SIG Writing Conference 2026 – Digital Writing Technologies & AI

A systematic review of the role of motivation in digital multimodal composing

Type: Single Paper
Presenting author: Yu, Nan , Beijing Normal University, China
Chair: Newton, Moira , University of Auckland and Papatoetoe West School, New Zealand
Keywords: learning and learner characteristics;meta-analysis;motivational and emotional factors in writing;multimodality
Domain: Motivational, Social and Affective Processes; See Session Schedule

Abstract

AbstractIndividual differences (IDs) have been shown to account for a substantialproportion of variation in learning outcomes in second language acquisition (SLA). Specifically, as L2 writing is a cognitively complex and challenging endeavor, it isimperative to explore the role of IDs in this domain. Among them, motivation hasreceived particular attention, since “L2 learning is fundamentally a motivationalpursuit” (Li et al., 2022, p. 113). Digital multimodal composing (DMC) has emerged as a popular pedagogicalpractice in SLA, offering learners and teachers new opportunities for engagement andmeaning-making. Among the IDs mediating L2 students’ participation and success inDMC, motivation plays a crucial role. Understanding how motivation isconceptualized and measured, and how DMC shapes or is shaped by students’ motivational states, can provide deeper insights into how DMC tasks could be betterdesigned and integrated to facilitate L2 writing development. Following the PRISMA guidelines, this systematic review investigates howmotivation has been conceptualized, measured, and influenced in DMC research. Drawing on 30 empirical studies, this review addresses three research questions:(1) What constructs of motivation in DMC research are examined?(2) What effects of DMC on L2 students’ motivation are found?(3) What influencing factors of motivation in DMC are identified?Thematic synthesis revealed that (1) most studies focused on a limited set ofmotivational constructs, namely, intrinsic/extrinsic motivation, task value, andself-efficacy, often measured by general educational instruments without cleartheoretical justification or task-specific adaptation; (2) DMC tasks, particularly digitalstorytelling, were generally found to enhance motivation, although the effects variedin magnitude and durability by learner profiles, task designs, and learning contexts; (3)a combination of intertwined factors were identified: learner-related features (e.g., curiosity, identity), task-related conditions (e.g., genre, duration), and context-relatedfactors (e.g., audience, collaboration). Taken together, these findings underscore the potential and complexity ofintegrating DMC in a meaningful way to support and sustain learner motivation. Thispaper calls for more theoretically-grounded, task-specific, and context-sensitive futureresearch on this line of inquiry.ReferenceLi, S., Hiver, P., & Papi, M. (2022). The Routledge handbook of second languageacquisition and individual differences. Routledge.

Assessing argumentative writing through students’ interactions with generative AI

Type: Single Paper
Presenting author: Conijn, Rianne , Eindhoven University of Technology, Netherlands
Co-author: Oliveira, Manuel , Eindhoven University of Technology, Netherlands
Co-author: Zednik, Carlos , Eindhoven University of Technology, Netherlands
Co-author: Bombaerts, Gunter , Eindhoven University of Technology, Netherlands
Co-author: Sadowski, Bert , Eindhoven University of Technology, Netherlands
Keywords: AI writing;assessment;Generative AI;writing processes
Domain: Assessment and Evaluation
Related Symposium: The future of writing education; See Session Schedule

Abstract

As generative artificial intelligence (genAI) increasingly produces text that is indistinguishable from human work, conventional assessments that focus solely on the written product are becoming an unreliable measure of student learning. In this presentation, we therefore introduce an assessment method that focuses on the writing process. We focus on two components of student–genAI interaction during argumentative writing. First, directive reasoning interaction, which captures how purposefully students steer the AI. This is important because passive acceptance of AI output is often associated with lower-quality writing. Second, visible expertise, which reflects the extent to which course-related conceptual knowledge becomes apparent in the interactions.Student–genAI interaction data and final essay grades were collected from 70 graduate students who wrote argumentative essays using a self-chosen genAI tool. All 1,450 prompts were annotated using our taxonomy, developed from the course learning objectives combined with indicators of directive reasoning interaction and visible expertise. The taxonomy contains three main categories: writing, content, and argument, and 35 subcategories.The results showed that students most often prompted genAI to improve or evaluate their writing, such as grammar and style (41%). GenAI was used less frequently to evaluate or improve content (29%) or argumentation (22%). Interactions indicative of high directive reasoning interaction and visible expertise were positively related to performance. For example, prompts asking genAI to revise a specific argument, based on a clear, conceptual critique; or to integrate information from a source into a premise, were associated with higher essay grades. In contrast, interactions showing low directive reasoning or low visible expertise, such as “write an essay on topic X” or requesting a summary to be inserted into the essay, were related to below-average essay grades.To conclude, evaluating the writing process through student-genAI interactions may be used to complement and even replace traditional essay assessment methods. Future work should examine the generalizability of our findings to other argumentative writing assignments and explore how the assessment approach might apply to other types of written assessment. Finally, as genAI evolves, it needs to be considered whether any interactions from our taxonomy might become obsolete.

