Some comments on the use of factor analyses