PreprintPDF Available

Exploring the potential of process-tracing technologies to support assessment for learning of L2 writing

Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

Assessment for learning (AfL) seeks to support instruction by providing information about students' current state of learning, the desired end state of learning, and ways to close the gap. AfL of second-language (L2) writing faces challenges insofar as feedback from instructors tends to focus on written products while neglecting most of the processes that gave rise to them, such as planning, formulation, and evaluation. Meanwhile, researchers studying writing processes have been using keystroke logging (KL) and eye-tracking (ET) to analyze and visualize process engagement. This study explores whether such technologies can support more meaningful AfL of L2 writing. Two Chinese L1 students studying at a U.S. university who served as case studies completed a series of argumentative writing tasks while a KL-ET system traced their processes and then produced visualizations that were used for individualized tutoring. Data sources included the visualizations, tutoring-session transcripts, the participants' assessed final essays, and written reflections. Findings showed the technologies, in combination with the assessment dialogues they facilitated, made it possible to (1) position the participants in relation to developmental models of writing; (2) identify and address problems with planning, formulation, and revision; and (3) reveal deep-seated motivational issues that constrained the participants' learning.
No caption available
Content may be subject to copyright.
Exploring the potential of process-tracing technologies to
support assessment for learning of L2 writing
Jim Ranalli
Hui-Hsien Feng
Evgeny Chukharev-Hudilainen
Assessment for learning (AfL) seeks to support instruction by providing information about
students’ current state of learning, the desired end state of learning, and ways to close the gap.
AfL of second-language (L2) writing faces challenges insofar as feedback from instructors tends
to focus on written products while neglecting most of the processes that gave rise to them, such
as planning, formulation, and evaluation. Meanwhile, researchers studying writing processes
have been using keystroke logging (KL) and eye-tracking (ET) to analyze and visualize process
engagement. This study explores whether such technologies can support more meaningful AfL
of L2 writing. Two Chinese L1 students studying at a U.S. university who served as case studies
completed a series of argumentative writing tasks while a KL-ET system traced their processes
and then produced visualizations that were used for individualized tutoring. Data sources
included the visualizations, tutoring-session transcripts, the participants’ assessed final essays,
and written reflections. Findings showed the technologies, in combination with the assessment
dialogues they facilitated, made it possible to (1) position the participants in relation to
developmental models of writing; (2) identify and address problems with planning, formulation,
and revision; and (3) reveal deep-seated motivational issues that constrained the participants’
Keywords: assessment for learning; writing processes; L2 writing; keystroke logging; eye
tracking; process tracing
1. Introduction
Assessment for learning (AfL) is assessment meant to support learning and teaching as opposed
to other types that support sorting, certifying, or accountability functions (Wiliam, 2011). While
AfL has been a central focus of curriculum reform in several countries for more than a decade,
AfL research in L2 writing remains scarce (Lee, 2017). Part of the problem may have to do with
the capacity for AfL, and L2 writing assessment more generally, to take account of writing
processes, such as planning, formulating, and revising.
Research has shown major differences between the ways skilled and unskilled writers engage in
these processes (Roca de Larios, Murphy, & Marín, 2002) as well as connections between
patterns of process engagement and writing quality. In some studies, nearly 80% of the variance
in writing quality was explained by the type and timing of the processes writers engaged in
(Breetvelt, Van den Bergh, & Rijlaarsdam, 1994; Rijlaarsdam & Van den Bergh, 2006). And yet,
most L2 writing instructors probably know very little about the way their students go about
producing the assignments they submit as part of classroom assessment: how much time they
spent, for example, and whether and how they engaged in specific processes. Even if instructors
adopt a “process” approach and assign multiple drafts, these drafts are still written products that
bear little information about the specific processes that went into their creation.
Meanwhile, the tools that researchers have used to study writing processes have been increasing
in power and sophistication. They now include technologies that are largely unobtrusive and
increasingly accessible and affordable such that they could be scaled up for use in instructional
settings. The present study investigated whether and how such process-tracing technologies
might support more meaningful AfL of L2 writing. The study contributes to the existing research
on L2 writing assessment by investigating whether new technologies can expand the focus and
potential benefits of AfL while also addressing calls for assessments that capture more of the
writing construct (Cumming, 2002; Deane, 2013). In line with the theme of this special issue of
Assessing Writing, it also demonstrates unique and significant advantages of computer-based
over paper-based writing.
2. Literature review
2.1 Assessment for learning and feedback in L2 writing
Assessment for learning has been defined as “the process of seeking and interpreting evidence
for use by learners and their teachers to decide where learners are in their learning, where they
need to go, and how best to get there” (Broadfoot et al., 2002, pp. 2-3). While AfL has become
an important component of curriculum reform in Australia, Hong Kong, the U.K., and the U.S.,
it has so far inspired little research in the area of L2 writing (Lee, 2017). Such research as exists
has been based in secondary and tertiary settings in EFL contexts in Hong Kong and Taiwan.
Many of these studies have focused on teachers’ motivations and strategies for implementing
AfL and factors that facilitated or constrained AfL initiatives (Lee, 2011; Lee & Coniam, 2013;
Lee & Falvey, 2014; Mak & Lee, 2014). Another line of research has examined the influence of
AfL on student writing outcomes and student attitudes about AfL innovations (Huang, 2012,
2016; Lee, 2011) such as providing indirect feedback to help learners develop more
independence in error correction.
Feedback is central to AfL because it is the means to convey information about where students’
abilities currently lie in relation to their goals and about the ways they can progress toward those
goals (Black & Wiliam, 1998). Hattie and Timperley’s (2007) influential work on feedback,
often cited in the general AfL literature, posits four types: feedback about the student him or
herself (FS), such as praise; feedback about the task (FT), such as correctness or alignment with
a rubric; feedback about the processing of the task (FP), such as the type of behavior needed to
make improvements; and feedback about self-regulation (FR), such as information to support
self-evaluation. Research shows that, whereas FS is the least effective, FP and FR can contribute
powerfully to deep processing and mastery of tasks. FT, which is the most common type of
feedback, is powerful when used in conjunction with FP and FR, although in practice this rarely
happens (Hattie & Timperley, 2007).
Much of the feedback provided in conventional L2 writing classrooms can be described in terms
of FT. Written corrective feedback (WCF) addressing linguistic issues may support focus on
form (Doughty, 2001) and thus promote learning beyond the current task (although the WCF
research on this question is mixed; see Bitchener & Ferris, 2012). The other common types of
feedback on writing—namely, instructors’ written or oral comments about content, organization,
and other higher-level concernsmay only serve to improve the current text; that is, it may be
too task-specific to generalize to future writing and thus less supportive of learning. FP and FR,
meanwhile, are difficult for writing instructors to address because of the lack of information
about either of these dimensions of students’ task engagement. However, information about
students’ task processing could be gathered by using some of the same methods employed by
writing process researchers (described below).
In AfL, it is vital that assessment information be communicated to students in ways that help
them clarify goals and understand evaluative criteria (Chong, 2017). AfL is informed by
motivation theory that emphasizes learning (i.e., mastery) goals as opposed to performance
goals, and when conducted appropriately, AfL is seen to both support and be supported by
learner motivation (Lee, 2017). One way to ensure that feedback is conveyed in ways that
students can understand and make use of is for instructors to engage in individual “assessment
dialogues” with students (Carless, 2006), which can be mutually beneficial insofar as they may
also help instructors synthesize and interpret assessment information (Chong, 2017).
2.2 Processes in writing theory and research
Processes are foundational to the major cognitive models of writing (Bereiter & Scardamalia,
1987; Flower & Hayes, 1981; Hayes, 2012; Kellogg, 2008; Leijten, Van Waes, Schriver, &
Hayes, 2014), including the pair of models proposed by Bereiter and Scardamalia (1987) to
account for differences between novice and skilled writers, which, although based in L1
research, has informed theorizing about L2 writing (Weigle, 2002). In this distinction, the
knowledge-telling model describes the novice approach, wherein writing is a simple act of
retrieving information from memory and “telling” what you know. By contrast, skilled writers
exhibit knowledge-transforming, which involves an interaction between the author’s mental
representation of her ideas and a separate mental representation of the text, with discrepancies
between the two occasioning problem-solving and rethinking of the original ideas. (Kellogg
[2008] expanded this dichotomous conceptualization by adding a third, top-level model,
knowledge-crafting, which characterizes the work of professional writers.) According to Bereiter
and Scardamalia (1987), it is not possible to identify the underlying approach simply by studying
a given text because knowledge of topic, genre, and language affects writing outcomes. Rather, it
is cognitive processes and the different ways they are enacted that make such differentiation
possible. These processes include:
2.2.1 Planning
Planning typically encompasses generating ideas and organizing them into a structure. Unskilled
writers tend to forego planning whereas skilled writers use advanced (i.e., before writing) or
emergent (i.e., while writing) forms of planning or both (Cumming, 1989; Victori, 1999).
