Discuss statistics related things


by a.torelli » Tue Jun 08, 2021 7:52 am

I would like to ask you why, after normalising by logarithm the items with kurtosis greater than 1 in absolute value in kurtosis and skewness, the loadings of these items in a confirmatory factor analysis are very low, only of those items. If untransformed, they maintain high loadings.

Alternatively, how could I justify moving an item from factor 3 to factor 2, although in exploratory factor analysis a placement in factor 3 is recommended? The index modification for moving such an item from factor 2 to factor 3 in the confirmatory factor analysis is not the highest but is the third highest. Such an item if moved from factor 2 would bring down Cronbach's alpha.

Thank you very much.
Posts: 1
Joined: Tue Jun 08, 2021 7:44 am

Return to Statistics