Investigating second language writing processes
Kellogg's (1996) writing model conceptualizes writing as an interactive process comprising three key sub-processes: formulation, execution, and monitoring The formulation phase involves planning content by retrieving ideas from long-term memory or task input and creating a coherent structure for the text Translating these ideas into linguistic form encompasses three essential sub-processes: lexical retrieval, syntactic encoding, and the expression of cohesion.
In the execution phase, a handwritten or typed text is produced using motor movements
The final stage, monitoring, verifies that the written text accurately represents the writer's intended message When discrepancies arise between the text and the original content plan, L2 writers make necessary revisions The stages of formulation, execution, and monitoring are interlinked, leading to a complex interplay of cognitive processes.
There is substantial amount of research investigating the processes in which L1 writers engage, and the results overall confirm the writing stages outlined in Kellogg's model
Research on L2 writing processes and their connection to writing outcomes is limited, with most studies relying on a single method for analysis For instance, some researchers have focused solely on introspective techniques like the think-aloud method to investigate the cognitive processes of L2 writers (Roca de Larios et al., 2008) Others have depended exclusively on online recordings of keystrokes and mouse movements to gather data on writing processes (Leijten & Van Waes, 2013; Spelman) This narrow approach highlights the need for a more comprehensive methodology that incorporates multiple sources to enhance construct validity in understanding L2 writing.
Miller, Lindgren & Sullivan, 2008 for reviews)
Although these studies have yielded useful insights, there are clear advantages to combining various data sources in investigating writing processes (Leijten & Van Waes,
By integrating keystroke-logging and eye-tracking methodologies, researchers can gain a comprehensive understanding of writing behavior and reading activities of language users This combination allows for the observation of eye movements, which can indicate reasons for pauses, such as re-reading prompts or previously written text While this dual approach enhances the understanding of the writing process and cognitive operations, it does have a limitation: it does not provide direct insights into the conscious cognitive processes of writers during tasks To address this gap, incorporating introspective protocols alongside eye tracking and keystroke-logging could offer deeper insights into writers' cognitive operations.
To date, the combination of keystroke logging, eye tracking, and introspection in L2 writing research and testing remains unexplored Nonetheless, some studies have effectively integrated introspective data with keystroke logging to analyze cognitive processes in L2 writing For instance, Stevenson, Schoonen, and de Glopper (2006) investigated how attention to linguistic elements might hinder higher-level conceptual thinking in secondary school students writing in a foreign language Similarly, Van Weijen (2009) utilized these methods to examine and contrast the cognitive activities involved in L1 and L2 writing, focusing on planning, idea generation, and formulation The integration of these data sources in both studies led to more nuanced and accurate insights into the cognitive behaviors of L2 writers.
Bax (2013) and Brunfaut and McCray (2015) demonstrated that combining eye-tracking and introspective methodologies is effective for investigating the cognitive validity of IELTS and Aptis reading tests in the context of L2 reading.
This research aimed to investigate the cognitive writing processes of test-takers during Task 2 by employing a combination of eye-tracking, keystroke-logging, and introspective methods.
The IELTS Academic Writing Test serves a dual purpose: to enhance our understanding of second language (L2) writers' processing behaviors and to validate the cognitive aspects of the assessment By comparing the test-takers' writing processes to Kellogg's established model of writing, we aim to confirm the assessment's cognitive validity.
The second language writing process and product
The present study aimed to investigate the connections between the writing processes utilized by L2 writers and the resulting quality of their written work This relationship has been extensively examined in research focused on L1 writing, highlighting its significance in understanding the writing development of second language learners.
Despite limited research on the association between revision behavior and text quality in L2 writing, existing studies yield mixed results Stevenson et al (2006) investigated whether the type of revision behavior could predict text quality among 22 secondary school students who wrote essays in both L1 Dutch and L2 English The study hypothesized a negative correlation between lower-level revisions and text quality in L2 writing, suggesting that L2 writers might focus more on lower-level processes, thereby neglecting higher-level cognitive operations like revisions Contrary to this expectation, the findings revealed no significant relationship between the type of revision and the quality of the texts produced.
In their 2008 study, Spelman Miller, Lindgren, and Sullivan explored the relationship between writing behaviors and text quality among high school L2 English writers whose first language was Swedish Over three years, they collected one writing sample per student annually, analyzing fluency through metrics such as words per minute, burst (characters typed between pauses), and fluency during burst (writing time between pauses) Pausing was measured by mean pause length and the proportion of pause time to total writing time, with pauses defined as lasting two seconds or more The researchers also examined the extent of revisions made by the students Text quality was evaluated using weighted subscores for content, grammatical and lexical range, accuracy, and fluency Their findings revealed that burst and fluency during burst were significant predictors of text quality, while pausing and revision behaviors did not significantly influence text quality outcomes.
In conclusion, earlier studies suggest a correlation between text quality and fluency, while showing no significant link to revision or pauses Therefore, additional research is needed to better understand the broader applicability of these findings.
Working memory and second language writing
This research aimed to explore how individual differences in working memory relate to cognitive operations in L2 writing and the quality of the produced text Understanding these relationships can provide insights into the cognitive processes during writing performance As DeKeyser (2012) notes, inferring cognitive processes from the interaction between individual differences and linguistic variables is crucial In language assessment, examining the connections between cognitive abilities like working memory and test performance can enhance cognitive validity Specifically, identifying a relationship between working memory, writing behaviors, and text quality can reveal the cognitive processes engaged by writers during assessments For instance, a correlation between task-switching ability and writing outcomes suggests that test-takers utilized this executive function during the writing task.
The prominent working memory model established by Baddeley and Hitch in 1974 consists of a central executive and two specialized subsystems: the phonological loop and the visual-spatial sketchpad This model was later expanded to include a fourth component known as the episodic buffer, enhancing its complexity and functionality.
The phonological loop temporarily retains and manipulates verbal information, while the visual-spatial sketchpad specializes in storing and handling visual and spatial data The central executive oversees complex cognitive functions, including attention management, processing routine activation, and regulating information flow between short-term storage and long-term memory Additionally, the episodic buffer integrates multi-dimensional information to create cohesive episodes Notably, the phonological loop, visual-spatial sketchpad, and central executive all have limited capacities.
The role of individual differences in working memory capacity has been the subject of a growing number of studies in the field of SLA (e.g., Kormos & Sáfár, 2008; Révész,
2012; see Williams, 2012 for a review) Yet, only a very limited amount of research has looked into the relationship between working memory and second language writing
According to Kormos (2012), it is surprising that there is a lack of focus on the importance of working memory resources in the writing process, as these resources are crucial for the success of various stages Specifically, enhanced phonological short-term memory plays a significant role in effective writing.
