[R] Joint counts and spatial autocorrelation - binary data

[DBK]

Hi All, 

I am new to this form and new to R, having just initiated the analysis of my
first project using R. I have been working on a logistic model of land use
change and am concerned about 1) measuring spatial autocorrelation and 2)
including an autocovariate in my model.

Here is what I think I need to do. My dependent variable takes on the value
of either 1 or 0 depending on whether a particular pixel from a random
sample of pixels transitioned from undeveloped to developed during the
study’s time period. I’d like to measure spatial autocorrelation by first
creating a distance matrix which includes the distances from each of my
sample points to every other sample point using latitude and longitude
coordinates. I would then like to divide these distances into bins (e.g. 0
to 500m, 501 to 1000m etc.). I have done this using both the dist() function
in geoR and the earth.dist() function in fossil. Both seem to work. I would
then like to calculate the number of similar states (i.e. joint counts)
between pairs of sample points at various distances (i.e. bin size).  I can
then graph this as proportion of disconcordance in land use change versus
distance (i.e. disconcordance is when one point of the pair transitions and
the other does not) very similar to that done by McDonald and Urban (2006).
I’d expect that as distance increases, the proportion disconcordance
eventually reaches that of the entire sample or more correctly, that of a
totally random spatial process. How can I assess the joint counts most
easily? Are there other approaches folks would recommend? Remember, I am
somewhat of a newbie.

The last wrinkle is that I am using multi-level models, which with varying
intercepts and/or slopes seems to account for some spatial heterogeneity. I
am not sure how to think about the pairing of multilevel and autologistic
approaches. Any insights?

Sample data below
Long	Lat	Trans
-87.9424	47.46495	0
-88.0451	47.464	0
-82.7524	42.89477	1
-86.6905	45.5972	0
-87.7316	46.57988	0
-82.4769	43.05674	1
-83.4313	42.42828	0
-86.2598	44.37078	0
-86.2559	44.66841	0
-86.2467	44.67979	0

-- 
View this message in context: http://old.nabble.com/Joint-counts-and-spatial-autocorrelation---binary-data-tp26306008p26306008.html
Sent from the R help mailing list archive at Nabble.com.