Advanced planning has been found to ease the attentional burden on unskilled writers (Kellogg,
1988), and in a study involving timed writing tests, it corresponded with higher writing quality
(Worden, 2009).
2.2.2 Formulation
Also referred to as translation, this process involves the conversion of ideas into language. In L2
writing and among novice writers, this is typically the process to which the most time is devoted,
but as writing skills develop, other skills such as planning and revision increase in proportion
(Roca de Larios et al., 2008). Writing-skills development is associated with an increase in the
interruption of formulation by other processes (e.g., planning, evaluating, revising), which can be
interpreted as recursiveness indicative of problem-solving (Roca de Larios, Marín, & Murphy,
2001). This should be distinguished from the simultaneous enactment of formulation and
planning (or other processes), which in novices can easily lead to attentional overload (Kellogg,
1988). Some research shows writers process both form and concepts more when formulating in
an L2 than when writing in L1 because of the increased cognitive effort from working in a
language that is not yet automatized (Lindgren & Sullivan, 2006).
2.2.3 Evaluation
Evaluation involves rereading the text produced so far to determine whether it needs
modification. The process is understudied compared to its companion process, revision.
However, eye-tracking can facilitate greater research focus on this process (Wengelin et al.,
2009). Evaluation is worth considering in its own right since positive evaluations that do not
result in revisions may still be of interest, and since revisions may proceed from different sources
of evaluation (oneself, instructors, peers, or automated analyses). In addition, evaluations that
determine a need for major changes may give rise to redrafting (i.e., formulating parts of the text
anew) as opposed to revision, which in such cases is more cognitively demanding (Flower et al.,
1986). In a study comparing distributions of writing processes across test and non-test
conditions, time allocated to evaluation in the latter stages of writing was found to constitute the
only significant difference, with twice as much evaluation observed in the non-test condition
(Khuder & Harwood, 2015).
2.2.4 Revision
This well-researched process involves modifying previously written text. Research shows
unskilled writers in general focus more on local revision (i.e., changes at sentence level) whereas
skilled writers engage in both local and global revision (i.e., changes involving larger units of
discourse). Among L2 writers, aversion to global revision has been attributed to fear of
confronting linguistic issues they cannot remediate (Uzawa, 1996). While unskilled L2 writers
may engage in local revision randomly, skilled L2 writers tend to leave local revision until later
(Zamel, 1983). Writers working in an L2 have been found to make more pre-contextual revisions
(i.e., revisions at the point of inscription) involving both form and concept than when writing in
their L1 (Lindgren & Sullivan, 2006).
2.2.5 Task definition
Task definition, also known as task representation, is the process whereby a writer forms and
refines her understanding of a task’s requirements. This process can be conceptualized at
multiple levels, from a particular assignment to a genre to a mental model of academic writing
(Nicolás-Conesa, 2012). A study of EFL student writers at the beginning and end of a nine-
month writing program found that they maintained conceptualizations that were largely product-
based, although some by the end showed evidence of a process orientation as well (Nicolás-
Conesa, Roca de Larios, & Coyle, 2014).
One important underlying theme in L2 writing process research is the greater complexity that
writing tasks take on because attentional capacity must be devoted to language processing that
could otherwise be devoted to higher-level concerns, such as content and rhetoric (Roca de
Larios et al., 2002; Spelman Miller, Lindgren, & Sullivan, 2008). As the above research
summary indicates, student writers have been observed to engage in the same processes
differently depending on whether they are writing in a first (i.e., automatized) or second (i.e., not
yet automatized) language. Yet L2 writers have also been shown to exercise agency, reusing and
reshaping their L1 knowledge and experience differently depending on the task and audience,
albeit with these choices being influenced by L2 proficiency and amount of L1/L2 writing
experience (Rinnert, Kobayashi, & Katayama, 2015). This suggests efforts to develop
interventions aimed at L2 writing processes must (1) prioritize helping L2 student writers
manage cognitive load; and (2) adopt an individualized approach based on assessing the needs of
particular students, which is consistent with AfL theory.
Finally, we note that the process research summarized above has been accomplished over time
using a variety of methods. Older methods included direct observation, observation via
videotaping, and thinkaloud protocols (which are still frequently used) while newer methods
include keystroke logging and eye-tracking. Whereas the former are clearly impractical for
classroom applications, the same cannot be said for the latter, hence the motivation for the
present study.
2.3 Process tracing in writing research
As the preceding section suggests, relatively little research in the field of writing assessment has
investigated cognitive processes. This can be attributed in part to the historically predominant
form in the field: the timed, impromptu writing test, in which recursiveness, global revision, and
other important aspects of writing-as-process can play little part (Wolcott, 1987). Nevertheless,
there have been repeated calls for assessments that can measure processes (Cho, 2003; Deane,
2013; Graham, Harris, & Hebert, 2011; O'Brien, 1992; Wolcott, 1987) and thus expand the
coverage of the construct of writing ability (Deane, 2013). To achieve this, recent, large-scale
standardized assessments have experimented with keystroke logging (Almond et al., 2012;
National Assessment of Educational Progress, 2011).
At the same time, writing process (WP) researchers have been using different types of
visualization to analyze keystroke logging (and in some cases eye-tracking) data and report their
findings, which is necessary because the huge data sets that these technologies produce make it
impossible for unaided human faculties to identify patterns and regularities (Bécotte et al., 2015).
Despite their varied forms, WP visualizations typically convey information about the timing and
location of writing behaviors of interest. They include written text marked up with special
symbols, known as S-notation (Severinson Eklundh & Kollberg, 1996); timelines combining
keystroke and eye-tracking information (Wengelin et al., 2009), progression diagrams showing
number of revisions as a function of both timing and location (Perrin, 2003); dynamic, GIS-
based representations (Lindgren et al., 2007); and graph visualizations resembling network
diagrams (Caporossi & Leblay, 2011). One issue in the design of WP visualizations is how to
achieve sufficient amounts of macro- and micro-level detail so as to support analysis of larger
trends (e.g., to see how much time was allocated to formulation versus revision) while also
allowing text-level views (e.g., to see what type of revision was made at a particular point in
time). The solution used in the current study was to dynamically link a linear graph inspired by
those in Lindgren and Sullivan (2002) and Leijten and Van Waes (2013) with a replay
visualization of character-by-character reconstruction of texts like those described in Severinson
Eklundh and Kollberg (1996) and Leijten and Van Waes (2013). (See description in Methods.)
To date, it appears only a single line of research has used process-tracing technologies to provide
formative feedback on L2 writing. Lindgren and Sullivan used keystroke logs to generate a
visual playback of the entire production of student texts, which served as stimuli for self- and
peer assessment of composing behavior. In two small-scale studies (Lindgren, Stevenson, &
Sullivan, 2008; Sullivan & Lindgren, 2002), participants who watched replays of their own and a
peer’s composing reported gaining unique insights into their own writing habits, such as a
tendency to repeatedly revise a certain passage and, as a result, lose the thread of the larger
argument. No follow-up was conducted, however, to see if these insights resulted in changes to
the behaviors identified by the participants. Also, 30-minute writing tasks were used, which
made it feasible to derive insights from a linear viewing of the replay; this might not be the case
with longer replays of untimed writing.
2.4 The current study
The following research question guided the study: Do process-tracing technologies provide
insights about L2 student writing that can enhance instructors’ efforts to create more meaningful
assessment for learning? We focus on the role of instructors because student involvement in AfL,
sometimes referred to as Assessment as Learning (AaL), requires attention to metacognitive and
self-regulatory abilities that went beyond the scope of this exploratory study (although AaL
would be a next logical step). In addressing this research question, an underlying aim was to
produce information that could inform future attempts to scale up this approach were it to prove
3. Methods
3.1 Context
The study took place at a large research university in the Midwestern U.S. Because of constraints
imposed by the institutional review board, the research could not be integrated into ongoing
writing classes, so we recruited students from these classes and had them complete writing tasks
outside of their normal coursework. Specifically, we recruited from two sequenced,
developmental writing courses to which international students are assigned on the basis of an
English placement test. The lower-level course focuses on academic writing at the sentence and
paragraph levels while the higher-level course assigns essay-length tasks.
