EViews.com

After reading the Eviews manual my understanding is that the target method allows one to place restrictions on the loading matrix to help interpret the factors. I am using this method, restricting some of the loadings to be zero. However, once the rotation is completed, although the resulting score is in line with my priors, the loadings matrix itself does not seem to have zeros for the variables I restricted. Am I doing something wrong or is the target method not doing what I think it is?

In a previous post, I had asked about placing restrictions on the loadings and was told it is not possible, so am now slightly confused as it seems the target method fits the description of what I am after. Any help would be appreciated.

Sorry. I had asked our Factor guy who had forgotten that we had implemented target rotation as one of the rotation methods. However, you should be aware that target rotation may not quite do what you want.

As noted in Browne,

There is a salient difference, however. In
confirmatory factor analysis, specified factor loadings are forced to assume
the specified values of zero. Misspecified elements may only be detected
indirectly through examination of the overall measure of fit supplemented by
modification indices.

In target rotation, corresponding elements of the
rotated factor pattern matrix are only made as close to the specified zeros
as possible. Differences can be large so that misspecified zeros are easily
detected.

Ah, that makes sense! Much appreciated.