Spearman's original computation of g: A model for Burt?