On 22 Nov 2002 22:01:59 -0800, [EMAIL PROTECTED] (Jay Warner) wrote:

[snip.  including Example about counts and 'class intervals' ...]

> ...   I have a mass for each class, an amount on a continuous
> scale, which I can also report as a percentage of the total.  I can
> calculate an average and standard deviation for the entire histogram.

And what do you mean by 'standard deviation for the entire
histogram'?    Either you have an error term for each bar of
the histogram, or you have nothing that is useful for testing
the bars.

> I can calculate an expected frequency for a theoretical distribution
> with the same average and standard deviation.  I can report both
> observed and expected frequencies in percent of total, stick this into
> a Chi Square, and crunch the numbers.

Standard deviation, again?  
What does this refer to, " ... report observed ... frequencies" ?

I thought there was nothing observed as frequencies in
this problem....

It is also unclear from *this*  statement whether you are hoping
to compare the  *proportions*  in each class, or whether
the overall amount -- down below, the question seems to
be entirely about matching the proportions.


[ ... ]
> How would you recommend that I make the comparison between an observed
> and theoretical distribution, with continuous observations?  and where
> might I get a text that gives some examples, so I can puzzle it out?
> 
> BTW, why do I care, you might ask.  I need to establish a density
> function within a measurable range, so I can assert loudly the maximum
> amount of mass well away from the center of the distribution.  One
> option is to find a solution to the density envelope function and
> extrapolate that.  However, my loud assertions will not have to be so
> loud if I can relate it back to a distribution expected on physical
> grounds.


So, what is the error associated with the observed masses in
the extreme classes?  If you have the error information with 
known variances, and these are presumed to be pretty  much
independent and normal errors, then you can combine 
several error terms by the usual formula.  And test against
the fixed amount for a known distribution.

You really do have to be explicit about what you know about
the error (by density, cumulative, or whatever), if you want 
survey the possible tests.

If you have the fixed, known variance, SE  for one bar, you can 
compare it to theoretical amount by the obvious z-test.  Can't you?

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
.
.
=================================================================
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