(PSTM) span is likely to assist in forming longer and more complex syntactic structures, since PSTM determines the amount of verbal information one can store in memory
Strong visual-spatial short-term memory enhances planning and editing during writing, as it effectively stores visual and spatial information (Kellogg, 1996) Additionally, individuals with a superior central executive are likely to manage the increased demands of parallel processing more effectively when various writing stages, such as planning and typing, occur simultaneously While writing generally involves less time pressure and fewer demands on parallel processing compared to speech production, certain aspects of the writing process still operate in parallel and require careful coordination (Kormos, 2012).
Research indicates a significant connection between working memory and success in second language (L2) writing A study by Kormos and Sáfár (2008) demonstrated that L2 learners' performance in the writing section of the Cambridge First Certificate was influenced by their working memory capacity.
The examination results showed moderate, positive correlations with phonological short-term memory spans, but did not indicate a relationship with complex working memory capacity This finding aligns with the research on bilingual writers conducted by Adams and others.
Guillot (2008) found a notable connection between phonological short-term memory (PSTM) and spelling performance, while complex working memory capacity did not correlate with text quality This highlights the need for further research to clarify how various working memory components influence writing performance Additionally, there is a lack of studies investigating the relationship between different stages of the writing process and working memory, indicating a gap that warrants exploration.
In light of the above, this proposed project intended to investigate the following research questions
1 What is the nature of the cognitive processes in which L2 writers of L1 Mandarin background engage when completing a version of
Task 2 of the IELTS Academic Writing Test?
2 What is the nature of the online writing behaviours which L2 writers of
L1 Mandarin background display when completing a version of Task 2 of the IELTS Academic Writing Test?
3 To what extent is text quality related to cognitive writing processes and online writing behaviours, for a version of Task 2 of the IELTS Academic
Phonological short-term memory, visual short-term memory, and executive control significantly influence online writing behaviors and the quality of text produced in Task 2 of the IELTS Academic Writing Test These cognitive processes play a crucial role in how writers organize their thoughts, maintain focus, and effectively manage language while composing Understanding the relationship between these memory types and writing performance can provide valuable insights for improving writing skills in an academic context.
In the present study, L2 cognitive writing processes were operationalised in terms of participants' stimulated recall comments describing their internal cognitive processes
Online writing behaviors were measured through fluency, pausing, and revision, utilizing keystroke logging and eye movement recordings during pauses Text quality was assessed based on the linguistic complexity of the written content, as analyzed by automated text analysis software.
(b) task response, coherence and cohesion, lexical resource, and grammatical range and accuracy using IELTS rating criteria
Design
Thirty L2 English writers performed a version of Task 2 of the IELTS Academic Writing
Test Their online writing processes were recorded with a Tobii TX60 mobile eye- tracking system and the keystroke logging software Inputlog 6.1.5 (Leijten & Van
In a study conducted by Waes (2013), 12 participants described their thought processes during task performance after a brief break using stimulated recall Additionally, all participants completed a background questionnaire and underwent a series of working memory tests.
Participants
The study involved 30 international students from a UK university, all meeting the IELTS entrance requirement of an overall score of 7.0, indicating a high level of English proficiency To account for the influence of first language, Mandarin Chinese L2 users of English were specifically recruited The majority of participants were female, and their ages varied within a specified range.
18 to 34 years with a mean of 26.60 (SD=3.69) The majority were studying towards a Master's level degree (n$), five students were enrolled in a PhD program, and one participant was completing a Bachelor's degree.
Instruments and procedures
IELTS Test
A computer-based version of the IELTS Academic Writing Test Task 2 was utilized to evaluate second language writing processes, behaviors, and outcomes Students were presented with a specific IELTS essay prompt to address in their writing assessment.
Studying abroad for university can be an exhilarating opportunity, but despite its potential benefits, staying in one's home country may be a more practical choice due to the challenges students face when adapting to a new culture.
To what extent do you agree or disagree with this statement? Give reasons for your answer and include any relevant examples from your knowledge or experience
Stimulated recall procedure
After completing the IELTS writing test, participants engaged in a stimulated recall session where they described their thought processes during the task They were encouraged to pause the recording at any moment to articulate their thoughts, while the researcher also paused the recording during participant revisions or when they revisited previous text The sessions were conducted in English, and participants demonstrated a high level of proficiency, allowing them to effectively communicate their thoughts without difficulty.
Tests of working memory
Three components of Baddeley's (2000) working memory model were assessed: phonological short-term memory, visual short-term memory and executive control
• Phonological short-term memory was assessed by a Mandarin Chinese non-word span (NW) and a Mandarin Chinese digit span test (DS)
• Visual short-term memory was gauged by the Forward Corsi Block (CBF) Task
• Executing functioning was measured using the Backward Corsi Block (CBB),
Operation Span (OSPAN), Colour Shape (CS), and Stop Signal Tasks (SST)
The order of the working memory tests was counterbalanced across participants.
The Chinese non-word span test, based on Zhao (2013), utilized 48 one-syllable non-words that could be pronounced but lacked corresponding Chinese characters These non-words were randomized to create sequences ranging from 2 to 9 items, presented at a rate of one per second Participants were instructed to recall the sequences immediately after hearing the word 'okay,' with three trials conducted for each sequence length The test commenced with a brief practice phase featuring two- and three-non-word sequences, followed by the main assessment Participants' non-word span was determined by the longest sequence they could recall correctly from at least one of the three trials.
The digit span test was also adopted from Zhao (2013), and had a similar design to the non-word span test The test asked participants to recall sequences of 2 to 9 digits
Participants were presented with randomly generated digit sequences ranging from 11 to 99, displayed at a rate of one number per second Each sequence concluded with the word "okay," prompting participants to recall the entire sequence afterward.
The study involved three trials for each sequence length after a brief practice session with two- and three-digit sequences Participants' digit span was assessed by identifying the maximum number of digits they could recall at least once for each sequence length.
The Forward Corsi Block task, utilized to assess visual-spatial short-term memory capacity, was administered through Inquisit Lab 4 During the test, participants viewed patterns of nine blocks on a computer screen, with 2 to 9 blocks highlighted in each trial Participants were then required to click the blocks in the same sequence they were originally highlighted.
The number of the highlighted blocks gradually increased from 2 to 9 There were two trials for each sequence length
The Backward Corsi Block task evaluates the updating function of executive control by requiring participants to click blocks in reverse order from how they were highlighted Unlike the Forward Corsi Block task, this version calculates scores based on the number of trials and the block span, which is the highest number of blocks recalled correctly in at least one trial Research suggests that the total score offers a more reliable measure than block span alone, as it considers performance across both trials of equal length (Kessels et al., 2000, p 254).
4.3.3.4 Automated operation span task (OSPAN)
The updating function of executive control was assessed by the OSPAN test (Turner
In a study utilizing the Inquisit Lab 4 platform, participants engaged in a dual-task involving mathematical operations and memory recall of English letters They solved math problems displayed on the screen, followed by a letter presentation, which continued until they were prompted to recall the letters in the same sequence Set sizes varied from 3 to 7 letters, with three sets for each size presented in random order Participants were instructed to solve the math problems quickly and accurately, aiming for a minimum accuracy rate of 85%, as per traditional scoring methods Their performance was measured using the absolute OSPAN score, calculated based on the number of letters accurately recalled from each set For instance, if a participant correctly recalled all letters in smaller sets but missed some in larger sets, their OSPAN score reflected only the accurately recalled letters, demonstrating their cognitive capacity in multitasking scenarios.
