Background information on the IELTS Speaking Test
The IELTS Speaking Test is a one-on-one interaction between a candidate and an examiner, lasting 11 to 14 minutes It consists of three distinct parts, each serving a specific purpose in terms of interaction, task input, and candidate responses.
! Part 1 (Introduction): candidates answer general questions about themselves, their homes/families, their jobs/studies, their interests, and a range of familiar topic areas
The examiner introduces him/herself and confirms the candidate’s identity The examiner interviews the candidate using verbal questions selected from familiar topic frames
This part lasts between four and five minutes
In Part 2 of the speaking test, candidates receive a verbal prompt on a card and are required to discuss a specific topic They are allotted one minute to prepare their response, after which they must speak for one to two minutes Following the candidate's speech, the examiner will pose one or two concluding questions to further engage the candidate.
! Part 3 (Two-way discussion): the examiner and candidate engage in a discussion of more abstract issues and concepts which are thematically linked to the topic prompt in Part 2
Examiners are provided with comprehensive guidelines to enhance the reliability and validity of the IELTS Speaking Test A key instruction emphasizes that "Standardisation plays a crucial role in the successful management of the IELTS Speaking Test" (Instructions to IELTS Examiners, p 11).
The IELTS Speaking Test is governed by a strict examiner framework that must be adhered to, ensuring consistency across all candidates In Parts 1 and 2, the examiner's script is tightly controlled to provide uniform input, while Part 3 allows for some flexibility to match the candidate's proficiency level Examiners are instructed to avoid unscripted comments and to follow the established guidelines closely Research indicates that candidates commonly engage in various speech functions during the test, including providing personal information, expressing preferences, and summarizing, among others While additional speech functions may arise, they are not mandated by the test structure.
The IELTS bands are defined by detailed performance descriptors that outline spoken performance across nine levels, as outlined in the IELTS Handbook 2005 In 2004, scores were reported as whole bands based on the recorded tests from that year.
Fluency and coherence are essential skills in effective communication, encompassing the ability to speak with natural continuity, appropriate speed, and minimal effort while connecting ideas seamlessly Key indicators of fluency include speech rate and continuity, while coherence is marked by logical sentence sequencing, clear transitions in discussions or arguments, and the use of cohesive devices such as connectors, pronouns, and conjunctions to enhance the flow of ideas.
Lexical Resource encompasses the breadth of vocabulary a candidate employs and the accuracy in conveying meanings and attitudes Key indicators include the diversity of vocabulary, the suitability and relevance of word choices, and the capacity to paraphrase effectively, whether with or without noticeable pauses.
Grammatical Range and Accuracy evaluates a candidate's use of grammatical resources, focusing on sentence length, complexity, and the effective use of subordinate clauses Key indicators include the variety of sentence structures and the ability to rearrange elements for emphasis Additionally, grammatical accuracy is assessed by the frequency of errors in speech and their impact on communication.
Pronunciation is essential for producing clear speech during the Speaking Test, with key indicators including listener strain, unintelligible speech, and the influence of the first language (L1).
In the IELTS Speaking Band descriptors, the concept of "employed" is crucial for distinguishing between different proficiency levels It appears under 'Fluency and coherence', where Band 8 is characterized by the ability to "develop topics coherently and appropriately," while Band 9 demonstrates the capacity to "develop topics fully and appropriately." Additionally, at lower proficiency levels, it is referenced under 'Lexical resource', differentiating Band 3, which struggles with vocabulary for less familiar topics, from Band 4, where candidates can discuss familiar topics but only convey basic meaning on unfamiliar ones.
Research focus and significance
The overall aim is to reveal how topic is developed in the IELTS Speaking Test The main research question is:
1) How is topic developed in the three parts of the Speaking Test?