3.2 Participants
Of 14 participants recruited for the study, six completed all four writing tasks. From those six,
we selected two using purposive sampling; specifically, sequential sampling to identify
confirming and disconfirming cases (Teddlie & Tashakkori, 2008). Both participants were
Mandarin L1 speakers from mainland China. Other biodata is summarized in Table 1. The
participants were paid $15 an hour.
3.3 Writing and process-tracing tools
Writing and process tracing were conducted in a web-based tool called CyWrite, which was
developed by the authors. The CyWrite system features a text editor that provides a familiar
word-processing experience while also permitting capturing of the process of composition with
combined keystroke logging and eye tracking. As the user composes text, the CyWrite editor
unobtrusively records time-aligned logs of keystrokes, text changes, and eye fixations. Eye
tracking is performed by a low-cost device mounted under the computer screen. The editor
interfaces with the eye tracker via a protocol that provides a real-time feed of eye-fixation
The three logs are streamed live to a server where they are analyzed and persistently stored. The
logged events are then rendered in a post-session viewer (Figure 1), in which user activity is
reconstructed in a visualization called playback that resembles high-fidelity screen-capture
recordings with an overlaid gaze-point marker. Above the playback area, another visualization
called the process graph is generated, containing variables measured in characters on the Y-scale
and time measured in minutes on the X-scale. The plotted variables are: total number of typed
characters, including deleted ones (“process” in blue); total length of text (“product” in green);
offset in text of the character rendered in the top-left corner of the viewport (“scrolling” in pink);
offset in text of the fixated character (“fixation” in yellow); and offset in text of the cursor
position (“cursor” in red). Gaps in the plotted lines indicate periods when the writer has switched
focus to another window in the operating system, such as an online dictionary page, or when the
eye tracker was being recalibrated. The process graph and playback features are dynamically
linked such that moving the playhead to a point in the graph will show what was happening in
playback at that moment.
Figure 1: Post-session viewer showing process graph above and playback below
3.4 Materials and measures
Four writing tasks, adapted from prompts created by the Educational Testing Service, were
assigned. We chose argumentative tasks since these elicit more complex interplays of processes
(Roca de Larios et al., 2002) compared to narrative prompts, for example. The topics, which
addressed generation-gap disputes over traditions, controversial trends and fads, computers and
privilege, and the value of gap years, are provided in Appendix A. Each task required
participants to write an essay of at least 400 words for a specified audience (faculty and students
at the site university) and purpose (to persuade readers to “see the issue the same way as you”).
The participants were encouraged to take as much time and as many writing sessions as needed
to complete it to their satisfaction. So that we could capture as much information about their
writing processes as possible, the participants were told to do all of their planning, formulating,
and revising in the CyWrite files prepared for them. In addition, the prompt for each task was
displayed in copyable but uneditable text at the top of the CyWrite file so that participants'
interactions with the prompt could also be captured.
Along with the essays, the participants also wrote four reflection tasks (Appendix B), two at the
beginning of the project and two at the end, which were adapted from Nicolás-Conesa et al.
(2014). One pair of these reflections probed students’ mental models of academic writing while a
second pair probed a variable unrelated to the current study, hence its exclusion here.
3.5 Analyses
In addition to the automated analyses that produced the process visualizations, several other
analyses were performed automatically by the CyWrite system. Summary statistics were
calculated for time spent writing per session, number of characters typed and deleted, and word
The completed essays were exported from CyWrite into MS Word format and then evaluated
using the Comment and Track Changes functions by the first author, who also acted as writing
instructor. The evaluated essays, transcripts of the follow-up sessions, the pre- and post-project
reflections, and the sheets on which the participants’ process goals were assigned, were imported
into the qualitative software nVivo for coding and analysis. The process-trace data was screen-
recorded, segmented, and coded using TechSmith Morae, a usability research tool.
Four separate coding schemes were used for the analysis of these different forms of data
(Appendix C). The reflection-data codes were adapted from Nicolás-Conesa et al. (2014). The
coding schemes for the follow-up session transcripts and process-viewer data were developed
using a combined top-down and bottom-up approach, based on categories in Roca de Larios et
al. (2008) as well as new categories identified during repeated cycles of reviewing the data with
reference to the research question and previous literature. The coding scheme for the evaluated
essays was based on assessment features from the International English Language Testing
System (IELTS) rubrics
and the first and second authors’ combined experience in ESL writing
For all but one of the data types, some data was reserved for calibration and final refinements of
the coding scheme; the remaining data were then annotated by a second coder. The total number
of annotations, the percentage of annotations performed by the second coder, and reliability
coefficients, are provided in Table 2. Reliability was excellent across the data sources based on
the scale interpretations in Strijbos and Stahl (2007).
3.6 Procedures
The study began in the second half of the Fall 2016 semester and continued through Spring 2017.
Participants started by completing the questionnaire and initial reflection tasks. They then wrote
their first writing task in a computer lab on machines equipped with the CyWrite software and
eye trackers. With the help of project staff, they scheduled their writing sessions on days and
times that were convenient for them. Once they completed a task, a tutoring session with the first
author—referred to as a “follow-up session” to reflect its formative-assessment focuswould be
In the follow-up sessions, which lasted 60-90 minutes, the participant and instructor would first
discuss the finished essay, as would happen in a typical writing conference. The discussion
would then turn to writing processes, with the participant and instructor reviewing the process
Each essay was also rated using the public version of the IELTS Task 2 writing band descriptors
( This analytic
rating scale was used in place of the TOEFL holistic scale selected because its four categories (task achievement,
coherence and cohesion, lexical resource, and grammatical range and accuracy) provide more formative
information to learners and because IELTS Task 2 prompts resemble TOEFL prompts in their argumentative focus
and in the types of response they elicit. We attempted to validate the writing quality ratings assigned by the first
author as part of feedback, but variation in writing quality among the essays was below the sensitivity threshold of
the instrument. On a scale of 0-9, the average rating was 5.43, with a standard deviation of only 0.71. Further, when
the first and the second authors independently rated the essays, they only achieved reliability of αinterval = 0.27
despite appropriate calibration. Thus, although the quality ratings were useful for providing feedback to participants,
they were not reliable enough to incorporate into the analyses reported here.
data. (The first follow-up session was spent familiarizing the participant with the process graph
and playback capabilities, the writing processes, and practicing graph interpretation). Depending
on the outcome of the product and process reviews, the session might then include (1)
observation of process data for experienced “model” writers who had completed the same
writing task, (2) instruction in strategies for approaching a particular process or decoupling
processes, or (3) both. The sessions would then end with the instructor assigning process goals
for the next writing task that addressed the major problem areas pointed out during the session. A
sheet summarizing these goals would be given to the participant to have in front of her whenever
she was working on the next writing task.
Upon completing the final writing task and follow-up session, participants wrote their final
reflections and completed an exit survey.
4. Results
A summary of the participants’ process and product information across the four writing tasks is
provided in Table 3. This data is also referenced in the individual narratives that follow.
In these narratives, attributions are made to specific data sources where relevant by means of
codes (e.g., FS1 = the first follow-up session, RT2 = the second reflection task, WT3 = the third
writing task).
4.1 Zedong: Development constrained by limited motivation
Zedong came to the project with little academic writing experience, claiming never to have
written an essay of more than 300 words in English, and this only for exam-preparation
purposes. He described his primary motivation for participating as financial although he said he
also wished to improve his writing. A freshman studying computer science and assigned to the
lower-level writing course, Zedong was living and studying overseas for the first time. He said
he hoped to work and study further in the U.S., but he believed writing in English would not be
essential for his future career. The only form of writing he enjoyed was computer coding; other
types were “not that interesting” (FS4). He had scored a 6.5 overall on the IELTS test, which
translates to a TOEFL score of between 79-93, and which was one step above the minimum
score needed to enter the university.
His essay in response to WT1 consisted of a single-paragraph of almost the exact word length
specified in the assignment. In it, Zedong had strayed from the prompt. Rather than arguing
whether older people were justified in getting upset when younger people flouted traditions, the
text focused on negative consequences of an annual spring festival in his hometown (e.g.,
pollution, traffic). The composition was also critiqued for non-sequiturs, uneven support for
main ideas, and a statement at the end that seemed to contradict the thesis he had laid out at the
Analysis of the process data showed Zedong had completed the task in a single session lasting
about 90 minutes (Figure 2), despite encouragement to take as long as he needed and to spread
the work over at least two sessions. He had begun formulating after less than four minutes of
reading the prompt and in the absence of any external plan. He said he had come up with a few
ideas with which to get started and had expected more to occur to him as he wrote (FS1).