To assess task-switching ability, the colour shape task was utilised (Miyake, Emerson,
In a study by Padilla & Ahn (2004) conducted using Inquisit Lab, participants evaluated colored shapes by deciding on either their color (e.g., green vs red) or shape (e.g., circle vs triangle) In non-switching blocks, participants focused solely on color or shape decisions, while in switching blocks, they made decisions based on a cue letter indicating whether to assess color or shape.
Participants were tasked with identifying colors or shapes based on cue letters, with "C" indicating color (red or green) and "S" indicating shape To analyze the data, reaction times were adjusted to eliminate values outside two standard deviations from the mean The switching cost was calculated by comparing the mean reaction times between non-switching and switching blocks, as referenced in studies by Altgassen et al (2014), Friedman et al (2006), and Gold et al.
The stop signal task, utilized to assess inhibitory control, was incorporated into a series of working memory tests presented through Inquisit Lab During the task, an arrow stimulus appeared on the computer screen, prompting participants to respond by pressing a designated key.
Participants were instructed to press "D" for left-pointing arrows and "K" for right-pointing arrows, but to refrain from responding if an auditory signal (a beep) was present The study measured inhibitory control using the mean reaction time (SSRT) as outlined in previous research (Congdon et al., 2012; Enticott, Ogloff & Bradshaw, 2006).
This index was computed after reaction times (SSRT) were trimmed to two standard deviations above or below the mean.
Data collection
All the participants attended one individual session The session took approximately
Participants in the study spent an average of 2.5 hours for non-stimulated recall and 4 hours for stimulated recall sessions Initially, they reviewed an information sheet and signed a consent form to participate Following this, they filled out a brief background questionnaire The eye-tracking device used was a mobile Tobii X2-60, which boasts a temporal resolution of 60 Hz, and it was calibrated before the sessions began.
23” screen, with the participants sitting about 60cm away from the centre of the screen
A 9-point calibration grid was used to calibrate participants' eyes, and the experiment was presented with the help of Tobii Studio 3.0.9 software (Tobii Technology, n.d.)
After calibrating the eye-tracker, participants wrote an IELTS essay Following a brief break, those involved in the stimulated recall were introduced to the procedure and asked to articulate their thought processes during the essay writing Meanwhile, the other participants completed working memory tests The stimulated recall participants took the working memory tests after a short break post-session.
Data analysis
Analysis of stimulated recall comments
The analysis of the stimulated recall protocols comprised five distinct phases Initially, the comments from the stimulated recall sessions were transcribed Next, a researcher examined the test-takers' remarks to uncover the cognitive processes involved in their writing This review led to the identification of emergent categories, which were subsequently organized into broader categories based on Kellogg's (1996) model of L2 writing.
Tables 1 and 2 for examples of coding categories for pausing and revision respectively)
A subsequent researcher verified the emerging micro-categories and the overarching categories derived from the data, achieving a 97% agreement for micro-category coding and a perfect 100% agreement for identifying general categories Additionally, the comments categorized were totaled to establish a frequency count for each participant.
Table 1: Examples for stimulated recall comments: Pausing
Do I agree or disagree? Which position should I take? Which one is easy to write? Which side is easier to take?
I was thinking what examples I was going to write here What point should I make?
I am thinking what kind difficulties they encounter so I pause and think about difficulties.
While monitoring my word count, I noticed I had reached nearly 250 words, limiting my ability to expand on my argument I recalled using phrases like "first of all," which suggested the need for "secondly" or "furthermore." It became clear that I could only elaborate on one opinion in detail due to the constraints.
I was thinking how to structure the essay I don’t type all the main points for each paragraph I would give different paragraphs for different topics.
Because I've already used the word 'discussions' so I was trying to think of another word which has the same meaning
I wanted to say ‘if not facing the difficulties’ But I didn’t think the expression is precise I wanted to find another expression
Uh I was thinking whether I should treat 'study abroad' as a singular or plural form
Yes, because when I just first thought of using the word 'nationality',
I thought in my own language there would not be any articles
Yeah so, but I think about the grammatical structure in English I may have to add the article
I was thinking about linking words I should use 'Secondly' is boring one Should I use that?
When I was writing this, all the paragraph was in my head So I was thinking how to connect it better.
Mastering the art of effective communication is essential for clear writing By studying various academic papers, I've learned to appreciate the nuances of expression While I sometimes attempt to craft sentences in a more complex manner to achieve a professional tone, I find that simplicity often enhances clarity and impact in my writing.
I had a meaning in my mind that this is very small population of this kind of students it couldn’t represent whole population so I was thinking about wording
MONITORING I want to maybe go back to the beginning and check one time and whether I should include anything.
I finished the last paragraph and I went back to read whole essay.
I review from the beginning checking any grammar mistakes
Table 2: Examples for stimulated recall comments: Revision
I know I wanted to write a personal case of myself So I wanted to start a sentence to bring my case to the essay But later, you can see
Critical thinking is essential in research and academic writing, serving as a pivotal skill that distinguishes effective learners This importance becomes particularly evident when comparing educational approaches in the UK and China, highlighting the role of critical thinking in fostering deeper understanding and analytical abilities in students.
I realized that my typing resembles free writing, which made me recall the structured format required for the IELTS writing task, a test I have never taken Concerned that my informal writing style might not align with the expectations of formal writing, I paused to reflect and subsequently made changes by deleting and revising parts of my text to better fit the necessary structure.
Yeah because I when I was thinking about the second idea
To enhance the article's structure, I plan to clearly divide my ideas into key sentences that outline different aspects The first section will focus on studying, while the second will address living This approach will help the reader follow the content more easily and understand the distinct themes being discussed.
I didn’t want to use 'competitiveness' or 'competence' because I used them before I chose another word 'capacity'
Because I think it is little bit difficult for me to express the meaning of 'transfer' In Chinese, it is transfer but, in this case, if I use 'transfer', I don’t think it is appropriate I used 'overcome' difficulties it will be easier for examiners to understand my meaning
In the former sentence, I think I mentioned two things, first thing is
I have never cooked before and the second thing is I have to think about how much money I spent That means I talk about two things
Maybe I need change into plural
Because when I wrote this sentence, I didn’t notice the tense and
I examined it again and put the past tense
Because I think for the first sentence I used all in singular form but if
Using singular pronouns like 'he' or 'she' in every sentence can make the text feel awkward and less fluid Therefore, considering a shift to plural forms could enhance the overall readability and coherence of the article This change would allow for a more natural flow while maintaining clarity and engagement for the readers.