In answering this question, it is considered how topic as an interactional organisation is related to the overall architecture of interaction in the Speaking Test
Sub-questions are as follows:
This research question focuses on the nuanced differences in topic development among candidates with varying scores: high, mid-range, and low It aims to identify the specific strategies and techniques employed by candidates at each scoring level, highlighting how their approaches to topic development diverge By examining these micro-details, the study seeks to uncover the distinct practices that contribute to successful topic development across different performance tiers.
3) How does the examiner’s interactional style contribute to topic development?
The research literature has identified this as an area for investigation (Brown and Hill, 1998, p 15) This study provides a qualitative investigation of this question
Examiners often adhere closely to the provided briefs regarding the topic, ensuring that evaluations align with the intended guidelines However, when they deviate from these briefs, it can significantly affect the interaction, potentially leading to confusion or misinterpretation of the assessment criteria This divergence may also influence the fairness and consistency of the evaluation process, ultimately impacting the candidates' performance and experience.
A study by Seedhouse and Egbert (2006) revealed that most examiners closely adhere to briefs and instructions However, instances occur where some examiners deviate from these guidelines, inadvertently providing certain candidates with an advantage in their responses While the 2006 research concentrated on turn-taking, sequence, and repair, the current study aims to explore these dynamics in relation to topic management.
5) Do specific topics cause trouble for candidates? Do specific questions within a topic sequence cause trouble for candidates? If so, what is the nature of the trouble?
A study by Seedhouse and Egbert (2006) revealed that the question “Would you like to be in a film?” posed significant challenges for many candidates, indicating a need for deeper exploration of this topic.
Engaging with and developing a topic chosen by the examiner is crucial for achieving a high score in the Speaking Test This study offers the first comprehensive analysis of how candidates succeed or struggle in this aspect.
Methodology
The methodology used in this study is Conversation Analysis (CA), which focuses on the relationship between interaction organization and institutional goals, highlighting how this differs from free conversation CA aims to connect both the overall structure and specific interactional devices to the core objectives of the institution, suggesting that interaction organization is rationally derived from these goals While previous research by Sacks extensively examined topics in the 1960s and 1970s, interest in this aspect has diminished in recent years, as evidenced by Heritage's omission of topic from his six key areas of institutional interaction analysis This study aims to demonstrate the significance of topic in this institutional context and its role in organizing discourse.
The study by Seedhouse and Egbert (2006) examines the IELTS Speaking Test through an institutional discourse lens, revealing that "the organization of turn-taking, sequence, and repair is tightly and rationally organized in relation to the institutional goal of ensuring valid assessment of English speaking proficiency" (p 191) Utilizing Richards and Seedhouse’s (2005) model of “description leading to informed action,” this research connects the analysis of interaction, particularly topic development, to institutional objectives, ultimately proposing informed actions based on data analysis.
Data information
In January 2009, Seedhouse visited UCLES in Cambridge to obtain approximately 306 audio recordings of IELTS Speaking Tests, with Hong Kong data provided on CD The initial phase involved digitizing the audio, completed by February 2009, after which the electronic files were sent to UCLES on DVDs From the raw data, 130 high-quality recordings were selected, as many cassettes were deemed unusable due to poor sound quality or insufficient labeling Sampling was conducted alongside Fiona Barker at UCLES to ensure a diverse representation in the transcripts, considering factors such as gender, geographical region, task/topic number, and Speaking Test band score.
This study investigates how candidates with varying IELTS scores—high, mid-range, and low—develop their topics during the test To facilitate this analysis, specific recordings corresponding to test scores ranging from 9.0 to 4.0 were selected, focusing on five diverse tasks from the Test The data encompasses test centers from numerous countries, including Albania, Bangladesh, Brazil, and more, although details regarding candidates' nationality, first languages, or ethnicity were not provided.
