An important concept in SEMs is identification. Identification asks the question: Does the data provide enough information (sample moments-covariances) to find a unique solution for the structural parameters of the model? A just identified model, also known as an exactly identified model, is a model in which the number of equations is equal to the number of parameters to be estimated. Such models are perfectly reproduced by SEM (R2 = 1.0) and have no value to the researcher. Such models can never be rejected and have zero degrees of freedom. Overidentified models are models specified such that the model contains more information than is necessary to estimate path coefficients. Only overidentified models are appropriate to SEM. The model has one or more degrees of freedom.
Last Modified: 02/14/2019