First, I used 'while' because I wanted to compare in the UK where
I am forced to be independent and in China where I used to depend on parent and friends First, I used 'while' but finally 'but' is a better connection word so I used 'but'
I just I tried to rephrase the sentence to make it more academic
I revise the sentence into I think more proper way but I don’t know it is enough
Actually I was not satisfied with the last sentence I tried to revise it.
Analysis of online writing behaviours
To assess speed fluency, we employed four metrics: total writing time divided by the total number of words or characters, excluding pauses, resulting in minutes per word and characters per word Additionally, we analyzed the number of words or characters produced between pauses.
P-burst and characters per P-burst) The threshold for pausing behaviour was set at two seconds, following conventions in writing research (Wengelin, 2006) Pausing behaviour was expressed in terms of number of pauses and mean length of pauses Pauses were also categorised according to whether they occurred within words, or between words, sentences and paragraphs Revision behaviours, such as deletions and substitutions, were measured by comparing the number of words/characters in the final text before and after the revision Additionally, revisions were classified depending on whether they involved revisions below the word, word, below clause, clause or sentence level.
Analysis of eye-tracking data
To explore participants' online writing behaviors, we combined eye movement recordings with pausing and revision patterns obtained from keystroke logging software (Leijten & Van Waes, 2013) We identified pauses in the Inputlog files using a two-second threshold and aligned these pauses with the eye movement data through Tobii Studio 3.0.9 software Our analysis focused on participants' gaze behaviors during pauses, categorizing their eye movements based on whether they remained within the word, clause, sentence, or paragraph preceding the inscription point.
Occasionally, participants went back to the instructions or did not view the computer screen while they paused, these instances were coded as instruction and off-screen respectively.
Analysis of learner texts
The 30 texts produced by the participants were scored by an IELTS rater in terms of task response, coherence and cohesion, lexical resource, and grammatical range and accuracy, using IELTS rating criteria
The analysis of test-takers' texts focused on linguistic complexity, including lexical, syntactic, and discourse aspects, as well as accuracy According to Jarvis (2013), lexical complexity consists of six sub-constructs: volume, evenness, dispersion, rarity, variability, and disparity In his research on lexical diversity, Jarvis found strong correlations among volume, evenness, and dispersion Consequently, this study evaluated lexical diversity based on rarity, variability, and disparity, as all participants were instructed to produce texts of identical length (Mazgutova & Kormos, 2015).
Using the New General Service List (New-GSL, Brezina & Glabasova, 2013), rarity was expressed as proportion of the most frequent 500 (New-GSL 500), 501-1000
The analysis of the New-GSL 1000 and New-GSL 2500 texts focused on the rarity of words and the prevalence of formulaic expressions We specifically examined the use of formulaic sequences within the 1,000 and 2,000-word texts to understand their impact on language structure.
3,000, 4,000, and 5,000 word frequency bands (K1–K5) with the help of Martinez and
Schmitt’s (2012) Phrase list, which includes the 505 most frequent non-transparent formulae using the British National Corpus as a reference point.
Lexical variability was assessed using Malvern and Richards’ (1997) D-formula and the measure of textual lexical diversity (MTLD; McCarthy & Jarvis, 2010) The value
D is estimated utilising a probabilistic mathematical model which creates a series of randomly sampled tokens to form a type-token ratio curve against increasing token size
MTLD, or Mean Length of T-unit Strings, measures the average length of word strings that meet a specific type-token ratio threshold The D and MTLD indices are calculated using Coh-Metrix, a tool designed for text analysis.
Following Jarvis (2013), disparity was operationalised as a latent semantic analysis
The LSA index, derived from Coh-Metrix 3.0, assesses the conceptual similarity between sentences in essays by examining the semantic overlap of lexical items.
The syntactic complexity of the texts was assessed in terms of three types of indices: complexity by subordination; phrasal complexity; and overall complexity (Norris &
In Ortega's 2009 study, complexity by subordination was quantified using the ratio of clauses to t-units Phrasal complexity was assessed by dividing the total number of words by the number of clauses in each text Additionally, the mean number of complex nominals per t-unit was calculated to further evaluate phrasal complexity Overall complexity was expressed through the ratio of words to t-units and analyzed using the Coh-Metrix tool.
3.0 structural similarity index Except for this measure, all indices were obtained by the program SynLex.
To assess the discourse complexity of the 30 essays, cohesion indices were also obtained with the help of the Coh-Metrix 3.0 program (McNamara et al., 2005)
In particular, the texts were analysed for the use of various types of connectives
Connectives enhance textual cohesion by signaling the relationships between ideas (Halliday & Hasan, 1976) They can be categorized based on the type of cohesion they establish, such as causal (e.g., because), logical (e.g., therefore), additive (e.g., and), or contrastive (e.g., however) relationships.
(Halliday & Hasan, 1976) We employed Coh-Metrix 3.0 to generate an incidence score for these type of connectives
Accuracy was assessed in terms of the number of errors participants produced per
100 words Errors were identified by one of the researchers, and 20% of the data were also double-coded by a native speaker with a background in language teaching
Intercoder agreements was found to be high (91%).
Statistical analyses
Descriptive statistics were computed to address research questions 1 and 2, utilizing categories derived from simulated recall comments alongside data on writing behaviors collected through keystroke-logging and eye-tracking software.
In addressing research questions 3 and 4, Spearman correlational analyses were conducted, with correlations of 25, 40, and 60 categorized as small, medium, and large, respectively, in accordance with Plonsky and Oswald (2014) Due to the extensive number of correlations analyzed, a conservative alpha level of 01 was established.
What is the nature of the cognitive processes in which L2 writers engage?
Table 3 highlights the stimulated recall comments that were gathered to uncover the cognitive processes involved in participants' pausing behavior during the IELTS Academic Writing Task 2 Notably, 48% of these comments pertained to translation processes, while 35% focused on planning operations, and the remaining comments addressed monitoring behaviors.
While the overall trend shows a predominant focus on translation processes among stimulated recall participants, it's noteworthy that four L2 writers demonstrated a different approach These individuals paused more frequently to engage in planning, highlighting the diversity of strategies employed by language learners.
The distribution of planning and translation subprocesses yielded clearer trends
A significant majority of participants (29%) indicated that they paused more often due to content planning rather than organizational planning (6%) In terms of translation challenges, all students highlighted difficulties with lexical retrieval (33%) more frequently than with syntactic encoding (13%), while most reported encountering issues with syntactic coding more often than with cohesion (3%).
Table 3: Reasons for pausing: Summary of stimulated recall comments (N)
Planning Translation Monitoring Don't remember
Con Org Tot Lex Syn Coh Tot
Par = participant, Con = content, Org = organisation, Lex = lexical retrieval, Syn = syntactic encoding,
Coh = cohesion, Tot = total; *Due to rounding some totals do not add up to 100.
Table 4 summarizes the stimulated recall comments that reveal participants' thoughts during the revision process Unlike the findings related to pausing, revision-related comments showed a uniform distribution among participants Notably, all participants mentioned translation-related processes significantly more often (70%) compared to planning mechanisms, which accounted for only 14%.