The project advanced to the transcription of digitised audio data to CA standards, with experienced transcriber Andrew Harris as the Research Assistant The researchers prioritized producing high-quality transcripts that captured detailed interactional features, even at the cost of reducing the total number of transcripts This meticulous approach was essential for addressing the research questions, which relied on minute interactional details Unlike Seedhouse’s previous IELTS project in 2006, where less experienced transcribers were involved, this project focused on transcribing high and low-scoring tests, as well as those with unusual features Ultimately, 60 Speaking Tests were fully transcribed, resulting in electronic transcripts (Word format) that are copyright protected.
UCLES All extracts in this report derive from this corpus
This study utilized a data set from recordings made in 2004, supplemented by secondary data including paper materials related to the Speaking Tests on cassette This included examiners' briefs, induction, training, and instruction packs, which are crucial for understanding the institutional goals of the interaction, the examiners' orientations, and specific topic-related issues Notably, Research Question 4 focuses on aligning examiner briefs with their spoken interactions.
Literature review
This research project builds upon existing literature in three key areas: the concept of topic as a focal point for analysis, the exploration of topics from a conversation analysis (CA) perspective, and the limited prior research concerning topics in language proficiency interviews (LPIs), especially in relation to the IELTS examination.
The organization of conversation into topics is a fundamental aspect recognized by both participants and analysts Topic is often viewed as a "metapragmatic folk term," reflecting its technical use while stemming from a common understanding of conversational subjects Essentially, a topic represents the subjects being discussed, with multiple topics comprising the overall content of a conversation While participants may intuitively define a topic as "what is being talked about at any given time," this definition poses challenges for analytical perspectives, particularly in accurately describing and defining the specific topic within a conversation.
The analysis of "holidays" can be categorized into various topics, potentially leading to an endless array of classifications However, the method used in this study adopts a Conversation Analysis (CA) approach, which differs significantly from traditional categorizations and focuses on how participants understand and engage with the topic at hand.
Defining a topic and its analysis poses significant challenges, as noted in various research traditions (Levinson, 1983; Brown and Yule, 1983; Schegloff, 1990) According to Atkinson and Heritage, the complexity of topics makes them one of the most intricate conversational phenomena to study, often resisting systematic analysis (1984, p 165).
Research within the Conversation Analysis (CA) tradition emphasizes the management of topic initiations, shifts, and endings as a collaborative process in real-time interactions, prioritizing the participants' perspectives This approach contrasts with analyses that categorize topics from an external analyst's viewpoint The significance of understanding the participant perspective is highlighted by Sacks’ observation that “the way in which it’s a topic for them is different than the way it’s a topic for anybody else” (1992, p 75).
Conversational analysis (CA) research has identified two distinct methods for managing topics in dialogue (Sacks, 1992; Jefferson, 1984; Button and Casey, 1984) One method, known as stepwise topic transitions, facilitates a seamless flow from one topic to another According to Sacks (1992), "the best way to move from topic to topic is not by a topic close followed by a topic beginning, but by what we call a stepwise move," which involves linking the previous discussion to the new topic, despite their differences (p 566).
Disjunctive or marked topic shifts, as identified by Jefferson (1984) and Sacks (1992), involve explicit markers that signal the transition between topics, often indicating a greater distance between them In the context of LPIs and IELTS interviews, most topic shifts are marked, which Sacks argues can indicate poor conversational quality He posits that effective conversation should not merely consist of distinct blocks of topical discussion, and the frequency of marked topic introductions can reflect a lackluster dialogue Consequently, the reliance on boundaried topic shifts within the IELTS examination may undermine the quality of the conversations it aims to evaluate, particularly when compared to more natural interactions outside of this testing environment.
Early research in conversation analysis (CA) provided valuable insights into how topics are managed in conversation, but interest in this area has significantly declined, likely due to challenges in analyzing topics It is unexpected that the shift towards 'institutional' CA did not spark renewed interest in this aspect of conversation However, recent studies indicate that examining topics in institutional contexts, especially in educational settings, may be more feasible, as topics are frequently set by the educator (Stokoe, 2000).
