Публикация

It’ll always change a bit study to study, and I believe that is the correlation with the modeled latent trait not the test score itself (which the articles points out is nonsense). Most papers don’t do that latent trait modeling