[R] Confidence interval?

Dieter Menne dieter.menne at menne-biomed.de
Tue Apr 7 10:01:13 CEST 2009


Thomas Seth Davis <Thomas.Davis <at> nau.edu> writes:

>  I need help fitting/plotting a confidence interval to a frequency
distribution....
> 

In many medical journals, reviewers only want to see some error bars.
In 90% of the cases, these are wrong or misleading, but it is hopeless
to argue with medical professors playing statisticians. So I give them
the square-root of the counts, which is not much worse than anything else,
but I do not claim these to be confidence intervals.
Have some good excuse ready when some counts are low.

If you are working in a more serious field, you might consider giving
something like estimates of the parameters, but you cannot plot these
in your curves easily. Try fitdistr in package MASS for a starter.
 
> > 
> > You must know your password to change your options (including changing
> > the password, itself) or to unsubscribe.  It is:
> > 
> >   davis234

This is really a great password, which needed 0.03 seconds on my 
computer to crack.

Dieter




More information about the R-help mailing list