The longitudinal effects of repeating a timed writing activity on English as a Foreign Language (EFL) students’ second language (L2) writing development were investigated. Data for 46 students in two university classes (23 in each class), each with a different course objective, were collected 30 times in the same way over one year. The students’ compositions were analyzed for fluency, grammatical complexity, and lexical complexity. Text analysis using Coh-Metrix showed that task repetition had an overall effect on L2 writing development. The text analysis was supplemented with a visual analysis using moving min-max graphs. Grammatical complexity developed more prominently than the other aspects of writing in both classes. This counter-predictive result points to the significance of the writers’ reflective consciousness towards their own writing. This study also emphasizes that it is important to study the dynamics in L2 writing development with multi-wave data.
TopIntroduction
This study investigated the longitudinal effects of repeating a timed writing activity on English as a Foreign Language (EFL) students’ second language (L2) writing development. Studies on task repetition have shown that the quality of students’ spoken language improves when the same task is repeated, because learners can pay focal attention to form the second time (Bygate & Samuda, 2005). However, it is unclear what effects repetition of a writing task brings about, assuming that learners are less pressured in writing.
We have clarified a group of learners’ developmental changes and focused on the non-linearity of writing development. We assessed which of the three principal aspects of L2 writing (fluency, grammaticalcomplexity, and lexical complexity) develop in one year. To analyze L2 writers’ texts, we used the web-based computational tool Coh-Metrix (Graesser, McNamara, Louwerse, & Cai, 2004). The development of sophisticated computer programs such as Coh-Metrix coupled with a large corpus has made it possible to identify notable features of L2 texts (e.g., Crossley & McNamara, 2009a; McCarthy, Lehenbauer, Hall, Duran, Fujiwara, & McNamara, 2007). However, the corpora used in previous studies largely consisted of texts written by advanced L2 writers (e.g., high intermediate to advanced writers in Crossley & McNamara, 2009a and L2 scientists whose work had been published in academic journals in McCarthy et al., 2007). The students we focus on in this study are basic writers who had few chances to write in L2 outside the classroom. Their writing is usually short and contains many grammatical errors. Therefore, it is worth inquiring whether Coh-Metrix can capture the changes in such basic writers’ performance. We assessed if any textual features in their writing changed significantly during one year. These changes are further examined with a graphical method.