Re: Re: A little experiment on demagnetization

Mariss

I probably contributed to your rotten day so let me stand up and be
counted.

The statistical or experimental design that you used was a pretest
postest on a single motor. It was not a sample that compares across
presumed similar devices. Yours is a pure experiment with dependent and
independent variables. As you say, the second motor was only there if
you needed further info. It has no bearing on the discussion of the
findings with the first motor. Sampling would be involved if you
asserted that this result would be the same for all similar motors
because others could say that the motor that you selected was atypical.

There are many articles written that support experiments using small
samples. While n=1 is a special case, it is a common thing. Changes do
to the manipulation, in this case the removal of the rotor, will need to
be more pronounced in order for an experimenter to assert significance.
I would think that a 30% change would raise an eyebrow on most people who
regularly do these kinds of experiments -- even with only one tested. I
could sell a lot of cereal, or software, with that kind of result!

A non-controlled variable is the only thing that you can't rule out with
the test that you did. Some might assert that there was a cosmic event
during the time between the two tests of the single motor that caused the
loss of performance rather than the removal of the rotor. This might be
where the performance of the second motor, again using a pretest postest
model at the same or nearly the same time would be valuable. If the
second motor stayed the same, it would rule out a Southern California
cosmic event.

A second way that you could mix in a second motor to your experiment,
would be to do the same procedure with it and compare results using
statistical measures like the mean and standard deviation of the two
motors. Now the common tests for significance apply -- but it's a shame
to trash motors like this just to make a point on a list!

This is probably about as far on this topic as I can go here without
owing list mom a third feeding at NAMES. No sense starting one of those
(put your advanced whatever here) urinating contests.

Ray -da youpper, eh- Henry

On Tuesday 04 February 2003 10:35 pm, you wrote:
> Message: 23
>    Date: Wed, 05 Feb 2003 04:24:28 -0000
>    From: "Mariss Freimanis <mariss92705@...>"
> <mariss92705@...> Subject: Re: A little experiment on
> demagnetization
>
> Jerry,
>
> No umbrage taken. I did test the "victim" motor before I disassembled
> it; I'm too cheap not to. The purpose of the second motor was to act
> as a baseline because once the original one was taken apart, there
> would be no way to restore it should additional data be required.
> This was the case by the way when issues of inductance change came
> up; I was thankful I had an identical motor to check against.
>
> Again, no hard feelings and sorry for the sarcasm; it's been a very
> long day. I do appreciate your criticism, I realize it was well meant
> and it keeps my work honest.
>
> Mariss