The trends observed for the sub-processes were similar to those we found for pausing
The stimulated recall comments indicated a greater focus on planning content (15%) compared to planning organization (2%) Participants primarily attributed their revision behaviors to lexical retrieval processes (37%), followed by morphosyntactic construction revisions (23%) and cohesion-related features (10%).
Table 4: Reasons for revision: Summary of stimulated recall comments (N)
Con Org Tot Lex Syn Coh Tot
Par = participant, Con = content, Org = organisation, Lex = lexical retrieval, Syn = syntactic encoding,
Tot = total; *Due to rounding some totals do not add up to 100.
What is the nature of the online writing behaviours which L2 writers display?
Table 5 presents the descriptive statistics for the fluency, pausing, and revision behaviours of the participants First, we consider the fluency indices As shown in
Table 5, participants, on average, produced 20 words and 100 characters per minute excluding pauses (M=.05 min per word, M=.01 min per character), and typed almost
4 words (M=3.75) and more than 20 characters (M 47) between pauses
Turning to pausing behaviours, participants paused for the shortest period within words (M=5.19 s), followed by between words (M=5.34 s) and sentences (M=5.77 s)
Pause length was the longest between paragraphs (M=6.33 s) The majority of pauses occurred between words (M=1.08) Considerably smaller number of pauses were observed within words (M=.10), and between sentences (M=.05) and paragraphs
The analysis of revision behaviors showed that participants retained 79% of the words and 74% of the characters in their final drafts from their total production during the writing process Most revisions occurred at the word level (M=73), with fewer changes made to full words (M=0.07) and smaller units (M=0.97) In contrast, revisions of full clauses (M=3.07) and units longer than clauses (M=2.60) were infrequent.
Table 6 reveals the eye-movement patterns during pauses, showing that participants predominantly did not focus on the screen (M=.11) When they did fixate on the screen, they primarily concentrated within the clause (M=.09) or paragraph (M=.09) The next most common behavior was maintaining eye movement within the sentence (M=.08), followed by fixations on specific words or expressions (M=.07).
(M=.06), or elsewhere on the screen (M=.05)
Table 5: Descriptive statistics for fluency, pausing, and revision behaviours (N0)
Revision by location per 100 words
Table 6: Descriptive statistics for location of eye-gazes per 100 words (N0)
To what extent is text quality related to online writing behaviours?
Relationships between IELTS scores and online writing behaviours
Table 7 displays the descriptive statistics for IELTS scores, revealing that the mean total writing score among participants was approximately 7 (M=6.88) The highest sub-score was recorded in task response (M=7.37), followed by lexical resource (M=6.83), grammatical range and accuracy (M=6.73), and coherence and cohesion (M=6.57).
Table 7: Descriptive statistics for IELTS
Table 8 illustrates the Spearman correlations that evaluate the relationships between IELTS scores and writing behaviors Significant correlations were found in three fluency measures, four regarding pause frequency, and five related to eye-fixation locations Notably, all fluency measures included the minutes per word metric Participants who produced a higher number of words per minute, excluding pauses, received better ratings in task response, lexical resources, and overall scores, with medium effect sizes observed for lexical resources (rho=-.53).
In the IELTS assessment, a significant negative correlation was observed between the frequency of pauses and task response ratings (rho=-.61), indicating that participants who paused more often received lower scores Additionally, those who frequently paused within words had diminished ratings in task response (rho=-.51), lexical resource (rho=-.53), and overall performance (rho=-.50) Furthermore, participants with lower task response ratings tended to pause more frequently between paragraphs, highlighting the impact of pausing on overall scoring.
(rho=-.53) All of these significant relationships for pausing were of medium size
Eye-tracking data indicated that participants who frequently revisited sections of their writing during pauses produced less successful IELTS essays, particularly in terms of task response, with a correlation coefficient of rho=-.50.
In addition, participants looking away from the screen more frequently during pauses produced essays that were rated as less successful in terms of task completion
(rho=-.49), lexical complexity (rho=-.55), accuracy (rho=-.51) and overall quality (rho=-
.60) The strength of these relationships was of medium size, except for a large effect size for the link between number of off-screen eye-gazes and the IELTS total score
Table 8: Spearman correlations between IELTS scores and writing behaviours (N0)
Revision by location per 100 words
** Correlation is significant at the 0.01 level (2-tailed).
Writers exhibiting lower fluency, frequent pauses within words, and off-screen eye movements tended to produce less effective IELTS essays, particularly in task response These factors were linked to decreased lexical complexity and overall lower IELTS scores Additionally, less fluent writing behavior and increased off-screen gazing were predictive of lower overall IELTS performance.
Relationships between linguistic complexity and online writing behaviours
Table 9 presents the descriptive statistics for linguistic complexity measures, revealing that most words in the final texts were drawn from the 500 most frequent words on the new-GSL list, while the essays also incorporated some less frequent off-list words (M=8.58) Additionally, participants utilized formulaic expressions (M=9.47), predominantly sourced from the K1-K3 range according to the Phrase list.
The essays exhibited a high level of lexical variety and significant semantic overlap among words Participants demonstrated syntactically complex language, with an average t-unit length of nearly 18 words, incorporating multiple clauses The texts featured a substantial use of connectives, with additive and logical connectives being particularly prevalent.
Table 9: Descriptive statistics for linguistic complexity (N0)
Table 10 presents the Spearman correlation results between lexical diversity measures and writing behaviors, revealing eight significant relationships Participants who produced fewer words and characters per P-burst utilized a higher proportion of words from the New-GSL 1000 word list, with correlation coefficients of rho=-.56 and rho=-.53, respectively Additionally, those who experienced longer pauses—both overall and within and between words—demonstrated increased usage of New-GSL 1000 words, indicated by correlation values of rho=.49, rho=.48, and rho=.48 Overall, a greater frequency of pausing was positively correlated with the number of New-GSL 1000 words used.
GSL 1000 word list (rho=.56) Amount of sentence-level revision also had a positive relationship with the number of off-list words occurring in the IELTS essays (rho=.55)
Frequent eye movements towards recently typed words during pauses correlated with a greater use of New-GSL 1000 vocabulary in participants' texts (rho=.47) In conclusion, individuals with lower writing fluency tended to take longer and more frequent pauses, while also focusing more on the words they had just written, resulting in a higher inclusion of common vocabulary in their IELTS essays.
On the other hand, infrequent words were more often utilised by writers if they made more higher-level revisions
The Spearman correlations between syntactic complexity measures and writing behaviors revealed five significant relationships Notably, a negative correlation was found between words produced per minute and clauses per t-unit (rho=-.48), suggesting that more fluent writers tend to create IELTS essays with higher clausal complexity Additionally, increased pausing between words correlated positively with the structural similarity index (rho=.53) and negatively with words per t-unit (rho=-.52), indicating that writers who paused frequently utilized less diverse syntactic structures and produced simpler language Furthermore, a significant connection was observed between phrasal complexity and eye fixation frequency; participants who spent more time fixating on instructions during pauses were likely to produce essays with enhanced phrasal complexity.
