On Jan 29, 2011, at 5:21 PM, peterfran...@me.com wrote: > > Sorry for not being more explicit in my emails, the code I am using > was not generated by me therefore I was not in the position to share > it. > > Sorry if the information i have provided is insufficient or if I > have wasted peoples time. > > Basically I have the information listed below I.e n and the > quantiles of the random distribution and was wondering if a mean > could be calculated from this or if I needed to modify the code to > output the mean? > > > the next columns are quantiles of the randomised PD distribution for > significance testing > # the probabilities are given as column names and are > designed to allow 1- and 2-tailed > # testing at the 0.05,0.01 and 0.001 levels, plus a > spread at 0.1 intervals across the range.
With the distribution characterized at that degree of specificity it should be possible to come up with a fairly reasonable estimate for the mean. If it is reasonably symmetric, the midpoint of the the 20th and 80th percentiles is a pretty good estimate (You also have the median within your output.) -- David > > > Sent from my iPhone > > On 29 Jan 2011, at 21:48, David Winsemius <dwinsem...@comcast.net> > wrote: > >> >> On Jan 29, 2011, at 9:39 AM, Peter Francis wrote: >> >>> Dear all, >>> >>> The data is generated from 1000 random samples of a phylogenetic >>> tree to calculate phylogenetic diversity. I sampled the tree 1000 >>> times at with various species communities (600) to get a random PD >>> per community. I then want to test my observed PD with that of a >>> random sample to test for significance. >> >> Color me puzzled. I thought you already had a test (the p value) >> and you wanted an estimate for a mean given that you had a 95% CI. >> >>> However the script i used, output >>> >>> q0.005 q0.01 etc upto q0.995 >> >> Those look like names or labels for quantiles. This would move >> along a lot faster if you gave a more complete listing of the >> output and the code used to generate it. >> >> -- >> David. >>> >>> But i wanted to know the mean PD value per community based on the >>> output, and that is where i am struggling >>> >>> Thanks, >>> >>> Peter >>> On 29 Jan 2011, at 02:16, Joshua Wiley wrote: >>> >>> Hi Peter, >>> >>> Do you know the formula used to calculate the confidence >>> interval? I >>> suspect it is possible with minimal algebraic manipulation of the CI >>> formula to find what the mean is. Assuming a normal distribution >>> (as >>> David), then it is certainly possible to find. This wikipedia page >>> might help: >>> >>> http://en.wikipedia.org/wiki/Confidence_interval >>> >>> And no, this is not really the correct place to ask. My basic >>> rule of >>> thumb is, "Does my question have anything to do with R? If my >>> answer >>> is, "No." then I usually look for somewhere else to post. Of >>> course, >>> for a comprehensive list, see the posting guide. >>> >>> If you are wondering if there is a function to do it for you, I am >>> not >>> sure, but it would be trivial to programme and if you show us the >>> formula for it (the mean from the CI), we can certainly give you >>> pointers for how to write your own :) >>> >>> Cheers, >>> >>> Josh >>> >>> On Fri, Jan 28, 2011 at 2:15 PM, Peter Francis >>> <peterfran...@me.com> wrote: >>>> Dear List, >>>> >>>> I am not sure if A) this is possible or B) the correct place to >>>> ask. >>>> >>>> I am looking to find the mean - i have n, and the two-tailed >>>> confidence intervals 0.95 & 0.25 with a p-value of 0.05. >>>> >>>> Can i find the mean from this data ? >>>> >>>> Thanks >>>> >>>> Peter >> >> David Winsemius, MD >> West Hartford, CT >> David Winsemius, MD West Hartford, CT [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
