Proposing a Cognitive EFL Writing Model Based on Personality types and Narrative Writing Intelligence: A SEM Approach

• Main Authors: Sima Ziaei, Behzad Ghonsooly, Zargham Ghabanchi, Hesamoddin Shahriari
• Format: Article
• Language: English
• Published: Allameh Tabataba'i University Press 2019-12-01
• Series: Issues in Language Teaching
• Subjects: big-five personality traits inventory; exploratory factor analysis; l2 writing sem model; narrative writing intelligence; narrative writing intelligence scale
• Online Access: https://ilt.atu.ac.ir/article_11012_958501e2494a1940cd16afdbbde1b6be.pdf

The aim of this paper was to investigate the role of two cognitive factors, namely, personality traits and narrative writing intelligence (NWI) in L2 writing. The total of 416 English learners participated in this study in two different phases. For the purpose of this study, a narrative writing intelligence scale (NWIS) was designed and used to score the writings of the first group of participants which included 200 Iranian learners of English in an EFL institute. The first writing task, which was a film recounting, contributed to designing this scale. Randall’s (1999) definition of narrative intelligence was the guideline to design the primary draft of NWIS. All 200 written film recounting were scored by this scale. The scores were transferred to SPSS 18.0. and Exploratory Factor Analysis was run. NWIS’ construct validity and reliability were confirmed. Moreover, the underlying dimensions causing correlation among the observed variables were reduced in three factors which account for 57.47% of the variance of the scale. The factors were named as Unity of the plot, Identification (of characters, objects, and ideas), and Voice and Rhetoric. The second group, who were female university students of EFL, were given two tasks, namely, writing a memory and filling Big-Five Personality test. Their writings were evaluated twice; once by employing the NWIS and once by a tailor-made writing scoring guide taken from Weigle’s (2002) guideline. Transferring all scores to Amos. 20, a SEM model was proposed by the researchers. The result shows the proposed model has good fit indices.