Technical Definition Metadata 2008/2009

1. Each variable or derived variable should be described in terms of the following metadata in addition to any general metadata.

• Data Definition
• Purpose
• Relevant Collections
• Source Data
• Derived Variables
• Values
• Detailed Definition
• Sample Code

2. The name of the Data Object. The name should conform to any relevant business standards e.g. Data Model naming Standards, Variable Naming Standards etc.

3. What the object or data item is, in business language, avoiding jargon. Examples should be given where appropriate.

4. The definition should not contain the object name to avoid circularity and it should be as clear and concise as possible.

5. The purpose of the object, in plain English (not SQL or a definitional language) and what it is used for.

6. In some cases it may be of benefit if examples of what the object 'is not' are listed to improve business clarity and avoid any confusion.

7. A list of any relevant collections eg., ILR (LR).

8. A list of other data that may be used in the object's calculation, a textual definition (label) and the relevant Dataset (where appropriate).

9. The Derived Variables produced by the Data Definition, and their associated datasets.

10. A list of the allowed business values and associated label.

11. Where available, the physical code used to calculate the data object should be included. This maybe SPSS code or the SQL script.

12. A detailed definition (in plain English) explaining how the data object is calculated, references to other related objects (Source Data) and any business dependencies should be included.

Last date modified: 15th December 2008