However, this global measure may not be sensitive to
highly localized clustering or dispersion in breastfeeding
regulations—variations which are important to understand,
and which may hold implications both for policy and further
research—but which have been averaged over using
this global approach [15]. Therefore, in addition to the
global test statistic, we also calculated complementary
local Moran’s I values, allowing specific regions of spatial
autocorrelation to be identified as ‘clusters’ or ‘outliers’,
which might otherwise be overlooked using a global analysis.
In short, the local statistic assesses if there are
regional clusters of states with similar regulations, even if
these do not sum to an overall global correlation. We used
this technique to identify states that had high regulations
(standard fully addressed) that were surrounded by other
states with similarly high regulations (high–high). We also
identified states that had low regulations (standard not
addressed) that were surrounded by other states with low
regulations (low–low), states with low regulations that
were surrounded by states with high regulations (low–
high), and states with high regulations that were surrounded
by states with low regulations (high-low).
However, this global measure may not be sensitive tohighly localized clustering or dispersion in breastfeedingregulations—variations which are important to understand,and which may hold implications both for policy and furtherresearch—but which have been averaged over usingthis global approach [15]. Therefore, in addition to theglobal test statistic, we also calculated complementarylocal Moran’s I values, allowing specific regions of spatialautocorrelation to be identified as ‘clusters’ or ‘outliers’,which might otherwise be overlooked using a global analysis.In short, the local statistic assesses if there areregional clusters of states with similar regulations, even ifthese do not sum to an overall global correlation. We usedthis technique to identify states that had high regulations(standard fully addressed) that were surrounded by otherstates with similarly high regulations (high–high). We alsoidentified states that had low regulations (standard notaddressed) that were surrounded by other states with lowregulations (low–low), states with low regulations thatwere surrounded by states with high regulations (low–high), and states with high regulations that were surroundedby states with low regulations (high-low).
การแปล กรุณารอสักครู่..