An Exploratory Factor Analysis of the Revision Strategies Questionnaire in an Underachieved ESL Writers’ Context

• Main Authors: Wardatul Akmam Din, Siti Jamilah Bidin
• Format: Article
• Language: English; English
• Published: International Institute for Science, Technology and Education (IISTE) 2014
• Subjects: PN4500 Essays (Literary history); QA299.6-433 Analysis
• Online Access: https://eprints.ums.edu.my/id/eprint/30347/1/An%20Exploratory%20Factor%20Analysis%20of%20the%20Revision%20Strategies%20Questionnaire%20in%20an%20Underachieved%20ESL%20Writers%20ABSTRACT.pdf; https://eprints.ums.edu.my/id/eprint/30347/2/An%20Exploratory%20Factor%20Analysis%20of%20the%20Revision%20Strategies%20Questionnaire%20in%20an%20Underachieved%20ESL%20Writers%E2%80%99%20Context.pdf; https://eprints.ums.edu.my/id/eprint/30347/; https://www.iiste.org/Journals/index.php/DCS/article/view/18186/19091
• Institution: Universiti Malaysia Sabah

Although various aspects of the writing process have been studied extensively of late, research on weak students’ revision strategies has been notably lacking. This paper focuses on the first of three parts involved in this research. It is the scale development stage in which the scales used in the study were validated mainly through confirmatory factor analysis (CFA). Then a model was hypothesized and tested through structural equation modelling approach to reflect the relationships between students’ writing attitude, behaviour and difficulties in ESL writing. Finally an ESL writing strategies training program would be conducted to the participants to explore whether the writing strategies training programme embedded in a ‘normal’ course curriculum would have any impact on the students’ ESL writing performance. For the purpose of this paper, CFA was performed to test the reliability and validity of the constructs, including item loading, construct reliability, and average variance extracted (AVE). CFA was executed via Structural Equation Modelling (SEM) technique utilizing Analysis of Moment Structure (AMOS) computer programme version 21. It is a requirement that item loadings for every factor to exceed 0.50 to be considered as items having sufficient loading values to represent its expected factor (Hair, et al., 2010). There were three sets of CFA that had been examined: (i) writing attitude, (ii) writing behaviour, and (iii) writing difficulties. Results demonstrate that the standardized loadings for each item were above 0.50, Cronbach’s Alpha and composite reliability exceeded 0.70 and AVE values beneath 0.50 (Hair et al., 2010). Hence, the factors are reliable, and have a good convergent validity and reasonable to be used for the further analysis, that is in the structural model.