Re: ci on radius
- From: vontressms@xxxxxx
- Date: 2 Feb 2006 17:41:34 -0800
Ray Koopman wrote:
vontressms@xxxxxx wrote:
All
I have some data that is measured in double angle space: (r,t) where r
is sqrt(x^2+y^2) and t is an angle between 0 and pi radians. There are
all sorts of ways to compute confidence intervals on the average angle
for circular data by trasforming (xbar,ybar) - the centroid in
Cartesian cooridiantes - x=r cos(2t), y=r sin(2t).
However, r is variable in my data. I would like to find a confidence
interval on r. A hypothesis test that r is not different from zero
would also be nice.
This is not homework. It is astigmatism data from manifest refractions.
r is the cylinder and t is the axis of the astigmatism.
Thanks,
Mark
Would a true r of zero be equivalent to the true centroid of the (x,y)
distribution being at the origin?
yes. It would mean that all of the data would be centered close to the
origin. Hotteling's T^2 may be an equivalent test that r=0. I still
need a CI on r though.
Thanks.
.
- Follow-Ups:
- Re: ci on radius
- From: Ray Koopman
- Re: ci on radius
- References:
- ci on radius
- From: vontressms
- Re: ci on radius
- From: Ray Koopman
- ci on radius
- Prev by Date: Re: ci on radius
- Next by Date: Counting with repetition
- Previous by thread: Re: ci on radius
- Next by thread: Re: ci on radius
- Index(es):
Relevant Pages
|
