hello i need your help about something, i would really apreciate if someone replies.
I have a questionaire with about 200 variables and the number of questionaires is 200 also. I proceed to factor analysis and come up with 41 factor. For the purposes of my research this is a big number so i decided to do a second factor analysis. What i did is to take the means of the variables that consist everyone of the 41 factors and give them this value.So basically its like a made a questionaire with 41 variables and not 200. After this i did the second FA and come up with 7 factors which is very good result.
I want to ask if the procedure i made is possible or if it is complitelly wrong.
anyone can help me out??
I have never seen a factor analysis with an subject/item ratio of 1:1. You
based your analysis on a correlation matrix with about 20'000 entries. With
just n=200 this seems a bit disproportionate.
Maybe http://pareonline.net/pdf/v10n7.pdf could be helpful.
thanks for your answer it really helped me a lot. unfortunately i have one last question.
while procceding factor analysis i had no missing values but when i tried to make confirmatory FA using the Amos program i message appeared saying that i have missing data.
how is this possible? In EFA in Spss no missing data and in CFA in Amos using the same file from Spss missing data? Whats wrong?