Hi,
If you wanted to statistically test in a panel data set (units of observation are state-years) whether values of a given covariate are "similar"/"homogeneous" (or different) within regions or amongst physically adjacent neighboring observation units (state-years), would you recommend to a) compute Moran's I test, or b) to compute the standard deviation of the covariate of interest across each region's units, and include this as a covariate in a model that regresses a the outcome variable of interest on such computed standard deviation?
I would appreciate comments on how to implement any of the options listed above (or whatever alternative you might suggest) in Stata.
Thanks in advance!
cross-posted here: http://stackoverflow.com/questions/4...ns-or-morans-i
If you wanted to statistically test in a panel data set (units of observation are state-years) whether values of a given covariate are "similar"/"homogeneous" (or different) within regions or amongst physically adjacent neighboring observation units (state-years), would you recommend to a) compute Moran's I test, or b) to compute the standard deviation of the covariate of interest across each region's units, and include this as a covariate in a model that regresses a the outcome variable of interest on such computed standard deviation?
I would appreciate comments on how to implement any of the options listed above (or whatever alternative you might suggest) in Stata.
Thanks in advance!
cross-posted here: http://stackoverflow.com/questions/4...ns-or-morans-i
