Thanks for the lecture. 👏 By standardization of items do you mean to first standardize (subtracting mean and dividing by SD) all the items before putting them into EFA and CFA? if yes, then what to do if some items are on continuous and some are on categorical scales with different number of categories? Thank you.
I have been flowing your lectures, and I have one question: if I add latent variables to very complex path analysis, the latent will simplify my model? I really like your lessons, very didatic. Thank you
Hi, Thank you for the wonderful video! I am wondering if you can assist me with some questions I have about a rather complex CFA... I have a sample of 327 participants who completed 6 measures with a total of 140 items between them. I would like to complete a CFA to determine whether the measures are distinct from one another (each came from a separate validated scale). My questions are: (1) How do I calculate if I have sufficient power to complete this analysis? (2) Do you have any references for the factor loadings I should be using as a cut-off? I have seen people use > .300, >.400, etc. in previous stats courses, but I am not sure what is the "best practice".
Power is so tricky for SEM ... maybe simulation? I really don't have a good clue beyond that. Also tricky but .3 is related to a medium sized correlation, and I usually cite Preacher and MacCallum's Repairing Tom Swift paper.
@@StatisticsofDOOM Thank you for such a fast reply! I took a look at the Preacher & MacCallum (2003)-- it was super helpful! I was wondering if you knew of a paper which might reference a CFA and the relevant loadings?
Thanks for the lecture. 👏
By standardization of items do you mean to first standardize (subtracting mean and dividing by SD) all the items before putting them into EFA and CFA? if yes, then what to do if some items are on continuous and some are on categorical scales with different number of categories?
Thank you.
Hmm generally categorical data is analyzed differently than continuous data. Maybe treat them as separate?
I have been flowing your lectures, and I have one question: if I add latent variables to very complex path analysis, the latent will simplify my model?
I really like your lessons, very didatic. Thank you
I would only add the latent variable if it makes sense to do so (theoretical based reasons).
Hi, Thank you for the wonderful video!
I am wondering if you can assist me with some questions I have about a rather complex CFA...
I have a sample of 327 participants who completed 6 measures with a total of 140 items between them. I would like to complete a CFA to determine whether the measures are distinct from one another (each came from a separate validated scale). My questions are:
(1) How do I calculate if I have sufficient power to complete this analysis?
(2) Do you have any references for the factor loadings I should be using as a cut-off? I have seen people use > .300, >.400, etc. in previous stats courses, but I am not sure what is the "best practice".
Power is so tricky for SEM ... maybe simulation? I really don't have a good clue beyond that.
Also tricky but .3 is related to a medium sized correlation, and I usually cite Preacher and MacCallum's Repairing Tom Swift paper.
@@StatisticsofDOOM Thank you for such a fast reply! I took a look at the Preacher & MacCallum (2003)-- it was super helpful! I was wondering if you knew of a paper which might reference a CFA and the relevant loadings?