...
When initailly created, Codebook added a number of attributes to descriptive text fields that were intended to support future controlled vocabularies. It was thought at the time that controlled vocabularies would be listed internally as simple enumerations. As the technology changed and XML developed into the standardized us of schemas, the use of external controlled vocabularies became the dominant use. To support the use of external controlled vocabularies new structures were needed. The "concept" is the standardized means of providing the information needed to provide access to and validation of the use of external controlled vocabularies. Version 2.5 provided a means of noting the external vocabulary used at any point in the Codebook. Version 2.6 adds the ability to provide a direct link to the term within a controlled vocabuarly vocabulary used at a specific point. The standard form for controlled vocabularies identified within Codebook are based on the "concept" element. This may be expressed by an element of type="conceptType" or through the use of a "conceptualTextType" which includes the option for using a "concept" in conjunction with descriptive text. This has resulted in a number of options for expressing the use of external controlled vocabularies.
Best Practices:
Exisiting Existing documentation making use of various attributes intended to contain controlled vocabularies should be treated as terms with unspecified controlled vocabulary usage. If the controlled vocabulary is known, and the element supports the use of "concept", replicate the value in the attribute in "concept" and add the information identifying the controlled vocabulary used.
...
Repeat sumDscr for each set.
NCubes
Description
Dimension definitions
Location Map for defining location of cell contents in a data fileNCube is the DDI structure to describe dimensional data; cross-tabulations, tabular data, aggregations, etc. The nCube describes the structure which provides a label, universe, measure(s)types, dimensions, and attributes of the table as well as the relationship between the cells of table. Note that a table can be 1 to n dimensional. NCubes use variables to define each dimension of the table. Each cell in the table intersects with each dimension at one and only value. Note that some visual "tables" may be composed of multiple NCubes with a common dimension. This is generally done for display purposes and/or for consolidating printing space. The table below is composed of two nCubes; Age by Sex and Poverty by Sex.
Age < 18 years | Age 18 and over | Below Poverty | At or Above Poverty | |
|---|---|---|---|---|
Male | ||||
Female |
Dimension definitions
Each dimension is assigned a number and linked to the variable describing it using "dmns". The variable is described using a code list and the codes are used as the intercept value of cell with a dimension. The cell is identified with an array of the intercept value for each dimension in dimension order. This allows the user to reorganize the display of the table without changing the cell address. In the following table Age is dimension 1 and Sex is dimension 2. The cell address is presented in that order.
Age < 18 years | Age 18 and over | |
|---|---|---|
Male | 1,1 | 2,1 |
Female | 1.2 | 2,2 |
Location Map for defining location of cell contents in a data file
The location field of each nCube is expressed as locMap="IDRef" using the ID of the location map (locMap) defining the link between the nCube cell and its physical location of the stored value. The data item provides a reference to the nCube and each cube coordinate (dimension number and intercept value). Physical location is defined in the same way as used by a variable. The locMap can be used by both microdata variables and nCube cells making it a single consistent means of linking a description to a data item.