS show
http://www.pbs.org/wgbh/nova/earth/edge-of-space.html
was broadcast in 1080p, and is for sale on itunes and
on DVD, but the free versions all seem to have reduced
resolution.

I like this show because it portrays real scientists in a
somewhat realistic way. For a lot HS students, the only
scientists they can think of are the villains in horror
movies.

Also, it portrays science as being very much a team sport.

Also, it shows some real contemporary research. Unlike
(say) Higgs bosons, just about anybody can understand why
sprites are interesting. Thunderstorms make a good answer
for students who want an example of something that is not
hopelessly arcane but remains a topic of current research,
because it is not fully understood. Folks are not going
to learn any deep physics from the video, but it might
motivate them to dig up information from elsewhere.

Last but not least: Some universities strongly support
giving undergraduates jobs in the research labs. This
is not at all easy to do, but if it works it completely
solves the problem of teaching 'em how research is done.
http://www.nsf.gov/crssprgm/reu/reu_search.jsp

https://www.google.com/search?q=undergraduate+summer+research+site%3A.edu

=============

I leave it as a question: Does anybody have any good
suggestions for how to teach people how science is done?
_______________________________________________
Forum for Physics Educators
Phys-l@phys-l.org
http://www.phys-l.org/mailman/listinfo/phys-l