Differential Effects of a Tablet-Based Writing Intervention on Text Quality: An Intervention Study

Type: Single Paper
Presenting author: Michels, Viktoria , TU Chemnitz, Germany
Chair: Fitjar, Camilla , University of Stavanger, Norway
Keywords: intervention study;text composition;writing instruction;writing tools and writing technology
Domain: Instructional Design; See Session Schedule

Abstract

Differential Effects of a Tablet-Based Writing Intervention on Text Quality: An Intervention StudyStudents with weak writing skills struggle with text production and content learning, underscoring the need for early support (Becker-Mrotzek et al., 2014). Digital writing environments, such as those offering spell-checking and text-to-speech functions, may provide such support, particularly for weak writers (Graham & Harris, 2018). Yet despite their growing use in schools, little is known about which learners benefit most and how digital tools differentially affect the development of text quality.This study, conducted within the BMBF-funded EdToolS project, examines differential effects of a tablet-based writing intervention on text quality among 7th-grade students (N = 153) using a pre–post–follow-up design with a control group. Text quality was assessed using keyboard-written texts. The intervention comprised strategy instruction and training in the use of a word processor (spell-checking in EG1/EG2; text-to-speech in EG2), followed by a practice phase in which students wrote multiple texts (EG1/EG2: using tablet and tools, CG: handwritten). A language competence score derived via PCA was used to classify students into high- and low-performance groups.Linear regression models revealed that low performers in EG1 showed significantly greater short-term gains in text quality (pre–post) compared to the control group, whereas no differential effect emerged for EG2. Among high performers, text quality in EG2 remained more stable from pre to follow-up than in the control group.These findings provide insights into learner-specific benefits and limitations of digital writing tools. Given the increasing role of digital literacy, the study highlights the urgent need to align digital tools with differentiated writing instruction. Graham, S., & Harris, K. R. (2018). Evidence-Based Writing Practices: A Meta-Analysis ofExisting Meta-Analyses. In R. Fidalgo, Raquel, Harris, Karen R., & Braaksma, Martine (Hrsg.), Design Principles for Teaching Effective Writing (S. 13–37). Brill. https://brill.com/view/book/edcoll/9789004270480/B9789004270480_003.xmlBecker-Mrotzek, Michael, Joachim Grabowski, Jörg Jost, Matthias Knopp, und Markus Linnemann. „Adressatenorientierung und Kohärenzherstellung im Text -Zum Zusammenhang kognitiver und sprachlich realisierter Teilkomponenten von Sprachkompetenz“. Didaktik Deutsch, Nr. Jg. 19. (2014): 21–43.

Explicit instruction and rubrics for argumentative synthesis writing: Effect of Collaboration

Type: Single Paper
Presenting author: Gutiérrez-Bermejo, Eugenia , Universidad Autónoma de Madrid, Spain
Co-author: Cuevas, Isabel , Universidad Autonoma de Madrid, Spain
Co-author: Mateos, Mar , Universidad Autónoma de Madrid, Spain
Co-author: Martín, Ana , Universidad Autonoma de Madrid, Spain
Co-author: Luna, Maria , Universidad a Distancia de Madrid (UDIMA), Spain
Co-author: Martínez-Álvarez, Isabel , Madrid Open University, Spain
Chair: Stukker, Ninke , Groningen University, Netherlands
Keywords: collaborative writing;instructional design in writing;intervention study;reading and writing
Domain: Instructional Design; See Session Schedule

