Jeff, I suspect there are some important things you're not telling us
here. Are these comparisons being carried out over a variety of
conditions of some kind (temperature, perhaps, length of time, or
concentration of a catalyst, or pH, or ...)?
(If not, how does one observe multiple values of A/B ?)
What sampling distribution is believed to apply to A? to B? to A/B?
(If A and B are both chi-square, or similarly shaped, for example, the
ration A/B would be expected to be distributed as F.)
On Sun, 27 Aug 2000, Jeff E. Houlahan wrote:
> A colleague is looking at the relative amounts of two different types
> of fatty acids (say, fatty acids A and B) that are incorporated in two
> different types of tissues. He is comparing the ratio of A:B in the
> two tissues
Is he interested only in whether the ratio A:B is different (and in which
direction, if so) in the two types of tissues? Or does he also want to
quantify the difference, and perhaps predict the quantification from
other variables?
If the question is only "whether different" and not "how much?",
a simple sign test would suffice. Or Tukey's "compact" 2-sample test.
Neither of these entails distributional assumptions.
> but the data are heteroscedastic.....he has tried several
> transformations but nothing is stabilizing the variance. Is there a
> transformation that is specifically for ratios (the ratios range from
> 1:5 to 5:1)?
Has he plotted scedasticity (observed variances, e.g.) against A or B or
A/B, to see what sort of behavior it exhibits? (Or against any of the
other variables that I conjectured must exist?) The adjective
"heteroscedasitc" covers a multitude of sins, and can hardly be said to
be "descriptive" in any reasonable sense.
-- Don.
------------------------------------------------------------------------
Donald F. Burrill [EMAIL PROTECTED]
348 Hyde Hall, Plymouth State College, [EMAIL PROTECTED]
MSC #29, Plymouth, NH 03264 603-535-2597
184 Nashua Road, Bedford, NH 03110 603-471-7128
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================