Corrections for non-response bias can be very tricky.  A lot depends on 
what the uses of the survey results are and the particulars of the 
survey implementation.  Your hypotheses about direction of biases are 
clearly plausible.

Offhand, I would suggest that you report the obtained percentage and the 
margin of error (half the width of the 95% interval).  And then describe 
your adjustment to the point estimates (similar to a raking to totals 
approach).  Keep it clear that the uncertainty due to sampling will 
still exist and that the magnitude of the adjustment to the point 
estimate is just an educated guess about a bias in a known direction

Art
[EMAIL PROTECTED]
Social Research Consultants
University Park, MD  USA
(301) 864-5570



.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to