Here the research evidence is divided. There are studies which show no benefit from teaching grammatical analysis (Elley 1994; Elley et al 1979; Hillocks 1986; O'Hare 1973), contrasting with other studies which do show a benefit. In the latter studies the benefit can be shown in two ways: by comparing the writing of an experimental group who have received grammar teaching with that of a control group who have not, or by correlating subjects' writing skills with their explicit knowledge ('grammatical awareness'). The link to grammar teaching in the second case rests on the plausible assumption that this enhances grammatical awareness. The references which show benefits in these ways are: Bateman and Zidonis 1966; Bryant et al 1997, 2000; Gale 1967; Heap 1991; Kennedy and Larson 1969; Klotz 1996; Mason and Mason 1997; Mason, Mason, and Quayle 1992; Mccleary 1995; Mellon 1969; Nunes et al 1997a,b; Williams 1995. The least we can say is that, contrary to popular wisdom, the question is still open.