▪Categories. These correspond to the responses in the question on which the variable is based. For example, in the gender categorical variable, there are two categories Male and Female.

▪Special elements. These define special summary and statistics rows that are shown when the variable is tabulated, and intermediate values that are calculated during aggregation and are used in post-aggregation calculations (such as percentages and significance tests).

▪Quanvert-aware mode. Applications that understand the full range of special elements can access and utilize them.

▪Basic mode. Applications that do not understand the special elements can access the categories alone.

▪Categories column. This column contains the data for the categories. This column alone enables a simple tabulator to do rudimentary aggregation, because it contains the basic "slots" into which each case can fall. For a Quanvert categorical variable called varname, this column is called varname.

▪Extra column. This column contains the special Quanvert elements. For a Quanvert categorical variable called varname, this column is called varname.Ex and is stored in the metadata as a helper field.

▪Multiplier columns. These columns contain any numeric data associated with a categorical variable. One column is created for each element that has numeric data associated. These extra columns are not created for elements that do not have associated numeric data. For a Quanvert categorical variable called varname, these columns are called varname.multxxx, where xxx is the MDM mapped value of the element to which it applies. Like the extra columns, the multiplier columns are stored in the metadata as helper fields.