I think what you want to do is to go back, and see what _sort_ of numbers you
are playing with.

If you have a categorized item, at two levels (two different categories, such
as fruit - apples and fruit - kumquats: there is no such thing as a half
apple/half kumquat), then you can 'fudge' this into a continuous variable for
math analysis purposes by using 0 and 1 (or better, -1 and +1).  You will be
playing with boolean algebra, and the rules of multiplication, etc. do not
always match up with what you learned in 3rd form. (3rd grade).

If you now have 3 categories you wish to consider, such as fruit - apples,
fruit- kumquats, and fruit - tomatoes, you are asking to put these items onto
an interval or ratio scale.  Lots o' luck!  If you claim that kumquats are
better than apples, and tomatoes better than that, you have a chance.  You
can then claim to put them onto an ordinal scale, and that the intervals of
your ordinal scale are pretty much equally spaced.  This will give you an
interval scale, and you can do the math again.

This is what happens with a Likert scale (strongly disagree, disagree,
neutral, agree, strongly agree), when we try to calculate an 'average'
response.  An ordinal scale is forced into an interval scale.  (strongly
disagree=1, disagree=2, neutral=3, agree=4, strongly agree=5, so far as
'coding' is concerned)

I'm sorry I can't go back to your original explanation of what factors and
levels you are dealing with at the moment.  For each clearly categorized item
(fruit, in the 1st example above), you have a separate 'factor,' and you
can't just 'smush' (technical term:) it into an interval scale unless you
recognize what you are doing to it.  For each factor, you _may_ have
different levels that fit into an interval scale.  If so, have a good trip.

I'm sorry this doesn't exactly answer your question.  At a glance, it looks
to me as if you have a factor (a specific sensor), with different levels (0
and 1 as you put it), and you want to use 3 levels, such as forward, stopped,
backward.  Are those 3 levels on an interval scale?  Can you force it into
that form?  If so, then go for it!

Cheers,
Jay

Fahd wrote:

> Thanks a lot for the info, but I have one other simple question. Encoding
> activities as 0 or 1 will allow me to identify the sensor correlations for
> these two activities. But what if I have a third activity, is there a way
> to figure out the set of sensors that correlate to the three activities
> instead of looking at them in pairs or is there a way to combine the
> correlations from the pairs?. Thanks.
> Fahd
>
> .
> .
> =================================================================
> Instructions for joining and leaving this list, remarks about the
> problem of INAPPROPRIATE MESSAGES, and archives are available at:
> .                  http://jse.stat.ncsu.edu/                    .
> =================================================================

--
Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA

Ph: (262) 634-9100
FAX: (262) 681-1133
email: [EMAIL PROTECTED]
web: http://www.a2q.com

The A2Q Method (tm) -- What do you want to improve today?




.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to