KPG research
Towards the Validation of the KPG Activities of Oral Performance
Eleftheria Nteliou
PhD Thesis
Faculty of English Language and Literature
School of Philosophy
National and Kapodistrian University of Athens
The test takers’ proficiency in a foreign language is measured through their performance in tasks, designed to examine specific aspects of the construct of language ability, which is thoroughly described in test specifications and influences the construction of the scale of assessment criteria, which discriminate performance from level to level. Given that task design should conform to the test designers’ expectations regarding the test takers’ language ability at a certain level of proficiency, tasks are supposed to elicit specific linguistic features, thus affecting language production.
This study focuses on the oral tasks designed for Activities 2 and 3 in the speaking module of the English KPG exams, at levels B1 and B2. Its aim is to determine how specific task characteristics are expected to influence oral language production, by eliciting particular lexicogrammatical elements, which may differ from level to level. For that purpose, the first stage of the research deals with the linguistic description and analysis of the oral tasks designed for levels B1 and B2, in order to determine the lexicogrammatical characteristics that are expected to be elicited when test takers actually perform the speaking tasks. The theoretical background on which the analytical categories of oral task description are based draws from the systemic functional approach to language use, which also determines the construct of language ability in the KPG test specifications. The second part of the research aims at empirically specifying how the expectations at the oral task design stage are realized in actual test performance, thus providing evidence for the assumptions made in the first stage of the research. For that reason, this part deals with the transcription and discourse analysis of a number of simulated interviews at levels B1 and B2 and attempts to shed light on what really happens during oral production and mediation at these two levels.
Since the results of this study are based on empirical research, they may prove particularly useful for the KPG test designers (as well as for the test designers of other exam systems), because their work will be based on evidence of what B1 and B2 level candidates are really able to do with language, thus eliminating any kind of intuition regarding spoken language potentials at these two levels and leading to the creation of improved tasks. Moreover, this study will conduce to the ongoing validation of the KPG speaking tests, because it will explore the issue of inter-activity variability and will likely lead to a reformulation of the criteria on the assessment scale. Apart from language testing experts, the findings may be helpful to foreign language teachers as well, in their work to prepare students for B1 and B2 level exams as well as in their effort to maximize their students’ oral ability through the use of appropriate tasks at these two levels of proficiency.