Test for differences in percentage distributions?

[Dan Edelstein]

Hi,

This is probably more of a stats than an SPSS question, but I would
appreciate any help.  An engineering student came in with the following
data on different materials run through a grinding mechanism, and the
percentage distribution of weight that ended up in different sieve sizes,
representing different sized particles.



Weight percentage

Sieve diameter (mm)
black plastic
white plastic
rubber
19
0.9
0.0
0.0
12.5
4.2
1.3
13.3
9.5
18.2
17.7
26.1
8
20.6
20.8
28.7
4.75
39.2
48.6
29.6
2.83
11.7
10.0
2.3
2.36
1.8
1.0
0.0
1.7
1.4
0.4
0.0
1.18
1.0
0.1
0.0
0.85
0.5
0.0
0.0
0
0.4
0.0
0.0

It seems clear that the rubber broke into bigger pieces than either
plastic, as represented by a greater percentage of its weight falling
through the bigger sieves, while the distributions of the two plastics
seem fairly similar.  Does anyone know of an appropriate statistical test
for differences between these types of distribution, and if so how to
implement it in SPSS (or any other stats package you know of)?

Thanks in advance.

Dan

----------------------------------------------------
Daniel M. Edelstein
Academic Data Centre Manager
Leddy Library
University of Windsor
[phone], ext. 4722

Given the character of metadata, there is little likelihood of seeing a
metadatum in the wild.

[Maguin, Eugene]

All,

If there is a better place to post this question, please tell me and I'll
try there. A while back I posted a question about evaluating equivalence of
means and a kind person suggested a very useful article by Rogers et al. My
specific situation is that I want to evaluate the difference of several
treatments against a control and the equivalence of alternative treatments
with a specific treatment. Except for deciding on the width of the
equivalence interval, the equivalence problem would probably not be too
hard. The wrinkle is that I have covariates and so I need to work with
adjusted means. Emmeans gives the adjusted means and the SEs. Can I use
those data to conduct a test of equivalence? If so, how would I do that/can
someone suggest a citation?

Thanks, Gene Maguin