Testing Talk studies (selected)

Managing test tasks: interaction and assessment (2013b)

In a study using an existing corpus of 199 oral proficiency tests (Sandlund & Sundqvist, 2013b), we examine how test-takers at varying proficiency levels manage problems arising from understanding, negotiating, and completing an oral test task. Using conversation analysis, we examine specifically how student dyads manage so-called task-related trouble (TRT), and the different student strategies were subsequently compared to grades and subskill assessments assigned by three external raters. Findings revealed that while students with the highest scores used the pre-set topics merely as starting points, the mid-scoring students displayed a ‘verbatim’ orientation to the test topics, which meant that more interactional work became necessary when trouble arose. For the low-scoring students, we identified displays of task resistance, and an unwillingness of inability to understand and ‘play the test game’ (cf. Bachman, 1990, on testwiseness). We conclude that aspects of test interaction and task management not visible in the scoring rubrics do seem to play a role in assessment of oral proficiency.

Diverging task understandings (2013a)

A continued focus on interaction can be seen in Sandlund & Sundqvist (2013a), which specifically examines participants’ displayed understandings of the testing context and the oral tasks provided. We focus on sequences in which students and the examining teacher display divergent understandings of the test task, and where the teachers end up working hard interactionally to steer students’ topical talk in particular directions. By rejecting student turns that do not specifically answer a question posed in the pre-set topic, teachers show that particular kinds of topical talk is expected. The students, however, show persistence in pursuing their own topical angles, and the interactional trajectories following upon such displays of divergent understandings are extensive, and do not lead to increased student participation. The study has also been published in popular scientific form in Swedish (Sandlund & Sundqvist, 2012b).

Equity in the assessment of L2 oral proficiency (2016)

Following up on the relationship between patterns of interaction and assessments from a 2011 study, Sandlund & Sundqvist (2016) identify all test-takers in the Testing Talk (N=161) corpus who were assigned grades differing by at least two grade steps (A–F) by at least two of the raters, yielding a sub-corpus of 31 test-takers in 23 test recordings. With a conversation analytic approach, the tests were analyzed sequentially in search of explanations for why particular test-takers were assessed so differently by raters. Findings point to a range of issues that may have tapped into individual raters’ particular preferences, such as the display of appropriate moral stances, the use/non-use of explicit turn allocation devices, and the relative talkativeness of interlocutors.

Testing L2 Talk: a review of research (2016)

In our work with L2 oral tests, we were interested in compiling recent work on L2 oral proficiency testing, which was later published in the form of a review article (Sandund, Sundqvist & Nyroos, 2016b). In the review, we examine international studies on social interaction in L2 oral tests published between 2004 and 2014, and noted a steady increase in empirical studies on paired or small-group testing formats, while oral proficiency interviews (OPIs) still dominate the research base. We also noted that methodologically, conversation analysis is growing as an approach to L2 oral testing, but also that many other studies listed as ‘interactional’ in fact contained little or no interaction data.

Doing versus assessing interactional competence (forthcoming)

The notion of interactional competence (IC) in a second language has become increasingly popular in interaction-based second language research. In Sandlund & Sundqvist (forthc. 2018), we examine how teachers, in collaborative assessment meetings, construct the IC of two test-takers from the 2014 corpus, and contrast these interactional constructions to the test-takers’ interaction (i.e. IC in situ). With an interest in the clash between emic (participant-relevant) and etic (assessment rubrics and teacher discussions) perspectives on IC, we examine instances where students’ actions are treated by assessors as problematic in terms of IC, but where the test interaction shows orientations to local constraints. It is argued that if IC assessment should reflect students’ ability to participate competently in interaction, assessment criteria ought to be informed by analyses of authentic test interaction.

Code-switching and repair (2017)

In this conversation analytic study of code-switching in L2 oral proficiency tests (Nyroos, Sandlund & Sundqvist, 2017), we examine the systematic use of the Swedish conjunction eller in an otherwise all-English local context in test interactions. We examine eller as a repair preface, and analyze the main interactional work performed by this particle. Findings indicate that eller-initiated repair (EIR) functions to ‘hold’ the turn in progress by pushing forward possible turn transition, hence pre-empting potential other-initiated repair, and also works to display the speaker’s awareness regarding interactional trouble. In sum, eller functions as a meta-commentary that signals that modification of the current turn is necessary and in progress - an awareness display that may be particularly relevant in a language testing context. The other-languageness of the repair initiator fortifies the display of such an awareness.

The teacher as examiner of standardized tests (2017)

In Sundqvist, Wikström, Sandlund & Nyroos (2017, advance access), the aim was to examine teachers’/examiners’ practices and views regarding four aspects of the national English speaking test (NEST) – test-taker grouping, recording practices, the actual test occasion, and examiner participation in students’ test interactions – and to discuss findings in relation to issues concerning the normativity and practical feasibility of standardization. Self-report survey data from a national web-based survey of a random sample of teachers (N = 204) and teacher interviews (N = 11) were collected and quantitative data were analyzed using inferential statistics. Findings revealed that despite thorough instructions, teacher practices and views varied greatly across all aspects (further confirmed in interview data), and these variations challenge standardization of NEST. Several questionnaire and interview findings have also been published in popular scientific form, such as Sundqvist et al, 2013, 2014, 2015) in LMS Lingua.

Professional development for teachers – a workshop model (2016)

Our third objective concerned practice, and the development of a workshop model for teachers’ in-service training. In Sandlund, Sundqvist & Nyroos (2016a), we describe our workshop development through the theoretical lens of language teachers’ continuing professional development (CPD) by demonstrating how research on L2 speaking tests can be applied in professional training workshops on oral test interaction and assessment. Workshops were followed up with questionnaires. Findings reveal some noteworthy pedagogical implications of collaborative work with test recordings and assessment that can be implemented at local schools.

Textbook on oral proficiency in English: Forthcoming

Another important outcome of our third research objective is the textbook we are currently writing for Studentlitteratur (Sandlund & Sundqvist, forthcoming). In the textbook (titled Muntlig färdighet i engelska [Oral proficiency in English]), we will give readers (in-service teachers and teacher trainees) a theoretical basis for oral proficiency testing, offer transcripts from our studies, and make practice recommendations for classroom work on oral proficiency.