Research comparing the interactional organization of Language Proficiency Interviews (LPIs) to ordinary conversation has shown that LPIs function as a form of institutional interaction, sharing characteristics with interviews These characteristics include the predetermination of interview actions and a notable interactional asymmetry, which highlights the power dynamics and differences in speaking rights between participants This asymmetry is particularly significant in the context of the interaction.
In the IELTS Test, the examiner holds significant authority over the selection of conversation topics, which are predetermined by the test design As a result, candidates have limited opportunities to introduce their own topics, restricting their ability to showcase their conversational skills for assessment.
The significance of topic selection in Language Proficiency Interviews (LPIs) is crucial, especially regarding the appropriateness, introduction, management, and oversight of these topics However, there is a noticeable lack of research specifically analyzing the role of topicality within these contexts, as highlighted by the work of Kasper and Ross.
Research from 2007 indicates that interviewers frequently pose multiple questions on a single topic, a trend that is also prevalent in the IELTS context, as noted by Lumley and O’Sullivan.
Research has shown a significant link between topic selection and task difficulty in oral proficiency interviews (OPIs), particularly regarding gender bias (2005) Fulcher and Márquez Reiter (2003) also explored this relationship A recent study by Gan et al (2008) highlighted that peer group oral assessments facilitate topic negotiation among participants, resembling natural conversation dynamics However, concerns regarding test reliability and practicality limit the applicability of such assessments in the IELTS context This study aims to fill the existing gap in literature by systematically investigating topic management within the IELTS Test framework.
The characteristics of topic management in the Speaking Test
This section outlines how examiners handle topic shifts during the Speaking Test, highlighting their marking strategies and approaches to following up on candidate responses These aspects are crucial variations that influence the assessment process.
How examiners mark a topic shift
Unmarked topic boundary
The first is the zero option of an unmarked topic boundary, where the examiner moves to the next topic by asking a direct question without any additional interactional work
125 they progress to secondary school or or even university they they
126 er (0.4) they still need to learn english
128 E:! why did you choose to learn english and not another language
130 C: erm:: (1.8) in hong kong i think commercially (0.3) erm english
131 is very useful (0.9) for business and also erm (0.3) in hong kong (005756T130)
In line 128, the examiner simply asks the scripted question without any kind of additional interactional marking.
Generic marking of topic boundary
The second type of Topic Boundary Marker (TBM) is known as generic marking In this approach, the examiner indicates a shift in topic using generic, non-explicit markers, rather than making a clear, explicit statement about the change in topic.
271 have only time to study at night (0.3) hhh i’m still working
272 and this is my (0.8) the most important room in the whole house
275 E: ! okay (.) alright good thank you (0.7) hhh so erm (0.4) d’you
276 spend alot? of time in this room?
278 C: erm i have a erm classes from nine to four? an then i come home
279 (.) usually do the dinner and check my hu- er son’s er homework (000218T135)
In the extract, the examiner acknowledges the candidate's previous statement with "okay," signaling a shift in topic This is followed by markers such as "alright," "good," and "thank you," indicating an impending change After a brief pause, the phrase "so erm" further emphasizes the transition, projecting the introduction of a new topic.
Explicit marking of topic boundary
Explicit topic boundary marking is a key method used by examiners to indicate a shift in topics during assessments In Part 1 of the Test, these markers are primarily dictated by the script, although examiners occasionally deviate from it In the remaining sections of the Test, examiners have the flexibility to create their own explicit topic boundary markers.
129 E: ! m hm (0.3) okay (.) hhh let’s move on to:: talk about
130 !na::mes (0.8) er:: >can you tell me what your< name means?
The examiner adheres to the written script for the topic, utilizing the clear transition phrase "let’s move on to talk about…" Additionally, the examiner introduces a generic marker, "okay", as a pre-sequence to this scripted transition An example of an unscripted, explicit transition phrase follows in the next extract.