The study found a significant relationship between syntactic complexity and various viewing behaviors during pauses Specifically, greater syntactic complexity correlated with increased speed fluency and fewer pauses, while participants who frequently looked away from the screen during pauses produced essays with fewer subordinate clauses Overall, these effects were moderate in size, indicating that higher syntactic complexity is linked to more focused attention on instructions and reduced off-screen eye fixations.
Table 12 presents the Spearman correlation results between discourse complexity measures and writing behaviors, revealing six significant relationships Participants who frequently used clausal connectives tended to write fewer words and characters between pauses (rho=-.47) Additionally, a greater use of clausal connectives correlated with increased overall pausing and more frequent pauses within words (rho=.51; rho=.48) Conversely, participants who paused less between words produced a higher number of contrastive connectives (rho=-.51).
The study found a positive correlation between the overall use of connectives and the frequency of participants looking at the instructions during pauses (rho=.58), indicating a medium effect size In summary, increased discourse complexity, as measured by the use of causal connectives, was associated with reduced speed fluency and more frequent pauses Conversely, more complex discourse, operationalized through the use of contrastive connectives, correlated with fewer pauses Additionally, a greater number of gazes at the instructions predicted a more extensive use of connectives.
Table 10: Spearman correlations between lexical diversity and writing behaviours (N0)
Revision by location per 100 words
** Correlation is significant at the 0.01 level (2-tailed).
Table 11: Spearman correlations between syntactic complexity and writing behaviours
Revision by location per 100 words
Table 12: Spearman correlations between discourse complexity and writing behaviours
All connectives Causal connectives Logical connectives Contrastive connectives Additive connectives
Revision by location per 100 words
** Correlation is significant at the 0.01 level (2-tailed).
Relationships between accuracy and online writing behaviours
Table 13 gives the descriptive statistics for our accuracy measure, errors per 100 words
Participants, on average, produced highly accurate texts, they only committed three errors per 100 words (M=.03)
Table 13: Descriptive statistics for accuracy (N0)
Table 14 reveals that the Spearman correlations between accuracy measures and writing behaviors were not significant, suggesting that factors such as fluency, the length and frequency of pauses, and the extent of revisions did not influence the accuracy of participants' texts.
Table 14: Spearman correlations between accuracy and writing behaviours (N0)
Revision by location per 100 words
What is the nature of the relationship of phonological short-term memory, visual short-term memory,
short-term memory, visual short-term memory, and executive control to online writing behaviours and text quality?
Table 15 gives the descriptive statistics for the working memory measures, while
Tables 16 and 17 present the results of the Spearman correlations between the various working memory tests and the indices of writing behaviours and text quality
A study identified three key correlations between working memory skills and writing behaviors Participants with superior task-switching abilities tended to pause for shorter durations between sentences (rho=.59) Additionally, individuals with better information updating skills paused less often between paragraphs (rho=-.51) Conversely, those with less effective visual short-term memory were observed to glance at the instructions more frequently during pauses (rho=-.52).
Table 17 reveals three key correlations between working memory and text quality indices Notably, participants with lower task-switching abilities generated a higher number of New-GSL 1000 words (rho=.46) and utilized more logical connectives (rho=.48) Additionally, a strong relationship was identified between participants' non-word span scores and their use of words from the New-GSL 1000 list.
Those who had better span scores included a larger number of New-GSL 1000 words
Table 15: Descriptive statistics for working memory measures (N0)
Visual-spatial short-term memory
Colour shape task (task-switching ability) (ms)
Stop signal task (inhibitory control) (ms)
Table 16: Spearman correlations between working memory measures and writing behaviours (N0)
NWS DS CBF CBB OSPAN CST SST
** Correlation is significant at the 0.01 level (2-tailed).
NWS = non-word span, DS = digit span, CBF = Corsi block forward, CBB = Corsi block backward,
OSPAN = operation span, CST = colour shape task, SST = stop signal task
What is the nature of the cognitive processes in which L2 writers engage?
This study utilised the stimulated recall procedure to tap into the cognitive processes in which L2 writers engage when performing one version of the IELTS Academic Writing
In a study involving L1 Mandarin participants, individuals were asked to articulate their thoughts during pauses and revisions of their texts Consistent with Kellogg's (1996) writing model, the stimulated recall comments indicated that participants actively engaged in planning, translation, and monitoring processes throughout their writing.
Nearly half of the pauses during the translation process were linked to challenges in lexical retrieval, syntactic encoding, and cohesion Additionally, over a third of comments focused on planning operations, predominantly regarding content planning, while only a small fraction addressed organization Stimulated recall indicated that about 10% of pauses were related to monitoring processes In contrast, a significantly higher number of revision-related comments—70%—were tied to translation processes, emphasizing linguistic encoding Consistent with pausing, participants most frequently revised lexical elements, followed by adjustments to morphosyntactic and cohesive features.
Only 14% of the revision-related comments referred to planning, most of which concerned planning the content of the essay.
The findings indicate that the IELTS Academic Writing Task 2 demonstrates cognitive validity, as the cognitive processes utilized by L2 writers during the task align with those typically employed by L1 writers, as outlined in Kellogg's (1996) established writing model.
Our results also suggest that the cognitive processes of L2 writers completing the
The IELTS Academic Writing Task 2 effectively aligns with the assessment's objectives, as outlined in the IELTS Candidate Guide This guide emphasizes that the primary goal of the academic writing test is to evaluate candidates' capabilities in delivering suitable responses regarding content, organization, and the accuracy and variety of vocabulary and grammar The participants in this study demonstrated cognitive writing processes that reflect these key focus areas.
Table 17: Spearman correlations between working memory and text quality measures
NWS DS CBF CBB OSPAN CST SST
** Correlation is significant at the 0.01 level (2-tailed).
NWS = non-word span, DS = digit span, CBF = Corsi block forward, CBB = Corsi block backward,
OSPAN = operation span, CST = colour shape task, SST = stop signal task
What is the nature of the online writing behaviours which L2 writers display?
Keystroke logging and eye-tracking methodology were employed to examine the online writing behaviours of L2 writers when carrying out the IELTS Academic Writing Task 2
We evaluated various aspects of writing, including speed fluency, pause duration and frequency, overall revision counts by location, and the pattern of eye movements during pauses On average, participants produced 20 words in their writing tasks.
Participants typed at a speed of 100 characters per minute, averaging over 4 words and more than 20 characters between pauses The shortest pauses occurred within words, followed by those between words, sentences, and paragraphs, with most pauses happening between words Ultimately, 79% of the words and 74% of the characters produced were retained in the final draft, indicating that most revisions took place at the word level.
A comparison of our study's results on pausing with those of Spelman Miller (2000) reveals similar trends among L2 writers Both studies found that pause length increased with text level units, with the highest frequency of pauses occurring between intermediate constituents, akin to pauses between words Additionally, pause bursts were observed at nearly 4 words per minute in both studies, which is notably lower than the rate identified for native writers by Spelman Miller.
