Question on LV method and repeated indicator approach

[dhenderson17]

Hello all,

I have read over the posts on this forum and I have two questions about the mechanics of the LV method and the repeated indicators approach. My research model has 7 first order reflective constructs which lead into 2 second order formative constructs, which then leads into 1 third order formative construct.

With the LV approach, I have seen others use Principal Components Analysis (PCS) used to calculate the weights for the first order factor on the second order factor. My question is: is it better to use PCA in SPSS to calculate the weights for the first order factors on the second order factor? If SPSS were used to calculate the weights, would I then look at the Component Matrix to find the weights for the first order factors on the second order factors?

As for the repeated indicators approach, is it better to repeat the indicators for the second and third order factors using Mode A or Mode B? Because the first order constructs are reflective, I would use Mode A. However, the second and third order factors in my model are formative. From what I've seen on this forum, it seems like Mode B is preferable unless there is multicollinearity among the first order and second order factors. In that case, Mode A seems preferable.

Thanks for any help you can provide.

Thanks!

Dave

[viswadatta]

you need not use spss. solve the entire model in PLS