His need to generate ideas while formulating contributed to a fractured and slow overall process.
Figure 2 shows a shallow upward slope of the green product line, as well as numerous gaps
indicating Google and dictionary searches to find words to express his ideas or to check spelling.
Several instances of “flatlining” (i.e., stretches with no slope and no apparent keyboarding or re-
reading) are also evident, during which Zedong said he was “just thinking about what to write
next” (FS1).
Figure 2. Process-graph from Zedong’s first and only writing session in response to WT1.
In terms of revision, the process data showed only a brief (less than one minute) stage near the
end, during which he made two changes to the second sentence of the text: adding a second
headword to create a compound noun phrase and substituting the word “disappointed” for “sad”
to describe the feeling of losing one’s traditions. No rereading of the entire text was evident.
Asked to reconcile this with his project-initial reflection, in which he stated the importance of
revision and re-reading “to see is the logic make sense to the readers” (RT1), he laughed and
admitted: “I didn’t revision so much. Just I don’t have the habit to. I don’t like to revise” (FS1).
Our goals for Zedong thus focused on helping him transition from a linear, knowledge-telling
approach to a more recursive, knowledge-transforming approach including a fuller complement
of processes. Zedong was told to include an external planning stage at the beginning, rereading
of the prompt at regular intervals to ensure he stayed on track, and at least 20 minutes of
evaluation and revision at the end, with a global focus preceding a local focus. We also
suggested a “placeholder strategy” in which he was to write an L1 word, or his best
approximation of the spelling of an L2 word he was unsure about so that he could continue
formulating without losing his train of thought. These items, annotated with a symbol such as
“##” to make finding them again easier, could be looked up after formulation during a dedicated
resourcing or revision stage. One rationale for these goals was to allow Zedong to
compartmentalize processes so that he was doing only one thing at a time and thus avoiding
attentional overload.
In terms of planning, he progressed from simple plans consisting of a few main points (WT2) to
a longer outline that broke down his argument into "benefits" and "deficits" (WT4). His planning
technique became more interactive, but he also had a habit of dismantling his plan as he wrote,
cutting and pasting portions into the text proper to convert to parts of the essay so that the
complete plan was no longer available for purposes of further refinement or comparison during
evaluation. Despite this, he also showed increasing propensity to compartmentalize formulation,
planning, and resourcing (with the help of the placeholder strategy, which he used a great deal),
facilitating a more productive overall process with less flatlining.
Evaluation and revision were more challenging for Zedong. For WT2, he had to be encouraged
by project staff to return for a second session in which he could evaluate and review his work.
This session (Figure 3) lasted 19 minutes. The cursor line indicates two passes through the text,
which playback showed to consist of local changes.
Figure 3. Process graph from the second of Zedong’s two writing sessions for WT2.
To address this, we showed him process data of a model writer engaging in global evaluation and
discussed how the changes benefited the final product. We also provided specific questions to
focus his global evaluation on issues of task response, content, organization, and audience
awareness (Appendix D). Finally, we emphasized that writing is time-consuming even for skilled
writers working in their mother tongue and that spending more time would be a gainful strategy.
However, Zedong’s process data for WT3 and WT4 showed similarly abbreviated evaluation and
revision stages with predominantly local changes in evidence. Asked about this in the follow-up
sessions, Zedong said he was avoiding global evaluation of a complete draft because he feared
the amount of work required if he were to find major problems. He preferred instead to focus on
smaller chunks of text because these were “flexible [enough] to make changes” (FS2).
The lack of global-level evaluation and revision was particularly reflected in the critiques of
coherence problems on Zedong’s submission for WT4. Although he was better addressing the
prompt and generating better ideas, the way he organized and connected them in the essay still
fell short, and dealing with these problems required scrutinizing his work at a global level, which
he expressed unwillingness to do.
In short, while he made some strides, Zedong’s development seemed constrained by limited
motivation. Despite being compensated on an hourly basis, he never spent more than two
sessions, or much more than two hours, on any writing task. He seemed willing to take on board
strategies that created efficiencies and made writing easier (e.g., planning), but those addressing
writing quality that might entail substantial effort (i.e., global evaluation and revision) lost in his
cost-benefit analysis. These seemed likely to be a significant challenge for him going forward.
That being said, by the end of the project, he was using the process terms (e.g., formulation,
global evaluation) appropriately to discuss his writing and showed some capacity to interpret his
graphs and critique his process engagement.
4.2 Mingyu: Overplanning, under-evaluating, and tending to go off topic
Mingyu was also a freshman on her first international study experience who was majoring in
Computer Engineering. She was more proficient in English than Zedong, having scored a 98 on
the TOEFL examwell above the 71 needed for entrance to the universityand she had placed
into the higher-level writing course. She expressed uncertainty about her future plans beyond
“coding every day in front of the computer” (FS4). She believed training in academic writing
was important insofar as it developed “logical thinking” (FS4), but beyond this, she did not
anticipate a need for such writing in her future. She said she found it difficult to write according
to prescribed topics, but “when I just write whatever I want, that seems fine to me” (FS4).
Mingyu’s submission for WT1 was a five-paragraph essay consisting of 642 words. While it was
well organized, the essay was critiqued for having strayed from the prompt. Rather than arguing
whether or not adults were justified in feeling upset when young people flouted tradition, she
wrote a descriptive piece focusing on two psychological concepts that could explain people’s
attachment to traditions: “Consistency theory” and the “‘Knew it all along’ effect” (WT1). Most
of the essay was devoted to illustrating these ideas.
Analysis of the process data for this task showed Mingyu had completed it in a single session
totaling 88 minutes (two recordings with a small break in between), nearly the same amount of
time spent by Zedong. Unlike Zedong, however, Mingyu had begun by creating an external plan
consisting of ideas for major points indexed to numbered paragraphs in the essay. She started
formulating at the 13-minute mark and showed some ability to compartmentalize processes by
shifting between the evolving text below and her plan above, which she would update as new
ideas occurred to her and which she used to replenish her ideas at junctures in the formulation of
There were, however, only two instances where she returned to the prompt at the top of the file;
one of these (Figure 4 at the 45-minute mark) shows her reviewing it before an ex-post facto
attempt to connect her description of “Consistency theory” back to the topic of the assignment.
In addition, the process data depicted a final evaluation and revision stage of only nine minutes,
during which she made mostly local-level changes (substitutions for word choice and phrases
elaborated for clarity), although she did add a sentence to aid the transition between paragraphs.
These two process issues seemed clearly connected to the problems she had had keeping her
essay aligned with the prompt, and review of her project-initial reflection suggested this was not
an unusual occurrence for her. In discussing the importance of planning, she had written:
“Making an outline for your article can help you better control the direction of your article, so
that you will not be off topic” (RT1).
Figure 4. Process graph from the first recording of Mingyu’s two writing sessions for WT1
Because she already seemed aware of the benefits of planning, our goals for her initially focused
on avoiding topic drift by rereading the prompt at regular intervals during planning and
formulating as well as before conducting global evaluation and revision. We supplied her with
guiding questions to use during global evaluation (Appendix D) and told her to make sure this
stage preceded a change of focus to local issues of word choice and grammar.
In her subsequent tasks, she continued to display a penchant for planning, employing different
approaches (brainstorming, outlining, freewriting), with the size of her plans usually far
exceeding the text proper in the number of characters and in the time spent developing them.
Much of this material would not end up in the final product. At one point, she was working from
two distinct sets of plans, one labeled “outline” consisting of bullet-point items, and another
labeled “outline draft” containing more detailed ideas written in lengthy phrases or complete
sentences. Her composing was also highly recursive and nonlinear, indicative of a knowledge-
transformational approach. She wrote conclusions before body paragraphs and annotated
sections of plan or text with bracketed comments such as “Doesn't explain why” and “this is not
the topic!!!”
Despite these latter strategies, she continued to experience problems with drifting off topic. In
WT2, instead of writing about a popular fad that she disliked, she chose to write about
procrastination, again concentrating on psychological reasons for why people engage in it and
the harm it can do. In WT3, she was more on track, providing different perspectives on the
question of whether computers benefited a privileged few, but her thesis in the introduction was
contradicted by a statement in the conclusion. In the final task, she went off topic yet again,
spending most of the essay discussing the importance of having clear goals in college and
whether it was even necessary to attend college, while the notion of gap years, the topic of the
prompt, was mentioned only twice and never explained.