385 C: [(cooks all the] (0.3) food and (0.5) [((inaudible))]
388 E: ! okay hhhh erm let’s look into the fu:tur::e (.) what sort of
389 jo::bs do you think will be popular (.) in the future
391 C: m:!:::: (1.5) what sort of jobs? (3) hhhh:: this is
392 very hard to [sa::y ] (0.5) ah:::: (1.5) globally?
In the extract, the examiner employs an unscripted explicit marker, "let's look into the future," to indicate a shift from the present to a hypothetical future scenario This transition can lead to confusion for candidates if not clearly marked, as noted by Seedhouse and Egbert (2004, p 51) To enhance clarity, the use of explicit topic shift markers is advisable, especially for questions requiring a change in perspective.
In the data, there are some instances in which an unmarked topic shift appears to create interactional trouble for candidates, as in the following extract (candidate score of 7.0)
412 C: =i’m just a simple (0.7) person i wanted to live a simple life
414 E: ! i see (.) and are there a lot of people like you?
418 E: are there a lot of people with the same ide[as?]
419 C: [ah ] yes yes er one
420 of my friends (0.4) she is now:: (0.3) she has a family?
The authors recommend that examiners explicitly mark all types of topic shifts during assessments This practice would ensure clear and consistent signaling to candidates, addressing the current inconsistency observed in Parts 2 and 3 of the examination.
Examiner follow-ups to candidate responses
Move onto the next topic question from the script/frame
122 C: erm (.) since kindergarten i think most people in hong kong start
123 learning english (0.2) since kindergarten (0.7) erm (0.2) and
124 (0.7) and erm (1) they learn english continuously erm (0.5) when
125 they progress to secondary school or or even university they they
126 er (0.4) they still need to learn english
128 E: ! why did you choose to learn english and not another language
Demonstrate to the candidate that they expect more of a response
41 E: okay (0.6) so can you tell me some of the other good things
44 C: er the apartment is (.) very comfortable=
46 C: cause it has (0.5) two back doors
49 C: =which is er (0.3) overlooking the gardens i’ve m!a::"de
53 C: and er it has erm (0.3) good (exit? m: it comes there if there
58 C: and er it’s co- convenient for:: (.) cooking:: (0.6) for
59 playing because we have a playground at the ba[ck]
In lines 48, 51, and 56, the examiner uses "m hm" to subtly encourage the candidate to elaborate further This form of back channelling is commonly utilized by examiners to prompt additional responses, and candidates recognize its intent While "m hm" is the typical expression, variations like "m:" can also be observed, as illustrated in the following extract.
111 C: well i usually spend my holidays at "home
Employ a device to seek clarification or expansion on the candidate’s response
Such devices are a) scripted why? questions, b) unscripted why? questions and c) unscripted miscellaneous prompts a) Scripted why? follow-up questions (Part 1)
226 E: sure (.) sure (1) so:: erm do you think names are important?
231 C: names are important because (0.3) you are using your names in
In Part 1 of the exam, scripted why/why not? follow-up questions are essential for encouraging candidates to elaborate on their responses rather than giving simple yes or no answers However, examiners may not consistently pose these follow-up questions, which can affect the depth of the conversation.
78 E: =let’s go on to talk about fr- (.) clothes and fashion (0.4)
79 are you interested in fashion?
82 E: ! [hm m m hm] do you enjoy shopping for clothes?
84 C: ah i do:: (.) i do want and love to:: [erm ] shopping but=
94 E: =m:: m: .hh what kind of clothes do you like
96 C: er:m:: (1.6) i i am fond of wearing:: pants::=
100 E: ! m:: "m:::: (0.6) has the way people feel about clothes and
In the extract above, the examiner does not ask any why/why not? follow-up questions on two separate occasions, even though these are scripted
Examiners should consistently ask scripted follow-up questions, such as "why" or "why not," to encourage candidates to fully develop their responses on specific topics Additionally, incorporating unscripted "why" questions in Parts 2 and 3 can enhance the depth of the conversation.
190 C: m:::::::: (.) probably i think