Re: resampling methods are serious procedure?



Illiwhacker...

I have a problem grasping what you meant. You wrote...

Using re-sampling to teach classical statistical techniques, while it
may 'work', therefore only reinforces the problem, by giving credence
to these techniques in the first place.

In my experience teaching "comparative data" (inference) via
resampling strips off the mystique and levels the playing field. It
virtually eliminates the "gotchas" typically spewed by profs. to their
undergraduate students. By this I mean
touting how much they know as compared to their struggling students.

procedures go against intuition shows how 'difficult' the whole
subject is, and hence how much need we all have of their expertise.

Indeed classical methods (or is it classical teaching) propagates
"look at how difficult this is"
and "if you don't pay close attention to all the little rules and
regulations you will stumble and fall."

When I can use a method of teaching (no matter where it may come from)
that inspires students and makes them want to learn more I'll use it.
True whether I'm teaching "statistics" or certain difficult subjects
in reaction engineering.

Some years ago I was asked to teach at a science camp for middle-
school kids. This was a local thing, sponsored by our country school
system. My assignment was to do a seminar in "statistics".
The kids attending this are certainly much brighter than the average,
and many of their parents worked at for the same company and at the
same technical center where I was employed.

This was a challenge. The time allowed was, as I recall, no more than
1 hour.

I came to the idea of teaching estimation and comparative data
(comparing averages). With equations? I don't think so. Building on
some teaching methods I'd used for engineers, scientists, and managers
I decided to use a couple of shoeboxes with M&Ms in them... about 1000
- 1200 M&Ms in each box. It was resampling time!! I let the kids draw
M&Ms (blindfolded... no picking and choosing) and proceeded from
there. The fraction of red M&Ms in one box was different from the
fraction in the other box. It almost goes without saying, they were
excited. Especially when I told them they could eat the candy at the
end of the day (I did this seminar four times for different groups).

But more than the candy, they learned. I explained how they could use
resampling to compare "data" (averages, fractions, proportions, and
more) in real-life circumstances... things they see or do "every day".
They learned, and they were exciting about learning. No equations. No
fine print. Just"go for it."

The fun part came when some parents came to me days later (the camp
lasted 5 days) and asked me what I'd taught their kids. They said the
kids were wound up about "it" and tried to teach their parents what
they'd learned. One even wanted to do a science project built around
the concept of resampling.
What more could I ask?

I honestly believe that if profs. would teach basic statistics using
resampling methods and then explain that there are other (classical)
methods that involve using equations and some "fine print" they would
end up with some inspired students. Some might even get so inspired
that they want to take higher-level courses in statistics.

Why, oh why, do we insist that entry-level students learn things that
are truly not intuitive. I'd like to see them get inspired before
moving too far upstream.

I know this violates an important principle... "Spinach first, dessert
later".

The same is true when teaching certain subjects in engineering (I'm a
chemical engineer and also a statistician.) The best engineering
profs. I found at the undergraduate level taught without using
equations until the students were inspired and ready for the spinach.
By far the best prof. I ever had was in a first course in chemical
engineering. He taught and inspired and when it came time to take a
test the questions went somewhat beyond what we'd learned in the
classroom. In short, we had to add common sense and good judgment to
engineering technology. At first this was frustrating. But all who
attended his courses will agree "Dr. Littlejohn made us successful."
Among other things, we learned to never, never, accept "facts" at face
value until we have verified them.

I wish profs. teaching statistics would do the same.
In too many instances students are taught to use computers and
software and push data through "analysis" without asking questions
about how such data came to be and may be "wrong" with it.

Nearly all the data that comes to me is flawed in some significant
way. Thanks to Dr. Littlejohn I examine it, "play with it", and look
for gross errors. (Key-entry errors and things that just don't look
right".

I know this deviates a long way from the matter at hand... resampling.
But in too many instances I've seen well-intentioned people "draw vast
conclusions from halfvast data". Just read it out loud.

Be of good cheer. OMU

"Comforting the afflicted and afflicting the comfortable."


.

Relevant Pages

  • Re: Teaching Stats
    ... I note the conversation on teaching problems in universities. ... And my students were doing ... ... well as fancy (multimedia) teaching TOOLS. ... The university from which I retired did have classrooms in which each ...
    (sci.stat.math)
  • Re: Balanced vs Uncoordinated
    ... much of that comes from confusing the hell out of students during the ... much of achieving competent aeronautical understanding comes from ... followed in stages by further teaching and further experience, ...
    (rec.aviation.student)
  • Re: Balanced vs Uncoordinated
    ... much of that comes from confusing the hell out of students during the ... much of achieving competent aeronautical understanding comes from ... followed in stages by further teaching and further experience, ...
    (rec.aviation.student)
  • Re: Improvisation class
    ... begin using a jazz vocabulary to improvise with (sort of a "bridge the ... materials should still be available to you. ... potential students interested. ... workshop info.  My work schedule of teaching 60 plus hours a week also ...
    (rec.music.classical.guitar)
  • Re: Metric accents of 4-time
    ... I am dealing with this in my teaching. ... some students will learn because ... of ways in which people absorb information, and hopefully one is able to ...
    (rec.music.theory)