hi,
I'm facing the following problem: My model has an endogenous lv that is specified as formative, which unfortunately impedes the use of the stone geisser criterion (Q²) to assess the predictive relevance of the model(fornell/bookstein (1982). According to the literature, I should use jackknife standard errors instead (for example chin 1998: 318). 2 related questions:
1) Where do I get these standard-errors from or how do I have to calculate them?
2) Thus far I couldn't find any critical value for the jackknife standard errors (must be <x in order to ensure sufficient predictive relevance). Can anyone give me some more information on that?
Thanks a lot for your help,
Patrick