We calculated mean regulation score by state for each of the
four breastfeeding standards, for centers and homes, as well
as a sum of all regulation scores combined. To examine
spatial patterns, we mapped regulations for each breastfeeding
standard using ArcGIS 10 (ESRI Inc., Redlands,
CA). We assessed levels of global spatial autocorrelation, by
regulation, by sum of regulation score, and by facility type,
using global Moran’s I in GeoDaTM. Global Moran’s
I numerically describes the extent to which areas (states in
this case) with similar attributes (breastfeeding regulations)
tend to cluster throughout a geographic region (the US) [14].
Moran’s I values range from -1 (perfect negative spatial
autocorrelation, where areas with dissimilar attributes tend
to cluster), through 0 (random distribution of area level
attributes), to ?1 (perfect positive spatial autocorrelation,
where areas with similar attributes tend to cluster). The null
hypothesis states that there is no discernible spatial pattern in
breastfeeding regulations throughout the US as a whole