County selection in sashelp.zipcode dataset

[SDally]

I am using the sashelp.zipcode dataset to map zip codes to counties.  That dataset contains just one observation per zip code.  Does anyone know what algorithm SAS uses to select the county for zip codes that cross county lines?  Population?  Area?  Centroid?  Something else?  Thanks for your help. 

[HB]

The file has the lat and the long for the centroid of the ZIP.  I would assume that the county was determined off that centroid location, but I wouldn't know for sure.  Just seems logical.   

[SDally]

Thanks for your reply, HB.  I agree that centroid seems logical, but I would love to see some documentation that says so for sure.