Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: STATA Spatial autocorrelation LM tests


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: STATA Spatial autocorrelation LM tests
Date   Tue, 10 Feb 2009 19:39:59 -0000

-spautoc- does _not_ purport to carry out tests on residuals, which is
what is asked for here. Nor is it an LM test, if I recall correctly. 

There is no mapping from spatial data to time series that makes much
sense. Even if exceptionally your data were laid in one spatial
dimension (e.g. along a highway, river, rail, beach, etc.) the
properties of generating processes in time, which are one-directional,
would still differ from those of generating processes in space, which
usually aren't.  

Nick 
n.j.cox@durham.ac.uk 

Robert A Yaffee

   Type: findit spautoc
      You'll find a program to compute both Moran and Geary's spatial
autocorrelation.
In general, your dataset must have a time variable for time series
analysis,
rather than latitude and longitude coordinates for spatial analysis.
   
From: BALAS ALEXANDRU <balexandru1981@yahoo.com>
 
> Does anyone know what is the code for LM tests for spatial 
> autocorrelation in STATA? It seems that varlmar is for residual 
> autocorrelation, but for time series, not for spatial autocorrelation?

> Or if varlmar is the correct code, how do I transform the dataset from

> a spatial one to a time-series?

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index