By the end of her participation in the project, we had identified additional causes for Mingyu’s
problems staying on task. First, she tended to get carried away by her own ideas, the logic of her
thought process, and new information she discovered while resourcing. Her lengthy, complicated
plans further facilitated topic drift by making it difficult for her to evaluate her plan in relation to
the prompt. An additional cause was her approach to evaluation and revision once she had
completed a full draft of her essay. The time she devoted to these processes was less than her
case-study counterpart (Table 3), and playback showed her focus during these stages to be
almost exclusively local. Once all the components of the text were in place, she would begin
local-level evaluation and revision and then finish writing.
Asked about her reluctance to engage in global evaluation despite repeated encouragement and
the provision of questions to guide this process, Mingyu expressed doubt about its value in light
of the amount of planning she had done.
Mingyu: I didn’t ask [these questions] … after I wrote everything. I read these before I
start expanding my ideas …
Instructor: But you didn’t do this after you made a complete draft?
Mingyu: No.
Instructor: How come?
Mingyu: Because I think I already spent enough time on this. (FS3)
Instead of using the questions to evaluate a completed draft of her essay, she was using them to
develop and evaluate her plans. She seemed to assume that, as long as she had material in her
plan addressing the guiding questions, the final product would also manifest these components
and qualities, so there was no need to revisit the questions. “When I finished writing it, I read it
again basically to just check grammar, and I think I just assumed that I did well on the task. And
then, based on my ideas [in the plan], I read it again. Like, it seems good to me, so I didn't
analyze from another aspect” (FS4).
Not surprisingly, Mingyu’s work did not evidence much progress across the four tasks. She
seemed frustrated in the final follow-up session but also indicated that she was coming to the
realization that she could not rely on planning alone to guarantee good results. “When I was
writing this, I had a very clear outline, but when I look at it right now, I can’t remember the
outline. I think when I translated my outline to my article, I didn’t write so clear that I can
understand,” she said. “When I look at this now, I can’t see the logic in it” (FS4). Like Zedong,
however, she finished the project having acquired more terms and concepts for describing the
way she went about writing as well as some capacity to interpret the visualizations and make
observations about her process engagement.
5. Discussion
5.1 Summary of findings
This study investigated whether process data in the form of visualizations generated from
keystroke and eye-tracking logs provided insights about L2 writing that could enhance
instructors’ efforts to create more meaningful assessment for learning of L2 writing. The findings
clearly provide an affirmative answer. We summarize the findings here with reference to the
aforementioned definition of AfL as “the process of seeking and interpreting evidence for use by
learners and their teachers to decide where learners are in their learning, where they need to go,
and how best to get there” (Broadfoot et al., 2002, pp. 2-3).
5.1.1 Understanding where learners are in their learning
The process data and the collaborative dialogues they facilitated were seen to provide rich
information about where students are in their current stage of development of L2 writing skills.
This information can be understood from two perspectives. First, it provided macro-level
diagnoses of Zedong’s and Mingyu’s skills in relation to the developmental models proposed by
Bereiter and Scardamalia (1987), with Zedong representing the knowledge-telling end of the
spectrum and Mingyu displaying a knowledge-transforming approach, albeit one with gaps.
These more general characterizations were immediately useful in determining subsequent
attempts at intervention; in Zedong’s case, adding those processes that were not yet represented
in his repertoire, and in Mingyu’s case, honing in on a specific process (task definition) with
which she seemed to be having trouble. This connects to the second perspective: formative
assessment of how students were performing at implementing recommended changes to their
process engagement. These subsequent, finer-grained assessments were rich enough in detail to
allow the instruction to be individually tailored, with the result that less time was spent on those
issues that were more easily addressed (e.g., Zedong’s uptake of the placeholder strategy) and
more on those problems that proved more intractable (e.g., both participants’ reluctance to
engage in global evaluation).
5.1.2 Understanding where learners need to go and how best to get there
In these respects as well, the findings showed process data providing insights that conventional,
product-focused assessment for learning do not afford. The process goals provided manageable
sets of targets for participants to aim for, and because they were often working on the same
process or strategy across multiple tasks, there was a sense of connectedness and progression to
the instruction that contrasted with what can seem like a succession of detached, standalone tasks
in some writing courses. This allowed the participants to develop clearer understandings of
individual processes and of process in general as an important aspect of writing as the project
In addition, the model-writers’ process data, when contrasted with that of the participants,
showed in visual terms what desired performance looked like and how it could be achieved. The
dynamically linked process graph and playback afforded macro- and micro-level views of model
approaches, and because these were recorded data they could be viewed multiple times across
follow-up sessions as needed. This affordance is particularly notable in light of research showing
the positive effects of observational learning on writing skills development (Braaksma et al.
2004; Couzijn, 1999; Lindgren et al., 2008).
Finally, the process data afforded the instructor the ability to draw from other aspects of the WP
literature beyond developmental models in setting goals and recommending specific strategies
for achieving them. For example, we knew of ways to address Zedong’s disfluencies in
formulating because of WP studies (1) showing connections between disfluencies and absence of
external planning (Kellogg, 1988) and (2) describing skilled L2 student writers’ use of a
placeholder-like strategy (Zamel, 1983). To be sure, this knowledge base may not provide
solutions for, or even documentation of, every issue instructors may encounter in assessing
students’ process engagement; for instance, we had trouble finding reference to anything
resembling Mingyu’s problems with topic drift. However, the potential for process tracing to
make this copious body of knowledge about, among other things, differences in skilled and less
skilled writers more immediately relevant to L2 assessment and instruction is an important
5.1.3 Other insights from process-tracing
Beyond alignment with these core definitional properties, the study also found process-tracing to
present additional affordances for AfL, including as a way of understanding specific, problematic
features of written products. A case in point was Mingyu’s awkward, after-the-fact attempt to
connect a body paragraph back to the topic of the prompt, evidenced by an isolated downward
spike in the process graph when she scrolled up to reread the instructions. Had her work on WT1
been assessed only on the basis of the final product, the poorly integrated sentence in question
would likely have been regarded simply as a lapse in organization and forgotten. Instead, it was
the first indication of an ingrained problem of topic drift related to task definition and global
evaluation. The potential for problematic features of written products to be rendered not only
more understandable but also instructional by salient events in the process data corroborates
what Hattie and Timperley (2007) assert regarding feedback about the task (FT) and feedback
about processing of the task (FP): namely, that FT is most useful when it supports FP. Specific
product-process connections such as this are one of the more intriguing findings of this study,
and one that merits more research.
Finally, we note how process data provided a useful means of triangulating or validating process
information from other sources; namely, student self-reports as expressed in reflective writing. It
is probably not uncommon for learners to profess one thing while actually doing something else,
as was the case with Zedong’s statement valorizing revision in his initial reflection, which
contrasted with the almost complete absence of revision in his work on the first task. Process
data can reveal such discrepancies, which we hasten to acknowledge do not necessarily result
from dishonesty. The cognitive demands of complex writing tasks make it difficult for learners
to accurately monitor their engagement in writing processes (Couzjin, 1999), so process data
may be useful as input for reflection. Mingyu’s reflection about getting off topic was rendered
more meaningful by the process data showing her to have a recurring problem in this regard.
This raises the question of how process tracing can support not only instructors but learners in
AfL, which is discussed below.
5.2 Implications for future applications of process-tracing to AfL of L2 writing
The findings summarized above show clearly that process tracing can support more meaningful
assessment for learning of L2 writing. There are several implications of these findings for current
practices in L2 writing assessment and instruction. First and most obvious is that the potential
benefits of process tracing may be hard to ignore. The experience of assessing writing in terms of
both product and process will make it difficult to go back to assessing writing on the basis of
product alone. This is because of the depth and breadth of perspective that process data can
provide on why written products are the way they are. By assessing products only, instructors
may be addressing only symptoms of underlying issues, resulting in lost opportunities to support
student learning. The appeal of process-tracing is strengthened by calls for writing assessments
that take account of processes (e.g., Deane, 2013) and the availability of a significant body of
knowledge at the ready to inform and support its application to AfL.
That being said, we must recognize that incorporating a process focus into AfL would require
major changes to current practices. Time would need to be created for this work in syllabuses
and timetables. Teachers and students would require significant preparation and training. As with
all assessments, applications for particular contexts would necessitate domain analyses and the
development of grading criteria and scoring guides. Given the large-scale nature of the
changes—including significant teacher preparation and students’ openness to allowing their
work to be scrutinized in this wayvalidity arguments would need to be developed and
validation research conducted to demonstrate that the interpretations and uses of process-based
assessments for learning are warranted.
