Models with latent variables such as factor analysis and LISREL (LInear
Structural RELationships) are fitted on several samples. It is shown theoretically
that the standard methods developed for normal and independent samples
can be applied to non-normal and correlated samples when a particular model
parameterization is followed. The results can be used, for example in social
and medical studies, when correlated populations are compared, or when
one population is observed over time. The theoretical results are
supported by simulation studies and are also applied to real data collected
in a study about the effectiveness of the Head Start Summer Program. The
advantages and the efficiency of the proposed method in comparison
to other methods are discussed.