Assessing Speaking

Abstract

While there have been some significant advances in our understanding of spoken language, in terms of both production and interaction, over the past decades, there remain a number of areas of significant concern to the test writer. These are most notably construct definition (what exactly we are assessing), predictability of task response (task description), the effect of characteristics of the test taker on performance, the effect of characteristics associated with the interlocutor (the person with whom the candidate is interacting) on performance, the appropriateness of the scoring system (i.e., rating scale validity and reliability, and the reliability of the rating process). Before exploring these concerns, it is important to acknowledge the central role in test development of validation by first making explicit the model of validation that underpins this chapter. Thus, the chapter outlines a view of testing speaking that is based on a model of test validation. The chapter also addresses issues in testing speaking such as the test taker (in terms of physical characteristics, psychological characteristics, experiential characteristics, and cognitive aspects), the test system (performance parameters such as live, recorded, and automated formats, linguistic demands, and test administration), and the scoring system (theoretical fit, accuracy of decisions, and value of decisions).

The effectiveness of computer‐assisted pronunciation training: A meta‐analysis

“…Pronunciation is a sub‐component of speaking, in addition to others such as grammar and vocabulary, all of which play a key role in determining the extent to which mutual understanding occurs among communicators. O'Sullivan () postulated that pronunciation is a key aspect of assessing speaking. According to his analytic rubric, language learners with a high level of linguistic competence should display no obvious examples of mispronunciation, while learners with a low level of competence often have unintelligible pronunciation and a heavy accent that may result in the meaning being misunderstood.…”

Assessing the Legal English Skills of Italian Lawyers in a Perspective of Lifelong Learning - The EFLIT End-of-course Exam and University Testing: An Attempt to Cater for Real World Needs

“…32 However, rather than being a reason for discouragement, this has had a positive washforward effect as it has turned out to be a good indicator of what has been achieved, and what needs attention and remedial work. 33 Timemanagement appears to be a major issue in both tests, perhaps due to previous experiences in taking tests, 34 as examinees in Italy are usually granted more time: the written exams taken by high school leavers to obtain their final qualification lasts four to six hours on average, and the exam for admission to the bar lasts eight hours. Moreover, candidates are not allowed a dictionary in the Pre-test or ILEC, whilst they can consult reference texts (e.g.…”

Assessing Speaking Proficiency: A Narrative Review of Speaking Assessment Research Within the Argument-Based Validation Framework

“…This prolonged growth in speaking assessment research warrants a systematic review of major findings that can help subsequent researchers and practitioners to navigate the plethora of published research, or provide them with sound recommendations for future explorations in the speaking assessment domain. Several review or position papers are currently available on speaking assessment, either reviewing the developments in speaking assessment more broadly (e.g., Ginther, 2013;O'Sullivan, 2014;Isaacs, 2016) or examining a specific topic in speaking assessment, such as pronunciation (Isaacs, 2014), rating spoken performance (Winke, 2012) and interactional competence (Galaczi and Taylor, 2018). Needless to say, these papers are valuable in surveying related developments in speaking proficiency assessment and sketching a broad picture of speaking assessment for researchers and practitioners in the field.…”