The scale of the changes to current practice that might be involved therefore recommends a
gradual and incremental adaptation of current practices. The first author is already experimenting
with this in an ongoing writing course in his own work context. Adaptations to the approach used
in the study so far include (1) limiting individual consultations with students to two: one at the
beginning of the term, after the first writing assignment, and a second at the end of the term
while students are working on the final assignment, with a combination of written commentary
and visualization feedback provided for those assignments in between; (2) forgoing the use of
eye-tracking so that students can write at times and in locations that are convenient for them; and
(3) limiting process feedback to a narrower range of problems that are common to at least four of
five students so that the use of class time to address these issues can be justified. At the time of
writing, the adaptations have been restricted to sections of the course taught by the first author,
but in the near future we plan to include additional sections with the help of cooperating
instructors who will also serve as research participants in a study about factors that enable and
constrain the implementation of process-focused AfL. We expect that previous studies of
implementation of AfL-related innovations in existing curricula (e.g., Huang, 2016; Lee &
Falvey, 2014) may prove informative in these efforts.
5.3 Limitations and future research
Some limitations must be borne in mind while interpreting these findings. We obviously make
no claims about the representativeness of the participants or the generalizability of our findings
to larger groups. While we tried to enhance ecological validity wherever possible, the
requirement that the participants complete the writing tasks in a campus computer lab may very
well have influenced the way they went about it. Another important limitation was that it was
impossible for them to plan or take notes in their L1, due to the CyWrite software's inability to
accommodate the use of Chinese characters. Some research has shown benefits for students
planning in their L1 (e.g., Jones & Tetroe, 1987).
In addition to those research ideas already mentioned, we see the need for further related studies
as follows. First, as a priority, research should evaluate what effects, if any, process-focused AfL
has on relevant outcomes such as writing quality, self-efficacy for writing, and mental models of
writing. In addition, design-based research (Design-Based Research Collective, 2003) should be
used investigate which types of visualizations, or other means of conveying information about
processes, are more effective for different audiences and purposes. Finally, and relatedly,
research should be conducted to see whether, and under what conditions, process-tracing can be
used directly by students to enable self-assessment that informs their learning (i.e., assessment as
learning, or AaL; e.g., Lee, 2017).
6. Conclusion
The motivation for this study was to explore whether process-tracing technologies used in
writing research could be applied to assessment for learning of L2 writing. These technologies
have shown potential not only for helping students better manage the behaviors through which
writing is necessarily achieved but for allowing instructors to capture copious amounts of
information about students’ actual writing abilities with which to inform instruction. They will,
we predict, become commonplace in the next five to 10 years. In the meantime, there is still
much work to do in building a pedagogy that incorporates them, including much more piloting,
evaluation, refinement, and scholarly discussion on the part of practitioners and researchers, as
well as professional development on the part of teachers. Based on our relatively brief
experience so far, however, we feel process-tracing for AfL holds considerable promise and
venture so far as to say there may come a time when writing teachers, and perhaps language
testers as well, may wonder how writing was ever assessed without it.
The authors would like to thank the anonymous reviewers and editors for their helpful comments
on earlier versions of this paper. We also gratefully acknowledge the National Science
Foundation, which supported the project under Grant No. 1550122.
Almond, R. G., Deane, P., Quinlan, T., Wagner, M., & Sydorenko, T. (2012). A preliminary
analysis of keystroke log data from a timed writing task. ETS Research Report Series,
2012(2), i-61. doi: 10.1002/j.2333-8504.2012.tb02305.x
Bécotte, H.-S., Caporossi, G., Hertz, A., & Leblay, C. (2015). Writing and rewriting: Keystroke
logging’s colored numerical visualization. Unpublished manuscript.
Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Bitchener, J., & Ferris, D. R. (2012). Written corrective feedback in second language
acquisition and writing. New York, NY: Routledge.
Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education:
Principles, Policy & Practice, 5(1), 7-74. doi: 10.1080/0969595980050102
Braaksma, M. A., Rijlaarsdam, G., Van den Bergh, H., & van Hout-Wolters, B. H. M. (2004).
Observational learning and its effects on the orchestration of writing processes. Cognition
and Instruction, 22(1), 1-36.
Breetvelt, I., van den Bergh, H., & Rijlaarsdam, G. (1994). Relations between writing processes
and text quality: When and how? Cognition and Instruction, 12(2), 103-123. doi:
Broadfoot, P. M., Daugherty, R., Gardner, J., Harlen, W., James, M., & Stobart, G. (2002).
Assessment for learning: 10 principles. Cambridge: University of Cambridge School of
Caporossi, G., & Leblay, C. (2011). Online writing data representation: A graph theory
approach. In Proceedings of the 10th International Conference on Advances in Intelligent
Data Analysis X (Porto, Portugal, October 29-31), 80-89. DOI: 10.1007/978-3-642-
Carless, D. (2006). Differing perceptions in the feedback process. Studies in Higher Education,
31(2), 219-233.
Cho, Y. (2003). Assessing writing: Are we bound by only one method? Assessing Writing, 8(3),
165-191. doi:
Chong, S. W. (2017). Three paradigms of classroom assessment: Implications for written
feedback research. Language Assessment Quarterly, 1-18. doi:
Design-Based Research Collective. (2003). Design-Based Research: An emerging paradigm for
educational inquiry. Educational Researcher, 32(1), 5-8. doi: 10.2307/3699927
Couzijn, M. (1999). Learning to write by observation of writing and reading processes: effects
on learning and transfer. Learning and Instruction, 9(2), 109-142. doi:
Cumming, A. (1989). Writing expertise and second-language proficiency. Language Learning,
39(1), 81-135. doi: 10.1111/j.1467-1770.1989.tb00592.x
Cumming, A. (2002). Assessing L2 writing: Alternative constructs and ethical dilemmas.
Assessing Writing, 8(2), 73-83. doi:
Deane, P. (2013). On the relation between automated essay scoring and modern views of the
writing construct. Assessing Writing, 18(1), 7-24. doi:
Doughty, C. (2001). Cognitive underpinnings of focus on form. In P. Robinson (Ed.), Cognition
and Second Language Acquisition (pp. 206-257). Cambridge: Cambridge University
Flower, L., & Hayes, J. R. (1981). A cognitive process theory of writing. College Composition
and Communication, 32(4), 365-387. doi: 10.2307/356600
Flower, L., Hayes, J. R., Carey, L., Schriver, K., & Stratman, J. (1986). Detection, diagnosis,
and the strategies of revision. College Composition and Communication, 37(1), 16-55.
doi: 10.2307/357381
Graham, S., Harris, K., & Hebert, M. (2011). Informing writing: The benefits of formative
assessment. Washington, DC: Carnegie Corporation of New York.
Hattie, J., & Timperley, H. S. (2007). The power of feedback. Review of Educational Research,
77(1), 81-112. doi: 10.3102/003465430298487
Hayes, J. R. (2012). Modeling and remodeling writing. Written Communication, 29(3), 369-
388. doi: 10.1177/0741088312451260
Huang, S.-C. (2012). Like a bell responding to a striker: Instruction contingent on assessment.
English Teaching: Practice and Critique, 11(4), 99-119.
Huang, S.-C. (2016). No longer a teacher monologue: Involving EFL writing learners in
teachers' assessment and feedback processes. Taiwan Journal of TESOL, 13(1), 1-31.
Jones, S., & Tetroe, J. (1987). Composing in a second language. In A. Matsuhashi (Ed.),
Writing in real time: Modelling production processes (pp. 34-57). New York: Longman.
Kellogg, R. T. (1988). Attentional overload and writing performance: Effects of rough draft and
outline strategies. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 14(2), 355-365. doi: 10.1037/0278-7393.14.2.355
Kellogg, R. T. (2008). Training writing skills: A cognitive developmental perspective. Journal
of Writing Research, 1(1), 1-26.
Khuder, B., & Harwood, N. (2015). L2 writing in test and non-test situations: Process and
product. Journal of Writing Research, 6(3), 233-278.
Lee, I. (2011). Bringing innovation to EFL writing through a focus on assessment for learning.
Innovation in Language Learning and Teaching, 5(1), 19-33.
