 |
 |
Hi everyone I have a question about spatial statistics. In an analysis I am doing, my outcome variable is a proportion, let's say it is p_i at location i. If I then do inverse distance weighting, I can estimate the value of p at each point, based on a weighted average of other data points in the surrounding area (and I can choose within a certain radius or a certain number of points). However, I would like to weight my estimates further to account for the underlying sample sizes at the other points. Because I am estimating a proportion, a place with a small denominator will have a poorly estimated value of p and I would like each estimated p (through IDW) to use the surrounding values less when they are based on small sample sizes. Does anyone know how I can do this? Or if not for IDW, have any other spatial methods (e.g. clustering methods) been adapted to adjust for this type of weighting of denominators when your outcome is a proportion? And more practically, can this be implemented in any software? Especially ArcMap or R. Thanks in advance Helen Helen Jenkins Boston University School of Public Health You may leave the list at any time by sending the command SIGNOFF allstat to [log in to unmask], leaving the subject line blank.