thank you very much Christian. Will you also do a tutorial for calculating the invariance in CFA with second-order factors? It would be very useful, thanks!
Dear Christian, As always, your videos are very clear and I have learned a lot with your help. I have a doubt, in a model with second order factors, are the criteria for assessing convergent and discriminant validity different? Would you please point me to some document to read? My model is: G1 BY F1 F2 F3 F4 F5; G2 BY F6 F7 F8; where G are the second order factors and F are the first order factors. Sorry for my English. Best regards, Flavio
Thanks for the interesting videos. I am still so vague about the current video. If we assume F1-F4 as longitudinal data, how this model adds to our knowledge from Growth curve modeling? what is the advantage of this model? how the correlations of F1-F4 with F5 are helpful to understand the trait?
Hi Zahra, Thank you for watching! Please check out my other videos below that more specifically explain the ideas behind latent state-trait models. ruclips.net/video/RX7ozo6eLIc/видео.html ruclips.net/video/CT2gLZDXVZA/видео.html ruclips.net/video/qKGImLbjERg/видео.html In LST models, the covariances among the first-order (state) factors represent the stability of individual differences across time. These covariances/stabilities are accounted for by the second-order trait factor. The trait thus reflects stable (person-specific) individual differences. LST models are a special case of second-order latent growth curve models when there is no growth/change (slope factor with zero mean and zero variance). LST models allow us to partition observed score variance into trait (person-specific) variance, state residual (situational and/or person-situation interaction) variance, and measurement error variance. I hope this helps! Christian Geiser
thank you very much Christian. Will you also do a tutorial for calculating the invariance in CFA with second-order factors? It would be very useful, thanks!
Thank you very much for watching and for your suggestion. I might do a video on this topic in the future! Best, Christian Geiser
@@QuantFish that would be great, for the moment can you suggest me some articles / scripts to understand how to set them? Thanks again
Dear Christian,
As always, your videos are very clear and I have learned a lot with your help.
I have a doubt, in a model with second order factors, are the criteria for assessing convergent and discriminant validity different? Would you please point me to some document to read?
My model is:
G1 BY F1 F2 F3 F4 F5;
G2 BY F6 F7 F8;
where G are the second order factors and F are the first order factors.
Sorry for my English.
Best regards,
Flavio
Thanks for the interesting videos. I am still so vague about the current video. If we assume F1-F4 as longitudinal data, how this model adds to our knowledge from Growth curve modeling? what is the advantage of this model? how the correlations of F1-F4 with F5 are helpful to understand the trait?
Hi Zahra, Thank you for watching! Please check out my other videos below that more specifically explain the ideas behind latent state-trait models.
ruclips.net/video/RX7ozo6eLIc/видео.html
ruclips.net/video/CT2gLZDXVZA/видео.html
ruclips.net/video/qKGImLbjERg/видео.html
In LST models, the covariances among the first-order (state) factors represent the stability of individual differences across time. These covariances/stabilities are accounted for by the second-order trait factor. The trait thus reflects stable (person-specific) individual differences. LST models are a special case of second-order latent growth curve models when there is no growth/change (slope factor with zero mean and zero variance).
LST models allow us to partition observed score variance into trait (person-specific) variance, state residual (situational and/or person-situation interaction) variance, and measurement error variance. I hope this helps! Christian Geiser
@@QuantFish Thanks, I watch all the great materials. The method is almost clear now.