Place Facets= before any other specifications referencing facet numbers.This specifies how many different facets interact to make the set of observations.

 

Facets in data = label facets, data facets

Label facets = highest numbered facet in Labels=

Data facets = number of facets in the data file (including facets implied by Dvalues=) - not required if same as Label facets. Useful with Entry-order=

 

Example 1: For judges, persons, items:

Facets=3

 

In complex data sets, every facet may not be involved in every observation. For instance, some observations may involve persons with items and tasks, while other observations involve persons with judges and items. Since there are four facets altogether (persons, items, tasks, judges), specify four facets, even though no single observation involves all four:

Facets=4

 

A rating scale (or partial credit) category or response is not a facet. Data are the result of interactions between the facets. Thus "Mary responds to item 7 with a `3'" indicates two facets (persons and items), not three (persons, items and responses).

 

Example 2: Two facets in Labels=, but three in the data: paired comparisons.

Facets = 2, 3

Entry = 1,2,2

Model = ?,?,-?,D ; "-?" reverses the sign of the second person measure