Similar to the stimulated recall comments, the keystroke logging indices, as well as the eye-gaze data, provide further confirmation of the cognitive validity of the IELTS
According to Kellogg's model of writing, Academic Writing Task 2 requires test-takers to utilize a range of cognitive processes, encompassing both lower- and higher-level writing operations This is evident as participants engage in pauses and revisions across different text levels, indicating their focus on previously produced texts Research indicates that pausing and revising at lower text levels correlates with lower-level writing processes, while similar actions at higher levels are associated with higher-level writing processes (cf Révész).
Kourtali, & Mazgutova, 2017; Stevenson et al., 2006).
To what extent is text quality related to online writing behaviours?
A series of Spearman correlations were performed to explore the relationships between text quality measures and writing behavior indices, revealing several significant connections These findings are summarized in Table 18, organized by different measures of writing behaviors.
Less fluent writing, expressed in terms of minutes per word, was associated with lower
In the IELTS assessment, task response, lexical resource, and overall scores are influenced by lower syntactic complexity and reduced fluency, characterized by fewer words and characters per P-burst This decrease in fluency correlates with diminished lexical complexity, marked by a higher frequency of New-GSL 1000 words and an increased reliance on causal connectives.
Longer and more frequent pauses, both within and between words, are linked to a less sophisticated vocabulary, characterized by a higher use of New-GSL 1000 words Increased pausing correlates with a greater percentage of these basic words in the text, indicating a simpler lexis Additionally, individuals who pause more often tend to use more causal connectives However, extensive pausing within words is associated with lower scores in IELTS task response, lexical resource, and overall performance.
Increased pausing within words also predicted more frequent use of causal connectives
Research indicates that increased pauses between words correlate with a reduced use of contrastive connectives Additionally, more frequent pauses between sentences are associated with lower syntactic complexity, characterized by greater structural similarity and shorter t-units Furthermore, participants who paused more between paragraphs tended to write less effective essays according to IELTS task response criteria Conversely, those who engaged in more revisions at the sentence or higher level produced more sophisticated vocabulary.
Research indicates that participants' gaze during pauses can predict text quality; specifically, those who frequently looked at previously written words tended to produce essays with less sophisticated vocabulary.
Greater number of eye-movements staying within the same paragraph predicted lower
The analysis of IELTS task response scores revealed that reviewing instructions during pauses led to increased phrasal complexity and a higher frequency of connectives used Additionally, it was observed that participants who frequently looked away from the screen tended to achieve lower scores.
IELTS task response, lexical resources, accuracy and total scores they received They also produced less complex sentences with fewer clauses
Recent observations indicate several key trends in writing quality and IELTS scores Firstly, less fluent writing correlates with lower IELTS scores, characterized by less sophisticated language and a higher frequency of causal connectives Secondly, frequent pauses between text units are associated with less sophisticated vocabulary Thirdly, increased pause frequency is a predictor of lower IELTS scores, less sophisticated lexis, reduced syntactic complexity, and a predominance of causal over contrastive connectives Fourthly, a higher incidence of revisions correlates with more sophisticated language use Lastly, looking back at previous text or off-screen during pauses is linked to lower text quality, while revisiting instructions is associated with greater syntactic and discourse complexity.
The findings of this study contrast with those of Stevenson et al (2006), who reported no correlation between revision behaviors and text quality This discrepancy may stem from Stevenson et al using broader criteria for text quality, such as content and language ratings, which may not have been sufficiently nuanced to identify potential connections.
Our research partially corroborates the findings of Spelman Miller et al (2008), demonstrating a positive relationship between fluency and text quality However, we diverged from their conclusions by identifying significant correlations between text quality and certain pausing metrics Similar to Stevenson et al (2006), Spelman Miller et al (2008) utilized a comprehensive measure of text quality, encompassing content, range, complexity, and accuracy, which may explain the differences in outcomes between the two studies.
Table 18: Significant links between writing behaviours and text quality
Writing behaviour Text quality rho
IELTS task response IELTS lexical resources IELTS total
Clause/t-unit New-GSL 1000 Causal connectives New-GSL 1000 Causal connectives
New-GSL 1000 New-GSL 1000 New-GSL 1000
New-GSL 1000 Causal connectives IELTS task response IELTS lexical resources IELTS total
Causal connectives Contrastive connectives Structural similarity Words per t-unit IELTS task response
New-GSL 1000 IELTS task response Words per clause All connectives IELTS task response IELTS lexical resource IELTS accuracy IELTS total Clause per t-unit
To what extent are phonological short-term memory, visual short-term memory, and executive control
visual short-term memory, and executive control related to online writing behaviours and text quality?
Our analysis of the relationship between working memory measures and writing behaviors revealed a few significant correlations Participants with stronger phonological short-term memory generated a greater number of New-GSL 100 words Additionally, those with better visual-spatial span tended to look at the instructions less often during pauses Lastly, individuals with enhanced updating ability exhibited fewer pauses between paragraphs.
Finally, less advanced task-switching skills predicted longer pauses between sentences and the use of less sophisticated lexis and fewer connectives
The findings of this study contradict the positive relationship between phonological short-term memory and text quality identified by Kormos and Sáfár (2008) and Adams and Guillot (2008) However, our results reveal significant positive correlations between executive control and certain measures of text quality.
Variations in study outcomes may stem from the diverse backgrounds of participants and the differing assessments of working memory and text quality used Additional research is essential to better understand the relationship between working memory, writing behaviors, and the quality of written texts.
In addition to the stimulated recall, keystroke-logging, and eye-gaze data, the working memory results supply evidence for the cognitive validity of the IELTS Academic Writing
According to Kellogg's model of writing, phonological short-term memory, the visual-spatial sketchpad, and executive functioning are positively associated with various measures of text quality and writing behaviors This indicates that these components of working memory play a crucial role during the writing process.
Table 19: Significant relationships of working memory measures to writing behaviours and text quality
Working memory measure Writing behaviour/
Non-word span New-GSL 1000 60
Visual-spatial short-term memory
Corsi block forward Eye-fixations at instruction during pauses -.52
Colour shape task (task-switching ability)
Pause frequency between paragraphs Pause length between sentences New-GSL 1000
The results of this study provide evidence from various data sources for the cognitive validity of a version of Task 2 of the IELTS Academic Writing Test Following Field
(2009), we set out to establish cognitive validity by comparing the processes in which
L2 test-takers utilize writing strategies similar to those of native writers when completing real-life writing tasks By referencing Kellogg's (1996) model of first language writing, we found that the cognitive processes observed in Task 2 of the IELTS Academic Writing Test align closely with the stages outlined in this model Additionally, data from stimulated recall comments, keystroke logging, and eye-gaze tracking provide further evidence of this alignment, confirming the relevance of Kellogg's model to the writing behaviors of IELTS test-takers.
