Page:Geographic Areas Reference Manual (GARM).pdf/271

 • Polygon shapes were measured to eliminate extremely narrow slivers as potential census blocks. This was done by comparing the ratio of the perimeter of the polygon to the area of the polygon, with the ratio of the perimeter of a circle to its area; in addition, the polygon had to have an estimated width of at least 70 feet. In addition, the Census Bureau developed automated processing routines to selectively recognize various geographic attributes of the polygons within which it would assign census block numbers.
 * Features were ranked according to their importance as census block boundaries based on (1) the type of boundary, (2) the feature with which it coincided, (3) the existence of special land use areas (such as military reservations), and (4) the presence of governmental boundaries, in particular, State boundaries. Boundaries were assigned a ranking preference according to these four factors (Seesee [sic] ).
 * In GBF/DIME-File areas, the existing 1980 census block designations were preserved where the census block boundary features had not changed between the 1980 and 1990 censuses.
 * At least one side of a potential census block had to be a road feature.
 * Extensions from dead-end roads/streets were used to split oversized polygons into separate blocks; such extensions were made wherever road features protruded into a large polygon and ended within 300 feet of non-road features, such as shorelines and railroads.

A major limiting factor in the delineation of census collection blocks was the range of three-digit numbers available within each BG, n00 through n99 (where n was the BG number). The Census Bureau reserved block numbers n00 and n98 for possible special uses, and it used the n99 block for water areas) [sic]. Each BG therefore could include no more than 97 census blocks. Within BGs with more than 97 polygons, the program grouped sliver or small-block polygons with adjoining larger polygons for purposes of assigning block numbers and reducing the number of census Census Blocks and Block Groups11-11