The DDI-4 Variable Cascade and the Datum-Centered Approach

One of the features that makes metadata management so useful is the idea of reuse. Reuse allows an organization to record some metadata once and use it many times. For instance, the metadata for a variable in a recurring survey is linked to each iteration of the survey that uses it. This is powerful. Not only does it reduce the effort of the supplier of metadata, it helps the user by automatically uncovering similarities and linkages between data sets.

In DDI-4, the developers recognized that some aspects of the description of variables are reused more often than others. This led to the development of a hierarchy for describing variables, separated into levels, called the variable cascade. Each level down the cascade represents more specificity, until the metadata for each recorded datum is reached. We will illustrate the variable cascade through examples.

The ability to describe a datum provides additional power. Through reuse every copy of a datum brings all the related metadata of the original, it’s possible to follow each observation as it goes through the processing life-cycle, and it’s possible to track every data set to which the copies of a datum belong. This is presented in detail.