datatypes - USG-SCOPE/data-dictionary GitHub Wiki

Here's a more complete explanation of the datatypes used in the data dictionary spec, version 1.0.

• The nominal datatype is for categorical data where there is a limited number of categories and the list of values is not naturally ordered. Examples: sex, marital status, or the set of categories at any level of NAICS or SOC.
• The ordinal datatype is a categorical type where the values have an order defined. Examples are preference scales, age ranges, income ranges, or educational attainment.
• The interval type is for quantitative data for which it is meaningful to take differences but not to multiply by scalars. Examples: temperature measured in Celsius or Fahrenheit; age in years.
• The ratio data type for quantitative data for which scalar multiplication is meaningful. Most quantitative measures are ratio data, such as economic indexes, and temperatures measured in Kelvin. (Absolute zero is 0K, and 50K is twice as warm as 25K.)
• The text data type is for free form strings, and is the datatype to use when the others do not apply. For text data, there is often no limit to the number of values that are acceptable, no standard category system, and there are no arithmetic operations. Examples include names of people or organizations. Dates can be put in text-data fields, and if they are formatted as yyyy-mm-dd they will sort alphabetically.