L2 C level Writing Proficiency: KPG Integrated Tasks and Language Interaction in Rated Scripts

Postgraduate Thesis uoadl:2886670 277 Read counter

Unit:
Κατεύθυνση Γλωσσολογία: θεωρία και εφαρμογές
Library of the School of Philosophy
Deposit date:
2019-12-02
Year:
2019
Author:
Pappa Eleni
Supervisors info:
Βασιλική Μητσικοπούλου, Αναπληρώτρια Καθηγήτρια, Τμήμα Αγγλικής Γλώσσας και Φιλολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Έλλη Υφαντίδου, Καθηγήτρια, Τμήμα Αγγλικής Γλώσσας και Φιλολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Άννα Χατζηδάκη, Λέκτορας, Τμήμα Αγγλικής Γλώσσας και Φιλολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Original Title:
L2 C level Writing Proficiency: KPG Integrated Tasks and Language Interaction in Rated Scripts
Languages:
English
Translated title:
L2 C level Writing Proficiency: KPG Integrated Tasks and Language Interaction in Rated Scripts
Summary:
Research into L2 writing assessment has largely focused on mapping textual features
onto rater-judged candidate performance, exploring issues related to writing quality
and rater reliability.
However, due to issues such as the ambiguous wording of the marking scale
(Lumley, 2002) or the raters’ difficulty to score borderline essays (Gebril & Plakans,
2014), very little has been found with respect to fine distinctions between adjacent
levels of language proficiency, C level (C1-C2) in particular.
In this line, the current research aimed to investigate KPG C level rater-judged
candidate performance in integrated tasks of two types, an intralingual and an
interlingual mediation task. Using a sample of 66 rated scripts (33 candidates), three
points were addressed: a) the effect of two different types of texts, an expository blog
and a narrative encyclopedic entry with an expository task requirement, on language
realization, b) interrater variation and c) cohesion and coherence as a potential
candidate performance differentiating language criterion within C level.
Quantitative analysis results indicate that, first and foremost, Coherence &
Cohesion and, second, Vocabulary Range, can allow for distinctions within C level.
What is more, their accentuated predictive strength when combined with the
Appropriacy criterion can yield a more global (in terms of tasks) account of observed variance. Furthermore, although in sum interrater differences were deemed negligible,
inter-activity variance tied Vocabulary Range with the intralingual mediation task and
Grammaticality/ Accuracy with interlingual mediation.
Qualitative analysis targeting Cohesion & Coherence using Coh-Metrix
suggested that it is mostly surface level features, and word information indices in
particular, that differentiate C1 and C2 situated performance.
These findings can inform current KPG research into task analysis and
ongoing rater training programmes. Additionally, they are in support of KPG’s
practice of applying different gravity status to language criteria depending on the level
of proficiency, while recommending that the task process in question be taken into
consideration as well.
Main subject category:
Language – Literature
Keywords:
KPG, C Level (C1-C2), quantitative analysis, integrated tasks, read-to-write tasks, textual functions, rated scripts, interrater reliability, qualitative analysis, Coh-Metrix, cohesion and coherence
Index:
No
Number of index pages:
0
Contains images:
Yes
Number of references:
77
Number of pages:
123
pappa-dissertation.pdf (2 MB) Open in new window