🚀 go-pugleaf

This message was posted last week on SAS-L. Due to the dearth of responses I am
posting it here as well ...

I am analyzing a dataset with a very small sample size (my largest n is 11).
The data are concentrations of various environmental pollutants. The client
thinks his data are lognormally distributed. He wants me to compute: mean,
standard deviation, confidence limits of the mean, tolerance limits etc.

{Personnally, I think the patient was dead on arrival and that I am doing a
statistical post-mortem. I have pleaded with the client for more data to no
avail. Worse still, a statistician was not consulted before collecting data.}

I am following methodology suggested in chap. 13 of Richard O. Gilbert 1987.
Statistical Methods for Environmental Pollution Monitoring. These methods are
rather tediuos, even with a small dataset as they involve cubic interpolation
(4-point Lagrangian interpolation).

Does anyone know of a package that would do these computations automagically?
I'd also appreciate SAS code that does this. One individual has suggested that
the data be log-transformed which will make it normal and then utilize the
'usual' formulas form  normally distributed variables. After the various
intervals have been computed, the data are then retransformed back. I do not
think is appropriate -- especially with small sample sizes.

SAS/QC (which I do not have here) does some stuff with the lognormal
distribution, in 6.06, but apparently it does not compute intervals for
lognormal data. Is this correct?

Finally, I have been away from STAT-L for the past 3 months. I guess I have a
lot of catching up to do. Happy New Year!