Lee, I. (2017). Classroom writing assessment and feedback in L2 school contexts. Singapore:
Lee, I., & Coniam, D. (2013). Introducing assessment for learning for EFL writing in an
assessment of learning examination-driven system in Hong Kong. Journal of Second
Language Writing, 22(1), 34-50. doi:
Lee, I., & Falvey, P. (2014). Perspectives on assessment for learning in Hong Kong writing
classrooms. In D. Coniam (Ed.), English language education and assessment: Recent
developments in Hong Kong and Chinese Mainland (pp. 221-236). Singapore: Springer.
Leijten, M., & Van Waes, L. (2013). Keystroke logging in writing research: Using Inputlog to
analyze and visualize writing processes. Written Communication, 30(3), 358-392. doi:
Leijten, M., Van Waes, L., Schriver, K., & Hayes, J. R. (2014). Writing in the workplace:
Constructing documents using multiple digital sources. Journal of Writing Research,
5(3), 285-337.
Lindgren, E., Stevenson, M., & Sullivan, K. P. H. (2008). Supporting the reflective language
learner with computer keystroke logging. In F. Zhang & B. Barber (Eds.), Handbook of
research on computer-enhanced language acquisition and learning (pp. 189-204).
Hershey, NY: Information Science Reference.
Lindgren, E., & Sullivan, K. P. H. (2002). The LS graph: A methodology for visualizing writing
revision. Language Learning, 52(3), 565-595. doi: 10.1111/1467-9922.00195
Lindgren, E., & Sullivan, K. P. H. (2006). Analysing online revision. In K. P. H. Sullivan & E.
Lindgren (Eds.), Computer Keystroke Logging: Methods and Applications: Studies in
Writing, Vol. 18 (pp. 157-188). Oxford: Elsevier.
Lindgren, E., Sullivan, K. P. H., Lindgren, U., & Spellman Miller, K. (2007). GIS for writing:
Applying Geographical Information Systems techniques to data mine writings' cognitive
processes. In M. Torrance, L. Van Waes & D. Galbraith (Eds.), Writing and cognition:
Research and applications (pp. 83-96). Amsterdam: Elsevier.
Mak, P., & Lee, I. (2014). Implementing assessment for learning in L2 writing: An activity
theory perspective. System, 47, 73-87. doi:
National Assessment of Educational Progress. (2011). Writing 2011: National assessment of
educational progress at grade 8 and 12 National Assessment of Educational Progress.
Washington, DC: Institute of Educational Sciences.
Nicolás-Conesa, F. (2012). Development of mental models of writing in a foreign language
context: Dynamics of goals and beliefs. PhD dissertation, University of Murcia.
Nicolás-Conesa, F., Roca de Larios, J., & Coyle, Y. (2014). Development of EFL students’
mental models of writing and their effects on performance. Journal of Second Language
Writing, 24(0), 1-19. doi:
O'Brien, C. W. (1992). A Large-Scale Assessment to Support the Process Paradigm. The
English Journal, 81(2), 28-33. doi: 10.2307/819969
Perrin, D. (2003). Progression analysis (PA): investigating writing strategies at the workplace.
Journal of Pragmatics, 35(6), 907-921. doi:
Rijlaarsdam, G., & van den Bergh, H. (2006). Writing process theory: A functional dynamic
approach. In C. A. MacArthur, S. Graham & J. Fitzgerald (Eds.), Handbook of writing
research (pp. 41-53). New York: Guilford Publications.
Rinnert, C., Kobayashi, H., & Katayama, A. (2015). Argumentation text construction by
Japanese as a foreign language writers: A dynamic view of transfer. The Modern
Language Journal, 99(2), 213-245. doi: 10.1111/modl.12210
Roca de Larios, J., Manchón, R. M., Murphy, L., & Marín, J. (2008). The foreign language
writer's strategic behaviour in the allocation of time to writing processes. Journal of
Second Language Writing, 17(1), 30-47. doi:
Roca de Larios, J., Marín, J., & Murphy, L. (2001). A temporal analysis of formulation
processes in L1 and L2 writing. Language Learning, 51(3), 497-538. doi: 10.1111/0023-
Roca de Larios, J., Murphy, L., & Marín, J. (2002). A critical examination of writing process
research. In S. Ransdell & M. L. Barbier (Eds.), Studies in writing, Volume 11: New
Directions for research in L2 writing (Vol. 11-47). Netherlands: Kluwer Academic
Severinson Eklundh, K., & Kollberg, P. (1996). A computer tool and framework for analyzing
online revisions. In C. M. Levy & S. Ransdell (Eds.), The science of writing: Theories,
methods, individual differences, and applications. Hillsdale, NJ, England: Lawrence
Erlbaum Associates, Inc.
Spelman Miller, K., Lindgren, E. V. A., & Sullivan, K. P. H. (2008). The psycholinguistic
dimension in second language writing: Opportunities for research and pedagogy using
computer keystroke logging. TESOL Quarterly, 42(3), 433-454. doi: 10.1002/j.1545-
Strijbos, J.-W., & Stahl, G. (2007). Methodological issues in developing a multi-dimensional
coding procedure for small-group chat communication. Learning and Instruction, 17(4),
Sullivan, K. P. H., & Lindgren, E. (2002). Selfassessment in autonomous computeraided
second language writing. ELT Journal, 56(3), 258-266. doi: 10.1093/elt/56.3.258
Teddlie, C., & Tashakkori, A. (2008). Foundations of mixed methods research: Integrating
quantitative and qualitative approaches in the social and behavioral sciences. Los
Angeles: Sage Publications.
Uzawa, K. (1996). Second language learners' processes of L1 writing, L2 writing, and
translation from L1 into L2. Journal of Second Language Writing, 5(3), 271-294. doi:
Victori, M. (1999). An analysis of writing knowledge in EFL composing: A case study of two
effective and two less effective writers. System, 27(4), 537-555. doi: 10.1016/s0346-
Weigle, S. C. (2002). Assessing writing. Cambridge: Cambridge University Press.
Wengelin, Å., Torrance, M., Holmqvist, K., Simpson, S., Galbraith, D., Johansson, V., &
Johansson, R. (2009). Combined eyetracking and keystroke-logging methods for studying
cognitive processes in text production. Behavior Research Methods, 41(2), 337-351. doi:
Wiliam, D. (2011). What is assessment for learning? Studies in Educational Evaluation, 37(1),
3-14. doi:
Wolcott, W. (1987). Writing instruction and assessment: The need for interplay between
process and product. College Composition and Communication, 38(1), 40-46. doi:
Worden, D. L. (2009). Finding process in product: Prewriting and revision in timed essay
responses. Assessing Writing, 14(3), 157-177. doi:
Zamel, V. (1983). The composing processes of advanced ESL students: Six case studies.
TESOL Quarterly, 17(2), 165-187.
Appendix A. Writing tasks used in the project
Writing task
Topic and prompt
Breaking Traditions: Many adults become upset when young people break with
traditions of the past. Do you think that these adults are justified in reacting this way?
Why or why not? Support your position with evidence from your own experience or
the experiences of people you know.
Fads and Trends: Briefly describe a fad or trend that you dislike. Explain why it has
attracted so many followers and why you dislike it. Develop your point of view by
giving reasons and/or examples from your own experiences, observations or reading.
Computers and Privilege: Some people say that computer technology gives an unfair
advantage to a privileged few. To what extent do you agree or disagree with the
statement above? Support your views with specific examples from your own
experience, observations or reading.
Gap Year: At least one major United States university officially recommends that high
school students take a year off a so-called gap year before starting college. The
gap year idea is gaining popularity. Supporters say it helps students mature and focus
on their goals. Detractors say taking a year off from school will get students off track
and that many will never go to college if they don't go right away. Do you think taking
a gap year is a good idea? Why or why not?
Appendix B. Reflection tasks
Assigned at the beginning of the project.
Thinking back on what you have learned in your previous schooling, write a reflection in which
you try to explain to a prospective international student from your home country what good
academic writing is and what it involves. Your reflection should consist of 250-300 words, or 2-
3 paragraphs.
Asssigned at the end of the project.
Now that your participation in this project has come to an end, write another reflection in which
you explain to a prospective international student from your home country what good academic
writing is and what it involves. Your reflection should consist of 250-300 words, or 2-3
Appendix C.
Coding scheme for reflection tasks focusing on participants' mental models of writing
(adapted from Nicolás-Conesa, 2012)
Dimensions of writing
E.g., "a good academic writing will help the writer
illustrate their idea effectively. The audience should get
the writer's idea from the essay."
