We carried out an experimental study with 2 groups of fourth graders through 10 weekly workshops of 90 min each. The experimental group consisted of 66 fourth graders (male = 26, female = 40), the control group consisted of 48 fourth graders (male = 23, female = 25), and both groups included less skilled and skilled writers. The experimental group children were taught and practiced "linguistic games" in which they had to modify narrative texts according to a specific rule or by introducing a new element, with the aim of creating a new and coherent text. The results showed an improvement in the experimental group children's ability to modify texts and an increase in their liking of narrative. Although differences in posttest writing tasks emerged between less skilled and skilled writers, no difference emerged either in liking of writing or in self-perception of competence. (Contains 4 tables.)