How the Construct validity, Face validity and Predictive validity of a pharmacological research ( treatment of artificially disease induced animal ) can be assessed ?
My understanding is that these three types of validity are not formally measured quantitively but rather qualitatively. For example, a model simply has good predictive validity if it reliably identifies treatments that work in humans. Alternatively it could have poor face validity if the model looks nothing like the analogous phenomenon in humans. But in neither of these cases would you necessarily put the validity on a quantitative scale. Although with some meta-analysis of many studies you might be able to do that... For now I think it is appropriate to discuss the validities as good/fair/poor rather than anything more rigid.