R help - Nov 2007 - Finding proportion of observations that are outliers from the left tail of the normal distribution

Hi fellow users

I have a new R problem i am hoping to get some pointers on. I have a
dataset that is approximately normally distributed but with a fat left
tail. I am interested in a good measurement on how much fatter the
left tail is than can be expected from a normal distribution. One
thing I'll  tried was fitting a two component mixture model with the
Rmix package but i am also interested in other ideas. Now to my
question.

I would like to calculate how many observations are "outside" the
normal distribution on the left side. I want to do this by fitting a
normal distributions and then use the proportion of observations not
inside the bell curve on the left side as a statistic. Since I am
relatively new to R I dont have very much success so far and would
love any pointers or code examples. Bonus points are given for ideas
on how to estimate standard errors.

Happy for any help.

On Mon, 19 Nov 2007, Thomas Fr??jd wrote:
> Hi fellow users
>
> I have a new R problem i am hoping to get some pointers on. I have a
> dataset that is approximately normally distributed but with a fat left
> tail. I am interested in a good measurement on how much fatter the
> left tail is than can be expected from a normal distribution. One
> thing I'll  tried was fitting a two component mixture model with the
> Rmix package but i am also interested in other ideas. Now to my
> question.

Look at the qvalue package.

More generally, methods for estimating false discovery rates usually rely 
on an estimate of the fraction you describe (but typically consider 
both tails). I would guess that there are several packages on CRAN or in 
the Bioconductor suite that address this.
>
> I would like to calculate how many observations are "outside" the
> normal distribution on the left side. I want to do this by fitting a
> normal distributions and then use the proportion of observations not
> inside the bell curve on the left side as a statistic. Since I am
> relatively new to R I dont have very much success so far and would
> love any pointers or code examples. Bonus points are given for ideas
> on how to estimate standard errors.
If these are independent observations, try the boot package.


HTH,

Chuck

>
> Happy for any help.
>
> ______________________________________________
> R-help at 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.
>
Charles C. Berry                            (858) 534-2098
                                             Dept of Family/Preventive Medicine
E mailto:cberry at tajo.ucsd.edu	            UC San Diego
http://famprevmed.ucsd.edu/faculty/cberry/  La Jolla, San Diego 92093-0901