home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!caen!uvaarpa!murdoch!usenet
- From: mk9y@faraday.clas.Virginia.EDU (Michael Kubovy)
- Subject: Regression and bootstrap query
- Message-ID: <1992Jul24.092353.3819@murdoch.acc.Virginia.EDU>
- Sender: usenet@murdoch.acc.Virginia.EDU
- Organization: University of Virginia
- Date: Fri, 24 Jul 1992 09:23:53 GMT
- Lines: 19
-
- I have a designed experiment in which each of 50 subjects responds to
- a figure in each of 180 orientations. For each subject I perform a
- robust linear regression of the subject's reaction time, rt
- (to a property of that figure) on the figure's orientation.
- I have one rt per orientation.
-
- I would like to use bootstrapping to estimate the s.e. of the slope
- and the intercept of each subject.
-
- Because the regressor is not subject to
- error, I must use fixed resampling. But I can't resample the residuals
- this way because I have only one residual for each level of the i.v.
-
- Any advice?
- --
- |\ /| / MICHAEL KUBOVY (kubovy@virginia.edu)
- | \/ |/ OFFICE: 804-982-4729; LAB: 982-4751; FAX: 982-4766
- | |\ Department of Psychology, Gilmer Hall,
- | | \ Univ. of Virginia, Charlottesville, VA 22903-2477
-