Abstract
Two or more dimensions unintentionally varied simultaneously are said to be confounded, but several theories in personality intentionally combine 3 or more distinct qualities. Researchers using these theories sum the qualities before testing predictions. How wise is this practice? The practice appears to derive from 2 distinct lines of reasoning. One of them assumes that the component dimensions converge on a single underlying quality (latent variable) that each reflects imperfectly. The other assumes a synergy among dimensions. Issues arising from each line of reasoning are illustrated by examining self-monitoring, attributional style, and hardiness. Conclusions are that (a) information is lost whenever a latent variable theory is tested solely by a composite and (b) a synergistic theory can be tested only through a statistical interaction.
