The seductive beauty of latent variable models: Or why I don't believe in the Easter Bunny

Seduced by their mathematical beauty, psychologists have been using latent variable models for more than a century. Whether discussing a general factor of cognitive ability, personality, or psychopathology there has been an unfortunate tendency to reify hierarchical structures without examining the utility of alternative models. To some of us, the use of latent variables was an unfortunate mistake. By emphasizing internal consistency rather than validity, parsimony of fit rather than function, the use of latent variables has led psychological measurement and theory down a beautifully seductive garden path rather than focusing on the real problem of actually being useful. I will address some of these alternatives and suggest that it is time to think more critically of the use of latent variable models in our theorizing and applications.