18 SPRING SPEAKER SERIES Oral Proficiency Interviews as Task-Based Assessment: Task Framing and Interlocutor Influences

Event Date: 
04/12/2018 - 3:30pm to 5:00pm
Location: 
New Cabell Hall 349

 

Steven. J. Ross

University of Maryland

 

Abstract

The oral proficiency interview has been in use for a half century and remains the primary method of assessing proficiency for more than 65 languages taught at US Government agencies.  The interview is a set of unscripted tasks articulated by a trained interviewer and rated by the interviewer and a second rater.   While the configuration of essential tasks varies across agencies, OPI tasks are designed to focus on various facets of proficiency such as conversation about familiar topics, narrations, giving instructions, reporting about current events, descriptions of persons, places, or objects, as well as transactional role plays or role plays with complications.  At higher ranges of proficiency, the tasks include non-routine role plays, supported opinions, discussion of abstract concepts, and tailoring language to the interlocutor. Performance on the various tasks provides evidence of a candidate’s interactional competence, fluency, accuracy, lexical range, coherence, as well as pragma-linguistic and socio-pragmatic abilities. 

The assertion that the interview assesses authentic conversational ability has been a focal point of criticism of the OPI.  The presentation will examine the basis of these critical claims from the perspective that the OPI is a form of institutional discourse, and as such is categorically different from conversational interaction. The presentation will introduce task framing as a crucial condition for candidates to perform with minimal ambiguity, and will make a distinction between off-task transitional talk and task-essential talk that forms the basis of assessment and mapping onto the Interagency Language Roundtable framework. The talk will conclude with a review of the recent and rapid development of speech recognition technology, considering the potential trade-offs associated with non-interactive speech testing and automated scoring.

 

 

Bio Data

 

Steven J. Ross completed the Ph.D. in Second Language Acquisition at the University of Hawai’i Manoa in 1995 and is Professor of Second Language Acquisition at the University of Maryland, where he teaches research methodology and language assessment courses.  He served as LPI Director at the International Institute for Business Communication in Tokyo for five years, where he undertook research on interviews and trained interviewers affiliated with the Test of English for International Communication (TOEIC).