Re: Comparison of means: Unknown distribution and Unequal Variances

student wrote:

I am doing an experiment, on certain random numbers with an extreme
value kind of a distribution, when the true underlying distribution is
unknown. The problem is something like this:

Interested in knowing the differences in means of min{X1,.....,Xn}
where Xi ~ iid unknown dist, for different values of n.

Obviously many assumptions of ANOVA are violated. To complicate things
further Levene's test p-value is 0.000, so we are nowhere in the range
in which the F-test remains robust. Box-cox transformation is not
working. Since, the distribution of Xi's is unknown, I was thinking
about some non-parametric test. But I am pretty sure that permutation
tests and there special cases (the Wilcoxon's tests) won't be valid
because of the high levels of variability. Does anyone has any
suggestions?


Yes.  Post this to sci.stat.

--
Stephen J. Herschkorn                        sjherschko@xxxxxxxxxxxx
Math Tutor on the Internet and in Central New Jersey and Manhattan

.

Relevant Pages

  • Re: the official point of view expressed by kernelnewbies.org regarding reiser4 inclusion
    ... distribution, I assume they enabled it knowing what they where doing because ... Or because they did 'allmodconfig' or 'allyesconfig'. ...
    (Linux-Kernel)
  • Re: Getting ahead: College grads wages drop 5.2% since 2000
    ... not in evidence. ... That's your straw man. ... difference between knowing something and not knowing something. ... you *MUST ASSUME* 50% occurrence and distribution. ...
    (rec.martial-arts)
  • Re: Sort benchmark and a classic paper by Gray
    ... That's where the Poisson distribution comes in. ... Yes, but I was going beyond that, and proving "with probability one" ... I was, of course, relying critically on the keys being a U ... Knowing the TRUE ...
    (comp.arch)
  • Re: calculating 95-percentile for a non-Gaussian distribution
    ... the description of a distribution could be piecewise and exactly describe at least the last 5 percent; in such a case you could calculate the 95-percentile without knowing the distribution of the rest. ... As a simple thought experiment on those lines, if your data set had 101 samples and you had examined 96 of them so far and had found they were all the same, then you do not need to examine the remaining data samples, as no matter what they are they would not be able to push the 95-percentile away from that repeated value. ... The known formula might be manipulable to calculate the mean and standard deviation in terms of a relationship between the unknown parameters, and you might know what the value of that relationship is without knowing the parameters themselves. ...
    (comp.soft-sys.matlab)
  • Re: Question about eth0
    ... If unsure always tell, which distribution ... Things like "Networking Control" are highly ... you've got some stale configuration files somewhere ... Can't tell you without knowing which distribution you use. ...
    (comp.os.linux.networking)