Writing Task 2 encourages test-takers to engage in cognitive processes that resemble those that native writers adopt, including both lower and higher-level writing processes
Participants demonstrated similar behaviors to first language writers, frequently pausing and revising at different levels of their texts, as noted in previous research (Stevenson, 2006) Additionally, our findings indicate that elements of working memory outlined in Kellogg's model—specifically phonological short-term memory, the visual-spatial sketchpad, and executive control—play a significant role for L2 users during Task 2 of the IELTS Academic Writing Test.
The findings indicate that the writing processes utilized by test-takers in this study closely resemble those of first language writers when creating written works.
This study presents valuable insights; however, it is limited by the homogeneous backgrounds of its participants, particularly regarding their first language (Mandarin) and proficiency in second language (L2) English This lack of diversity led to minimal variation in text quality, keystroke-logging, and eye-gaze indices, highlighting the need for more varied participant profiles in future research to enhance the findings.
The study's limitations highlight the need for broader research on writing behaviors, working memory, and text quality across diverse L1 backgrounds and proficiency levels Future investigations should involve multiple versions of the IELTS Academic Writing Task 2 and various writing assessments to enhance the generalizability of findings Additionally, collecting stimulated recall data from a larger participant pool would allow for more robust inferential statistics It is also essential to examine how individual differences, such as anxiety, creativity, and personality traits, impact writing processes and outcomes.
Adams, A.M & Guillot, K (2008) Working memory and writing in bilingual students
International Journal of Applied Linguistics, vol 156, pp 13-28.
Altgassen, M., Vetter, N.C., Phillips, L.H., Akgün, C & Kliegel, M (2014) Theory of mind and switching predict prospective memory performance in adolescents
Journal of Experimental Child Psychology, vol 127, pp 163-175.
Baddeley, A.D (2000) The episodic buffer: A new component of working memory?
Trends in Cognitive Sciences, vol 4, pp 417-423.
Baddeley, A.D & Hitch, G.J (1974) Working memory In G Bower (Ed.), The psychology of learning and motivation, (pp 47-90) New York: Academic Press.
Bax, S (2013) The cognitive processing of candidates during reading tests:
Evidence from eye-tracking Language Testing, vol 30, pp 441–465.
Brunfaut, T & McCray, G (2015) Looking into test-takers’ cognitive processes whilst completing reading tasks: A mixed-method eye-tracking and stimulated recall study
ARAGs Research Reports, vol 1, no 1 London: British Council.
Breetvelt, I., Van den Bergh, H & Rijlaarsdam, G (1994) Relations between writing processes and text quality: When and how? Cognition and Instruction, vol 12, pp 103-123.
Brezina, V & Gablasova, D (2013) Is There a Core General Vocabulary? Introducing the
New General Service List Applied Linguistics, vol 36, pp 1-22.
Congdon, E., Mumford, J.A., Cohen, J.R., Galvan, A., Canli, T & Poldrack, R.A (2012)
Measurement and reliability of response inhibition Frontiers in Psychology, vol 3, pp 1-10
Cushing Weigle, S (2002) Assessing writing Cambridge: Cambridge University Press.
DeKeyser, R.M (2012) Interactions between individual differences, treatments, and structures in SLA Language Learning, vol 62, pp 189-200.
Enticott, P.G., Ogloff, J.R & Bradshaw, J.L (2006) Associations between laboratory measures of executive inhibitory control and self-reported impulsivity Personality and
Field, J (2009) The cognitive validity of the lecture listening section of the IELTS listening paper IELTS Research Reports Vol 9 Canberra: IELTS Australia and London:
Field, J (2011) Cognitive validity In L Taylor (Ed.), Examining speaking Cambridge:
Friedman, N.P., Miyake, A., Corley, R.P., Young, S.E., DeFries, J.C & Hewitt, J.K (2006)
Not all executive functions are related to intelligence Psychological science, vol 17, pp 172-179.
Gold, B.T., Kim, C., Johnson, N.F., Kryscio, R.J & Smith, C.D (2013) Lifelong bilingualism maintains neural efficiency for cognitive control in aging The Journal of Neuroscience, vol 33, 387-396.
Halliday, M.A.K & Hasan, R (1976) Cohesion in English London: Longman.
Jarvis, S (2013) Capturing the diversity in lexical diversity Language Learning, vol 63, pp 87-106.
Kellogg, R.T (1996) A model of working memory in writing In C.M Levy & S
Ransdell (Ed.s), The science of writing: Theories, methods, individual differences and applications (pp 57-71) Mahwah, NJ: Lawrence Erlbaum
Kessels, R.P., Van Zandvoort, M.J., Postma, A., Kappelle, L.J & De Haan, E.H
(2000) The Corsi block-tapping task: standardization and normative data Applied
Kormos, J (2012) The role of individual differences in L2 writing Journal of Second
Kormos, J & Sáfár, A (2008) Phonological short-term memory and foreign language performance in intensive language learning Bilingualism: Language and Cognition, vol 11, pp 261-271.
Leijten, M & Van Waes, L (2013) Keystroke logging in writing research: Using Inputlog to analyze and visualize writing processes Written Communication, vol 30, pp 358-392.
Malvern, D & Richards, B (1997) A new measure of lexical diversity In A Ryan &
A Wray (Ed.s), Evolving models of language (pp 58–71) Clevedon, England:
Martinez, R & Schmitt, N (2012) A phrasal expressions list Applied Linguistics, vol 33, pp 299- 320.
Mazgutova, D & Kormos, J (2015) Syntactic and lexical development in an intensive
English for Academic Purposes programme Journal of Second Language Writing, vol 29, pp 3-15.
McCarthy, P.M & Jarvis, S.A (2010) MTLD, vocd-D, and HD-D: A validation study of sophisticated approaches to lexical diversity assessment Behaviour Research Methods, vol 42, pp 381-392.
McNamara, D.S., Louwerse, M.M., Cai, Z & Graesser, A (2005) Coh-Metrix (Version 1.4)
[computer software] Retrieved from http//:cohmetrix.memphis.edu
Miyake et al (2004) explore the role of inner speech as a cognitive retrieval aid for task goals, focusing on how different types of cues and the presence of articulatory suppression influence performance in a random task cuing paradigm Their findings, published in Acta Psychologica, reveal that inner speech significantly enhances task goal retrieval, emphasizing the importance of cue type in cognitive processing The study contributes to our understanding of the mechanisms underlying inner speech and its impact on task execution.
Norris, J.M & Ortega, L (2009) Towards an organic approach to investigating CAF in instructed SLA: The case of complexity Applied Linguistics, vol 30, pp 555-578.
Polio, C (2012) Second language writing In S Gass & A Mackey (Ed.s), Handbook of second language acquisition (pp 319-334) New York: Routledge
Plonsky, L & Oswald, F.L (2014) How big is ‘big’? Interpreting effect sizes in L2 research Language Learning, vol 64, pp 878-912.
Révész, A (2012) Working memory and the observed effectiveness of recasts on different L2 outcome measures Language Learning, vol 62, pp 93–132.