The author proposes a three-dimensional latent variable (trait) model for analyzing attitudinal scaled data. It is successfully applied to two examples: one with 12 binary items and the other with 8 items of five categories each. The models are exploratory instead of confirmatory, and subscales...