Article published in:Writing Assessment in Higher Education
Edited by Bert Weltens, Jos Hornikx, Wander Lowie, Petra Poelmans and Rebecca L. Present-Thomas
[Dutch Journal of Applied Linguistics 2:1] 2013
► pp. 108–125
An investigation into the writing construct(s) measured in Pearson Test of English Academic
Pearson Test of English Academic (PTE Academic) has six item types that assess academic writing either independently or integratively. This research focuses on evaluating the construct validity and effectiveness of the six writing item types. Exploratory Factor Analysis was performed to examine the underlying writing constructs as measured by the six item types. Item scores for different writing skills were subjected to Rasch IRT analysis. The difficulty of the item types was estimated and the effectiveness of each item type was evaluated by calculating the information function of each one. The results identified two writing constructs: an Analytical/Local Writing construct and a Synthetic/Global Writing construct. The study has implications for test developers on the use of multiple item types and their effectiveness, and for test users on how they can improve their writing skills.
Keywords: PTE Academic, item information, item type effectiveness, integrative skills, writing assessment, academic writing, Global Scale of English (GSE)
Published online: 06 May 2013