Abstract

Explicit instruction and rubrics for argumentative synthesis writing in Secondary Education: The effect of CollaborationGutiérrez-Bermejo, E.*, Cuevas, I.*, Mateos, M.*, Martín, A.* Luna, M** & Martínez, I**UAM*, UDIMA**Secondary education students must develop key competences to address current challenges, such as critical thinking and argumentative skills (European Commission, 2019). Writing an argumentative synthesis based on different texts presenting opposing perspectives on a topic is a complex task with great potential for promoting the development of these competences (Mateos et al.,2018). However, students struggle with identifying, contrasting, and integrating opposing perspectives, especially through weighing and synthesizing strategies, thus they require specific instructional support (Casado-Ledesma et al., 2021). The aim of this study is to compare the effectiveness of an instructional program for learning to write argumentative syntheses in the first year of secondary education, across different task settings (individual vs. collaborative writing). Instructional program includes learning activities based on explicit instruction (EI) and practice using an instructional rubric (PR), each adapted from Cuevas et al. (2024). Forty-nine students were assigned to two conditions (EI+PR vs EI+PR+C) and wrote three argumentative syntheses (pretest/mid-test/posttest syntheses). Results show that both conditions were effective in improving students’ synthesis quality. Additionally, in the practice session, students who wrote collaboratively achieved better results, although these differences were attenuated in the posttest. Findings are discussed, and we conclude with educational implications regarding the adaptation of task settings based on students’ profiles.Keywords: Argumentative Synthesis, Explicit Instruction, Rubric, Collaborative Writing.References.Casado-Ledesma, L., Cuevas, I., Van den Bergh, H., Rijlaarsdam, G.,Mateos, M., Granado-Peinado, M.,& Martín, E. (2021). Teaching argumentative synthesis writing through deliberative dialogues: Instructional practices in secondary education. Instructional Science, 49(4), 515-559. https://doi.org/10.1007/s11251-021-09548-3Cuevas, I. Mateos, M., Casado-Ledesma, L.,Olmos, R., Granado-Peinado, M.,Luna, M., Núñez, J.A. & Martín, E. (2024). How to improve argumentative syntheses written by undergraduates using guides and instructional rubrics. European Journal of Psychology of Education, 39, 4573–4596. https://doi.org/10.1007/s10212-024-00890-xMateos, M., Martín, E., Cuevas, I.,Villalón, R., Martínez, I., & González-Lamas, J. (2018). Improving written argumentative synthesis by teaching the integration of conflicting information from multiple sources. Cognition and Instruction, 36, 119–138. https://doi.org/10.1080/07370008.2018.1425300

Inputlog: New perspectives on keystroke logging

Type: Demonstration Session
Presenting author: Van Waes, Luuk , University of Antwerp, Belgium
Co-author: Van Horenbeeck, Eric , University of Antwerp, Belgium
Chair: Pogner, Karl-Heinz , Copenhagen Business School, Denmark
Keywords: AI writing;keystroke logging;keystroke logging data processing and analysis;on-line writing and digital media
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Inputlog is a widely used keystroke logging tool for observing and analyzing writing processes. This demo introduces the major new features of Inputlog 9.6.0 and outlines planned future developments.Versioning and Diary Function A new automatic versioning option allows users to save intermediate Word document versions at fixed intervals (e.g., every three minutes). Researchers can compare these versions to track document changes throughout the writing session.
An optional diary prompt in the closing wizard invites writers to comment on their session, facilitating the combination of process data and self-report.Expanded Logging Environments Because writing increasingly takes place outside MS Word, the logging environment has been expanded. Inputlog now offers dedicated logging modules for Google Docs and LibreOffice, broadening the range of authentic writing contexts that can be captured.Feedback Reports Inputlog generates student-centered feedback reports that visualize key process indicators, including process graphs and source interaction. Users may rely on the default template or customize report formats to meet instructional or research needs, such as the use of AI.Multilingual Logging New beta versions introduce preliminary support for logging Korean and Chinese script (via Pinyin). This extends Inputlog’s previous focus on Latin-based scripts and broadens its applicability in multilingual writing research. Copy-Task Dashboard Inputlog includes a standardized copy task designed to assess typing skills in thirteen languages using sentences, word triplets, and letter clusters. We also present a corpus of more than 5,000 anonymized copy-task recordings, accompanied by an interactive R-Shiny dashboard that allows researchers to explore the corpus, download data, and benchmark their own results.

Measuring the Quality of AI-generated Feedback? From Theoretical Modelling to Empirical Evidence

Type: Single Paper
Presenting author: Fürstenberg, Maurice , Ludwig-Maximilians-Universität (LMU), Germany
Chair: Gerth, Sabrina , University College of Teacher Education Tyrol, Austria
Keywords: AI writing;evaluation and assessment of writing competence;research methodology;writing at school
Domain: Assessment and Evaluation; See Session Schedule

Abstract

AI-generated feedback is widely used in schools without sufficient research having been conducted into its quality, particularly with regard to German students. This study therefore examines the quality of AI-generated feedback on German student texts, as well as how this quality is measured, from both theoretical and empirical perspectives. First, a theoretical model is developed based on international research (e.g. Fong, 2025; Jansen et al., 2025; Weidlich et al., 2025) which includes different producers and products. This model establishes the terminology used throughout the paper and illustrates that operationalising feedback quality poses a methodological challenge for empirical studies. Subsequently, a study compared feedback on three student texts in the form of a criteria-based assessment, an overall grade, and a short comment. This feedback was provided by 75 highly experienced Bavarian teachers and four AI systems. Finally, eight trained meta-reviewers assessed the quality of the human and machine feedback. In terms of overall grades, there was high inter-rater reliability (ICC = 0.7–0.9) between teachers and AI systems (with ten iterations). On average, AI models graded texts more leniently, but in the same order of ranking. The criterion-based assessment differed significantly. Regarding meta-feedback, an ordinal logistic model identified three criteria (explanation, concreteness and accuracy) as the strongest predictors of perceived usefulness, with the source (AI vs. teacher) having no significant influence. The results of the empirical study expand the area of research on real German pupils. The theoretical model helps to better systematise future studies and demonstrates the complexity of operationalising the central phenomenon of interest: the quality of AI-generated feedback. The many challenges involved in operationalising feedback quality are relevant for future studies. Fong, C. J. (2025). A renaissance in feedback science? Reviewing and reimagining feedback research methods. Contemporary Educational Psychology, 83, 102414.Jansen, T., Horbach, A., & Meyer, J. (2025). Feedback from Generative AI: Correlates of Student Engagement in Text Revision from 655 Classes from Primary and Secondary School Proceedings of the 15th LAK.Weidlich, J., et al. (2025). Teacher, peer, or AI? Comparing effects of feedback sources in higher education. Computers and Education Open, 9, 100300.

