LINGUIST List 17.3666

Mon Dec 11 2006

Diss: Applied Ling: Gebril: 'Independent and Integrated Academic Wr...'

Editor for this issue: Hannah Morales <hannahlinguistlist.org>


Directory         1.    Atta Gebril, Independent and Integrated Academic Writing Tasks: A study in generalizability and test method


Message 1: Independent and Integrated Academic Writing Tasks: A study in generalizability and test method
Date: 10-Dec-2006
From: Atta Gebril <amgebrwm.edu>
Subject: Independent and Integrated Academic Writing Tasks: A study in generalizability and test method


Institution: University of Iowa Program: Foreign Language and ESL Education Dissertation Status: Completed Degree Date: 2006

Author: Atta Gebril

Dissertation Title: Independent and Integrated Academic Writing Tasks: A study in generalizability and test method

Linguistic Field(s): Applied Linguistics
Dissertation Director:
Michael Everson
Dissertation Abstract:

This study investigates the generalizability of reading-to-write andindependent writing tasks. More specifically, the current study addressesthe effect of both tasks and raters on score variability, and investigatesthe relative impact of these two factors. For this purpose, 115 Egyptianuniversity students were randomly selected and assigned to one group. Theseparticipants were instructed to write on two independent writing tasks andtwo reading-to-write tasks. After data collection, three raters who have L2writing experience were selected to rate these writing samples. The studyfollowed a P• × I ◦ × R• design, which is a multivariate generalizabilityanalysis where persons and raters are crossed and tasks are considered as afixed facet. To analyse the data, the multivariate generalizabilityanalysis program mGENOVA (Brennan, 1999) was employed.

Results of the study showed that the reading-to-write tasks yielded asreliable scores as those derived from the independent writing tasks. Inaddition, the multivariate analysis suggested that a composite score ofboth the independent and integrated tasks is as reliable as the univariatescores of either the integrated or the independent writing tasks.Furthermore, the analysis indicated that having different raters score eachtask type would produce as reliable scores as having the same raters scoreboth task types. Additionally, the disattenuated correlation between thereading-to-write and the independent writing tasks was a perfect one. Afinal result indicated that score generalizability is very low when usingone task due to the large (pt) variance component. Implications andlimitations of the study as well as suggestions for further research areprovided.