Structural equation modeling: Difference between revisions

Revision as of 16:02, 10 June 2005

Revision as of 14:05, 14 June 2005

Line 3:

Among its strengths is the ability to model constructs as latent [[variable|variables]] which are not measured directly, but are estimated in the model from a number of manifest variables assumed to 'tap into' the construct. This allows the modeller to explicitly capture unreliability of measurement in the model, in theory allowing the structural relations between latent variables to be accurately modelled.

SEM encourages a confirmatory, as oppposedopposed to exploratory, approach to modelling. In other words it is normal to start with a [[hypothesis]], specify a model that reflects this and then set about operationalising the constructs of interest with a measurement instrument and test the model. Often the initial hypothesis requires adjustment in light of model evidence, but it is rare to see SEM used in a purely exploratory mode.

With an accepted theory or otherwise confirmed model, one can also use SEM in an inductive mode by specifying a model and using data to estimate the values of free parameters.