Page 1 of 1

Problem with bootstrapping for formative model

Posted: Tue Nov 14, 2006 1:56 pm
by sandra
Hi

Bootstrapping does not work if I include my formative measurement models. The program says: "A singular matrix occured during the estimation of the outer weights of a block. Using more indicatators could solve the problem".

However, I do not think that the number of indicators is critical because each latent variable is measured with 6 indicators.
If I invert the model (use reflexive measurement models) bootstrapping works finde (and the weights are significant).

Is it o.k. to do the bootstrapping with reflexive models and then use the bootsapping results for the "weights" (as if I had used formative models)?

Thank you very much!
sandra