Ambivalence of official statistics: some theoretical-methodological notes