Data were processed, reviewed, sorted, and recoded using STATA/SE version 16.0. To account for the effects of the survey’s complex sampling design or the hierarchical nature of the EDHS dataset, to restore survey representativeness, and to obtain reliable statistical estimates, the data were weighted via applying the STATA command "svyset." This command was prepended to each analysis in this study.
Arc GIS version 10.8 software was used to visualize the spatial distribution and locate hotspot areas (clusters).To measure the deviation of the spatial arrangement of the uninformed choice from randomness, the global spatial autocorrelation (Global Moran’s I) was calculated [18 (link), 19 (link)]. A positive value of Moran’s I represent positive spatial autocorrelation (cluster together), whereas a negative value indicates dispersed arrangement.
Arc GIS version 10.8 software was used to visualize the spatial distribution and locate hotspot areas (clusters).To measure the deviation of the spatial arrangement of the uninformed choice from randomness, the global spatial autocorrelation (Global Moran’s I) was calculated [18 (link), 19 (link)]. A positive value of Moran’s I represent positive spatial autocorrelation (cluster together), whereas a negative value indicates dispersed arrangement.
Free full text: Click here
