Latent trait theory, based on the principle of local independence, we may find it difficult to see why the notion of "factoring persons" ever arose in the first place. That is, it is fairly easy to understand a common factor as a latent trait such that in a subpopulation of persons for whom that trait is a fixed number the correlation between two tests is zero. It is hard to understand a factor, whose loadings are obtained by analyzing correlations between persons, as a latent trait such that, in a subpopulation of tests for which that trait is a fixed number, the correlation between two persons is zero. Thus, the main difficulty with the notion of applying the common factor model to a matrix of "correlations between persons" would be the logical difficulty of interpreting the residual covariances as partial covariances, the uniqueness of a person as the residual variance about the regression of the person on the factors, and so on. The position taken here is that the common factor model is a statistical model and not a device that is applicable "inversely" to "correlations" between persons. Nor, it seems, has any cogent need to apply the model in this way ever been demonstrated.
In juvenile delinquency, A Latent trait theory factor accompanying anxiety states is the pent-up hostility the individual is unable to release. This trait provides a clue to the generic origin of anxiety conditions. Anxiety-ridden individuals are hostile individuals--hostile to themselves, as being threatened, without being able to defend one or to manipulate the threatening situation to one's own advantage, produces antagonism toward such situations and the key personalities (or their ideal surrogates) who have figured in them. Such delinquents' furious resentment of all adult authority, grounded in chronic distrust, becomes readily understandable under these.........