E.g., "The grammar and spelling mistakes will make the
writing hard to understand. "
E.g., "a good academic writing usually include the thesis
statements, supporting points, and the conclusion"
E.g., "Before submit an academic writing we should
check several time about the mistake we usually make,
like tense, capital letter and space in front of the
E.g., "A good academic writing must contain several
parts. "
Coding scheme for process-visualization data
(Re)reading prompt
Planning - Brainstorming
Planning - Freewriting
Planning - Outlining
Planning - Replenishing
Planning - Updating
Revision - Local
Revision - Global
Resourcing - Language
Resourcing - Content
Placeholder strategy
"Notes to self" strategy
Coding scheme for evaluated essays
Adherence to task and topic
Issues with meeting task requirements as outlined in the
Audience and purpose
Rhetorical concerns, such as assuming cultural knowledge on
the part of the reader or failing to state a clear position in an
Definition of key term or
Issues related to central concepts or ideas; typically, leaving
these undefined or insufficiently elaborated
Language issue
Issues related to grammar or lexis
Logical connections
Problems such as gaps in reasoning or contradictions
Organization of information
Includes problems with unity in paragraphs, redundancy,
Support for main ideas
Issues with the relevance and quality of supporting ideas, the
diversity of types of support, relative amount of support
across main ideas
Thesis statement or controlling
Lack of clear thesis statements, topic sentences, or
connections between the two; positioning of controlling ideas
Lack of transitions, overuse of certain transitions,
inappropriate use of transitions
Coding scheme for follow-up session transcripts
Views on English and
E.g., "I don't think I need to do very much writing, but I
definitely going to do a lot of reading."
E.g., "This is like planning and formulating together."
E.g., "I think if something wrong, it's at the end, it's
difficult to change the order or make changes."
E.g., "Just this time at the end I didn't type a lot. Actually I
just can't write anymore. I don't know how to continue
when writing."
Language/ resourcing
E.g., "I was confused by justified. I was thinking it was
like another form of justice, so I just Googled it."
E.g., "I think I didn't make my plan very well. The
structure of my plan, I have some problem, like they didn't
Prompts/ topics
E.g., "The topic asked me to describe a fad or trend you
dislike. I actually don't have a trend or fad I really don't
Strategies (Placeholder,
Notes to self)
E.g., "Some method, like use a placeholder, like, make me
focus on the assignment and do the details later."
Task definition
E.g., "At the very beginning I was writing down my ideas
and then I went too far and I start talking about fairness,
and then I realized after I checked the topic then I get rid
of the fairness thing."
Technology issue
E.g., "The devices was running well last time. It's better
than the first time."
E.g., "I think I spent too much time in one article."
Appendix D. Goals sheet provided to one of the case-study participants.
ResearchGate has not been able to resolve any citations for this publication.
The influence of cognitive processing on second language acquisition (SLA), and on the development of second language (SL) instruction, has always been a subject of major interest to both SLA researchers and those involved in SL pedagogy. Recent theoretical research into SLA and SL pedagogy has shown renewed interest in the role of cognitive variables such as attention, short, working, and long term memory, and automaticity of language processing. This volume first examines the theoretical foundations of research into the cognitive processes underlying SLA, and then describes various implications for pedagogically oriented research and for SL classroom practice. The blend of research from the cognitive sciences and applied linguistics make it an excellent introduction to applied linguists and language teachers interested in the psycholinguistic processes underlying SLA.
Classroom assessment has always been an indispensible and integral part of any curriculum. In particular, assessment plays the role of reporting students' learning summatively (assessment of learning), providing diagnostic and formative information for teachers to inform their instruction (assessment for learning); more recently, Earl (2013) proposed the notion of assessment as learning, which puts students at the center of assessment. Students in this assessment paradigm act as critical connecters between assessment and learning through self-reflection and self-regulation. The first section of this article reconceptualizes summative and formative assessments into three assessment paradigms: assessment of, for, and as learning through incorporating Serafini's assessment models and Habermas's three human interests. In so doing, our understanding of the three paradigms is consolidated and enriched to encompass not only the pedagogical implications but also their philosophical and epistemological underpinnings. The second section of the article focuses on one particular kind of assessment method commonly used in language classrooms, which is written feedback. I summarize and categorize recent written feedback research with reference to the three assessment paradigms and suggest directions for future research.
While assessment and feedback tend to be treated separately in the L2 writing literature, this book brings together these two essential topics and examines how effective classroom assessment and feedback can provide a solid foundation for the successful teaching and learning of writing. Drawing upon current educational and L2 writing theories and research, the book is the first to address writing assessment and feedback in L2 primary and secondary classrooms, providing a comprehensive, up-to-date review of key issues, such as assessment for learning, assessment as learning, teacher feedback, peer feedback, portfolio assessment, and technology enhanced classroom writing assessment and feedback. The book concludes with a chapter on classroom assessment literacy for L2 writing teachers, outlines its critical components and underscores the importance of teachers undertaking continuing professional development to enhance their classroom assessment literacy. Written in an accessible style, the book provides a practical and valuable resource for L2 writing teachers to promote student writing, and for teacher educators to deliver effective classroom writing assessment and feedback training. Though the target audience is school teachers, L2 writing instructors in any context will benefit from the thorough and useful treatment of classroom assessment and feedback in the book. "A book on L2 writing assessment that truly is useful for teachers, school administrators, and students’ learning. Educators around the world, heed these well-founded insights." Alister Cumming, University of Toronto "This book carefully synthesizes two critical aspects of writing instruction--classroom assessment and feedback from various sources--that are often relegated to "separate" chapters or even books on how to teach writing to second language (L2) students. Professor Lee shows how a formative, supportive approach to both assessment and feedback provides the foundation for successful writing instruction and student growth. In her usual clear and compelling style, Lee links current and existing research trends to specific pedagogical techniques, providing a practical and accessible resource to in-service and pre-service teachers. Though the main audience is teachers at the primary and secondary levels, L2 writing instructors in any context will benefit from this thorough and useful treatment of these essential topics." Dana Ferris, University of California, Davis
This chapter presents the use of the Geographical Information Systems (GIS) for data mining and visualising information about cognitive activities involved in writing. The information can be collected from various sources, such as keystroke logs, manual analysis of stimulated recall sessions and think-aloud protocols. After an introduction to the GIS, an English as a foreign language (EFL) writing session is used to explain how to create the various GIS layers from the different information/analysis sources, and show how they can be easily data mined using the GIS techniques to improve our understanding of the cognitive processes in writing. The illustrative graphs used to provide an insight into the methodology are based on keystroke-logged data, manual researcher-based analyses and coded stimulated recall data that were collected after the writing session. Also a tool for visualisation and data mining, the GIS technique can support analysis of the interaction of cognitive processes during writing focusing on the individual writer, differences between writers or the writing processes in general. Depending on the research question, GIS affords the possibility to aggregate data to the level of writers, de-aggregate data in any way chosen or display data as attributes of individuals. © 2007 by Elsevier Ltd. All rights of reproduction in any form reserved.
Second language writing assessment has traditionally entailed a primarily summative function, where assessment information is used mainly for administrative and reporting purposes. If assessment is to make a stronger impact on student learning of writing, much has to be done to change the existing summative focus, which entails a move towards assessment for learning (AfL). This chapter discusses the principles associated with assessment for learning (AfL) particularly in the context of Hong Kong writing classrooms. After critiquing AfL and reviewing the data from Hong Kong writing classrooms, this chapter ends by highlighting the factors necessary for the successful implementation of AfL not only in the Hong Kong classroom between teachers and students, but also within the parameters of school, societal, and system contexts and wherever writing assessment is likely to be involved in changing paradigms. © Springer Science+Business Media Singapore 2014. All rights are reserved.
This article is concerned pragmatically with how recent research findings in assessment for learning (AfL) can bring about higher quality learning in the day-to-day classroom. The first half of this paper reviews recent studies within Black and Wiliam's (2009) framework of formative assessment and looks for insights on how pedagogical procedures could be arranged to benefit from and resonate with research findings. In the second half, based on lessons drawn from the review, the findings were incorporated into an instructional design that is contingent on formative assessment. The concept of teacher contingency is elaborated and demonstrated to be central to the AfL pedagogy. Attempts were made to translate updated research findings into an English as a foreign language (EFL) writing instruction to illustrate how teachers may live up to promises offered by recent developments on AfL. This AfL lesson, situated in L2 writing revision, made instruction contingent on and more responsive to learner performance and learning needs. As shown in an end-of-semester survey, learner response to the usefulness of the instruction was generally quite positive.