Two year position available for a statistician to provide data analysis support, consultation and collaboration to the Department of Surgery, University of Ottawa and the Ottawa Regional Cancer Centr
You might wish to start with Mosteller F, Tukey JW. (1977). Data analysis and regression. New York: Addison-Wesley. (there is a chapter on removing effects~ something like regression as exclusion) if
But IF you can avoid all this by taking logs, some think you are very well advised to. see page 194 Fisher RA The Design of Experiments 8th Edition 1971 (data on differences/ratios of heights of pla
The literature is extensive to say the least - a good start might be Y Lee and JA Nelder. Hierarchical generalized linear models. Journal of the Royal Statistical Society 58 (4):619-678, 1996. The c
In fact apart from the rather finnicky Yates' correction the Finnicky!! Getting the size of the test correct and not a function of the nuisance parameter? (slightly different than being "exact") If t
Below is a job posting at my institute that I would like to ecourage interested parties to consider. It might be ideal for some one who would like to work in an applied area for a couple of years bef
Shades of the "Problem of Moments" (circa 1800,s by Chebyshev) With a univariate sample it is perhaps easy to see that the standard bootstrap is equivalent to estimating a distribution function that
There is a sense in which it can't be done - Simulating from a quasi likelihood GLM - as quasi likelihood is a "dodge" from specifying a full (compound) probability model. But you can generate sample
Just noticed on page 3 of the new "Introducing S-PLUS 6" brochure the following historically unsupported claim. "Note the apparent reversal of the years at the Morris site, a data error that went und
ant2assign2C, where=1, immediate=T) Greets Diego -- Diego Kuonen http://wap.kuonen.com http://stat.kuonen.com http://www.Statoo.ch Ya era Ud. statouado? ht
Samuel Buttrey and Frank E Harrell Jr for their advise on the use of the Leaps funtion. I haven't had time yet to work on their ideas, but I'll do it as so
d by the exact same need(!) I wrote a silly function that picks up the printout and add it to the output value as an attribute called "printout": snatch <-
cut EXPLANATION was starred but perhaps could been more boldly put as "description under the (unrealistic?) assumption that variation is (best thought of as) random" although subject to some interpr
as well as linearity issue was about selection, not given some selection and some conditions essential equivilance ... In fact, some would argue that very different selections will likely be made us
This relates to a more general concern (raised a year or so ago) about likelihoods being under-emphasized in SPlus. If you want to know/see what the "weights option does" in a glm or gam you need to
For those who may still be able to attend ... OTTAWA WORKSHOP ON APPLYING STATISTICS 2000 Monday & Tuesday, May 15&16 2000 9am to 5pm Ottawa, Ontario Canada Sponsored by Department of Epidemiology an
-- This message was distributed by s-news@wubios.wustl.edu. To unsubscribe send e-mail to s-news-request@wubios.wustl.edu with the BODY of the message: unsubscribe s-news
cut cut -- and -- I found these two replies interesting particularily as they both implicitly involve the adoption of a Neyman null hypothesis (testing a main effect, say mean, in the presence of an
Two more comments I. Sometimes it might be worthwhile to evaluate the visualization against random noise or permuted data for a possibly interesting example see [1999] All maps of parameter estimates