7.4 Variable-level fields

Use one row per variable in a formatted table that can be filtered and reviewed.

Field	Purpose
Data element name	Client-friendly name
Dataset variable name	Exact name used in code or software
Definition	Detailed business meaning
Data type	Character, integer, decimal, date, logical, and so on
Format	Date pattern, precision, or code pattern
Allowed values or range	Valid categories or numeric limits
Default value	Value supplied when no other value is entered
Requirement status	Required, conditionally required, or optional
Constraints and rules	Uniqueness, business rules, keys, and restrictions
Source	Original field, secondary source, or derived field
Relationships	Links to identifiers, tables, or related variables
Example values	Safe examples that clarify interpretation
Missing-value representation	Blank, NA, Unknown, special code, and so on
Missing count or rate	Extent of missingness for the stated version
Unit	Dollars, minutes, percentage, count, and so on
Derivation	Reproducible rule for an engineered variable
Sensitivity	Privacy classification or handling restriction
Owner or steward	Person or role responsible for meaning or quality
Status and notes	Changes, ambiguities, corrections, or caveats

Not every variable will have a default, constraint, or relationship. Record None or Not applicable when the absence matters. Zero missing values should also be documented rather than left blank.