Скачать с ютуб MplusWizard: 31. Saving factor scores. в хорошем качестве

MplusWizard: 31. Saving factor scores.

This video is part of the mpluswizard.com course series. If you like this, please like and subscribe. ;) Course files can be downloaded from: https://www.dropbox.com/sh/qf9iz8ogqs... **Course notes:** When an SEM breaks down, due to high complexity or other issues, researchers sometimes use a 2-step approach: Step 1: Save factor scores from the measurement model. Step 2: Conduct a path analysis to test the structural model, using the saved factor scores as the new “observed variables” in the path analysis. To save factor scores from Mplus, you need to specify the following commands: SAVEDATA: SAVE = FS; !This tells Mplus what information you want to save. FILE IS factor-scores.dat; !This tells Mplus to create a file and put the factor scores there. To ensure that Mplus also exports the ID variable alongside your factor scores (so you can merge these with other variables), you also need to add the following command in the “VARIABLE:” section: IDVARIABLE IS idnumber; ! I used idnumber here because that is how my ID variable was named in my dataset, so choose the name of your own variable here. *** Warning 1: Factor scores should only be used for predictors, not dependent variables. Warning 2: Factor scores are heavily criticised because of an issue known as “factor score indeterminacy”. Factor score determinacy is the correlation between the estimated and true factor scores. It ranges from zero to one and describes how well the factor is measured. One is the best value. Mplus has an option to check the level of determinacy of your factor scores, when in the “OUTPUT:” command, you ask for “FSDETERMINACY”. This option is is available only for TYPE=EFA, TYPE=GENERAL, and TYPE=COMPLEX with all continuous dependent variables. Note: As an alternative to factor scores, you can try to save “plausible values”. *** You have reached the end of this introductory course. Well Done!