Abstract
The construction of artificial profiles, or stimuli, for judgment analysis (JAN) is discussed. Given a prespecified discrete relative frequency distribution with finite support for each predictor variable, the construction method is used to generate sets of artificial profiles such that (1) each predictor variable possesses precisely the prespecified discrete distribution and (2) the predictor variables are uncorrelated and (3) all the profiles are unique. Sets of profiles are provided which satisfy the three aforementioned properties and for which each of the predictors takes on five, seven, or nine levels with relative frequencies chosen to approximate normality.
Get full access to this article
View all access options for this article.
