It means a whole lot. It isn't too difficult to see that a spatially
correlated outcome
could be the consequence of a bunch of predictors, each of which has a
spatial pattern.
An entire discipline, called geography, used to talk about little else,
although the idea
will be familiar to others, say ecologists, geologists, epidemiologists
and even economists.
For example, if you travel through a large area, the natural vegetation
will change
as the rainfall and temperature change. The maps of rainfall and
temperature will show
patterns too.
Similarly the Walmart/Starbucks ratio (I steal from a certain blog) will
vary in your country
depending on all sorts of factors, most of which will have a clear
spatial pattern.
Whether it's a big deal as far as the statistical assumptions of your
model are concerned -- which may
be the entire focus of your real question -- is a rather different
story. The literature
I looked when I last worked a little in this area made assumptions about
the spatial properties
of the error term only, but spatial dependencies among predictors also
might show up as multicollinearity.
Just guessing.
Allan Joseph Medwick
How much sense does it make to look for spatial dependence in my
independent variables in addition to reporting the regular descriptive
statistics? If it exists, does it mean anything?
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
