Summary Data Description

Information about the collection situation, data format, universe, geographic coverage of the study, and unit of analysis.

Replication of the element geoBndBox is NOT recommended. The purpose of a bounding box is to support high level geographic point search systems. Most search systems of this type do not handle multiple instances of a bounding box. The bounding box should represent the full geographic coverage extent of the of the overall datasets being described by the Codebook instance. If there is a desire to provide the equivalent of a bounding box for each of multiple summary descriptions, use of the boundPoly is recommended. First provide a geoBndBox for the full area covered by the study. Then provide a boundPoly for each geographic area defined with a separate sumDscr. Note that when describing a bounding box using a boundPoly description the four corner points are described. The starting point and end point should match (closing the polygon).

Replication of sumDscr within the parent stdyInfo is useful when bundling specifics like timePrd, nation, and universe for specific samples within a larger project. A clear example of this is description for the various IPUMS project that harmonize multiple samples of census, health, and related data. IPUMS identified the coverage of individual samples within a project using a combination of these three elements to be able to differentiate between samples.

The examples in this description show the use of sumDscr to bundle specifics regarding timePrd, nation, and universe. A new sumDscr is used for each bundle.