Dear list, Apologies if anyone has received this message already. I've posted it to the list but it doesn't appear to have been posted and isn't archived.
I was hoping someone could please tell me how to run hierarchical exploratory factor analysis? I haven't used it before, but have read that I simply use the oblique factor scores from the initial EFA (incidentally i've calcualted these by simply summing the relevant items per factor), in my hierarchical "orthogonal" factor analysis. Is this correct and does it have to be orthogonal? If I do this though, given that I only have 3 lower-order factors it isn't surprising that they all load on one higher-order factor, is it? However, if I use 4 oblique factors in a hierarchical FA, 3 of them load on one factor and the other loads alone. Does this suggest I have one higher-order factor and what does it suggest about the factor that loads alone? I have also read that i could use the correlation matrix for the 4 oblique factors for the hierarchical FA. I'm not sure I have done this right though. I scored the 4 oblique factors by summing the items, then ran correlations using the code below which then saved my correlation matrix (also below) into a new data file. I then entered these 4 variables into a FA, but results of this hierarchical FA are different from when I use the factor scores (I'm not sure why or whether they should be, or which is the best approach to use in hierarchical FA - factor scores or correlation matrix). CORRELATIONS/VARIABLES=PAI_factor1 PAI_factor2 PAI_factor3 PAI_factor4/PRINT=TWOTAIL NOSIG/MISSING=PAIRWISE /MATRIX=OUT(*). PAI_factor1 PAI_factor2 PAI_factor3 PAI_factor4 1.0000000 .7326943 .6429660 .3696387 .7326943 1.0000000 .6780099 .4857973 .6429660 .6780099 1.0000000 .3509714 .3696387 .4857973 .3509714 1.0000000 Any advice greatly appreciated.ThanksKathryn
_________________________________________________________________
Celeb spotting – Play CelebMashup and win cool prizes
https://www.celebmashup.com/index2.html
Locked
