In D. Alamargot, J. Bouchand, E. Lambert, V. Millogo, & C. Beaudet (Eds.), Proceedings of the International Conference « de la France au Québec : l’Ecriture dans tous ses états », Poitiers, France, 12-15 November 2008 - [http://www.ecritfrancequebec2008.org/ ] Effects of a quality goal on writing strategies Caroline BEAUVAIS, Thierry OLIVE, & Jean-Michel PASSERAULT Université de Poitiers & CNRS Centre de recherches sur la Cognition et l‟Apprentissage (CeRCA) Abstract To determine if text quality is related to orchestration of the writing processes, the present experiment examined how a goal stressing text quality affects text quality and use of writing processes. Psychology students were instructed to think aloud while composing a narrative and an argumentative text. Half of the participants composed with a standard quality goal (no particular stress on text quality) while the other half composed with a high quality goal (high stress on text quality). Verbal protocols were categorized into four writing processes: planning, translating, reading and revising. Writers’ cognitive effort was analyzed through reaction times to auditory probes. Two sets of variables were measured: general characteristics of the composition (fluency, prewriting pause, cognitive effort…), and on-line management of the writing processes (number of episodes, mean length of episodes…). Two judges independently assessed text quality with two analytic scales. As expected, participants wrote better texts when quality was stressed. However, this stress did not affect on-line management of the writing processes. Only the prewriting pause was longer. Narratives and argumentations were composed with different strategies: narratives required more but shorter episodes of writing processes than argumentative. These findings indicate that writers change their writing behavior to adapt quality of their text to the goals of the task. Key words: Text production, Text quality, Writing strategies, Argumentation, Narrative What determines quality of a text a student has composed? Actually, more than knowing what to write in their text and how to formulate it, students have to know how to coordinate and regulate the attention paid to the writing processes (Kieft, Rijlaarsdam, & van den Bergh, 2008). Recent research has suggested that on-line management of the planning, translating and revising processes plays a decisive role in text quality (Levy & Ransdell, 1995; Breetvelt, van den Bergh, & Rijlaarsdam, 1994). In this framework, self-regulation of writing has stimulated a variety of studies, with instructions including both interventions and recommendations in writing with the main objective to improve text quality (Ferretti, MacArthur, & Dowdy, 2000; Graham, MacArthur, & Schwartz, 1995). What emerges from these studies is that writers benefit from training and explicit teaching of self-regulatory procedures of writing. For instance, one possible way used to enhance writing performance is to teach self-regulatory procedures that including goal setting in the course of the writing task. These goal setting procedures propose to writers to determine a precise goal to reach during writing and then to direct their attention to elements that help reaching that goals, such as for example the number of ideas to be included in the final text (Ferretti, MacArthur, & Dowdy, 2000; Graham, MacArthur, & Schwartz, 1995) or length of the final text (Hopman & Glynn 1989). Consequently, “goal setting is an important part of good writing” (Page -Voth & Graham, 1999, p. 230), and a clear demand about the text to compose seems to be a key factor of self-regulated behavior to reach the writing goal and thus to reach text quality. In this framework, the present experiment first examined whether a high quality goal has a positive impact on text quality. Skilled writers composed both a narrative and an argumentative text under two goal quality conditions (standard quality or high quality). More precisely, a condition in which a standard quality goal was given to writers, namely without explicit expectation about the level of quality to reach in the produced text, was compared with a condition in which a high quality goal was given to writers (i.e., putting a stress on the