Exploring the quality of enumeration-area data with graphic scripts

Awareness of uncertainty and variability is essential for the reliable use of geographic data tabulated for census tracts, counties, and other enumeration areas. Users of census information and other primary areal-unit data commonly ignore errors inherent in questionnaire design and census enumeration. And these users further increase uncertainty for themselves and their readers through classification, aggregation, derivative measurements, and other data-compression techniques. This secondary uncertainty calls for programmed graphic sequences, known as "graphic scripts', designed both to reveal variation hidden by analytical techniques and to identify likely blunders in the primary data. This paper uses a series of hypothetical screen snapshots to demonstrate how a graphic script addressing these issues can heighten an analyst's awareness of the variation, reliability, and meaningfulness of census results and other aggregated data reported by enumeration area. -Author