Writing quality from different latent profiles of revision subprocesses in upper-primary students

Abstract

Text revision is a complex process involving various subprocesses such as error detection, diagnosis and correction. These processes focus on various levels of text, from editing mechanical errors to substantial changes. The present study was designed with two main goals. First, to analyze the existence of homogeneous groups of upper-primary students according to how much they use the different revision subprocesses and their focus, assessed through a specifically created revision task. The distribution of these profiles was analyzed for individual characteristics, such as grade and gender. The second goal was to explore relations between the profiles in terms of text quality. 834 upper-primary students (age 9–13, 4th–6th grade) participated in the study. Students were asked to write a story to assess their narrative writing performance and to revise a prepared narrative text to detect, diagnose and correct six mechanical and six substantive errors. A four-profile model exhibited the best fit, classifying students as poor, mechanical, substantive and good reviewers. A gender effect was observed with more boys than girls in the poor reviewer profile, and more girls than boys in the good reviewer profile, with no effects of gender for the other two profiles. The results also indicated a clear progression in revising skills through schooling, with a higher percentage of poor reviewers in fourth-grade, mechanical reviewers in fifth-grade and good and substantive reviewers in sixth-grade. Finally, a relationship was found between text quality and student reviewer profiles, with poor reviewers writing lower quality texts and good reviewers writing higher quality texts. The identification of different revision profiles in upper-primary students has important theoretical and educational implications.