Promoting digital text production competences in primary education

Type: Single Paper
Presenting author: Heller, Kathrin , University of Education Karlsruhe, Germany
Co-author: Anskeit, Nadine , University of Education Karlsruhe, Germany
Chair: Gerth, Sabrina , University College of Teacher Education Tyrol, Austria
Keywords: instructional design in writing;intervention study;on-line writing and digital media;writing at school
Domain: Instructional Design; See Session Schedule

Abstract

The digital production of texts is considered a key competence in today's information and communication society (Frederking & Krommer, 2019). Familiarity with the writing medium is of great importance here, as it systematically influences text quality: fast typists produce better texts (Connelly et al., 2007; Gong et al., 2022). Initial pilot studies show that, in addition to keyboard typing, digital text production skills (e.g. simple word processing functions, navigation) are fundamental prerequisites to produce digital texts (Anskeit, 2022). Nevertheless, there is still a lack of comprehensive studies on the development of digital writing skills, especially in German-speaking countries and for primary school pupils (Gahshan & Weintraub, 2024; Schneider & Anskeit, 2017; Schüler et al., 2023). Addressing this gap, the project aims to develop instructional measures for digital writing and examine their effects on third-grade students’ text production.Building on a diagnostic laboratory study (n=16) using keystroke logging, the intervention study (n=121) investigates the effectiveness of a specially developed interactive learning pathway for promoting digital text production competences (keyboarding and word-processing functions) and compares it with a touch-typing course (focus on keyboarding). To evaluate both support measures, the typing behaviour (including speed and skills in simple word processing functions) of the learners will be assessed in a pre-post-test design using a procedure developed in the diagnostic study. In addition, effects on text quality (Lindauer, 2024) and text revision (Held, 2006) are analysed based on students’ independently written texts responding to a profiled writing task (Bachmann & Becker-Mrotzek, 2010).Initial results show that learners benefit from even short training sessions in terms of typing behaviour (see also Grabowski et al. 2007, Anskeit, 2022) and that the promotion of digital text production skills enables learners to utilise word processing functions. The extent to which this influences text quality and text revisions in the production of their own texts is determined using variance analyses (ANOVA with repeated measures) including covariates as reading comprehension and previous digital experience. The presentation will outline key findings from the diagnostic study, provide insights into the support material, and discuss the results of the intervention study.

Rubrics for Planning and Revising Argumentative Syntheses in Collaborative and Individual Settings

Type: Single Paper
Presenting author: Medina-Gutiérrez, María , Universidad Autonoma de Madrid, Spain
Co-author: Cuevas, Isabel , Universidad Autonoma de Madrid, Spain
Co-author: Olmos, Ricardo , Universidad Autonoma de Madrid, Spain
Co-author: Van Steendam, Elke , KU Leuven, Belgium
Co-author: Rijlaarsdam, Gert , University of Amsterdam, Netherlands
Co-author: Mateos, Mar , Universidad Autónoma de Madrid, Spain
Chair: Stukker, Ninke , Groningen University, Netherlands
Keywords: collaborative writing;intervention study;reading and writing;writing processes
Domain: Instructional Design; See Session Schedule

Abstract

Using Instructional Rubrics for Planning and Revising Argumentative Syntheses in Collaborative and Individual Settings: Effects on Text QualityMedina-Gutiérrez, M.*, Cuevas, I.*, Olmos, R*, van Steendam, E.**, Rijlaarsdam, G.*** & Mateos, M.*UAM*, KULeuven**, UvA***Integrating sources to write argumentative syntheses is a key academic skill, yet many undergraduates struggle, particularly during planning and revision (Vandermeulen et al., 2024). The current study examines the impact of an instructional rubric on the quality of students' argumentative synthesis tasks, with a special focus on reaching integrative conclusions, given their difficulties in integrating opposing perspectives through synthesizing strategies (Cuevas et al., 2024; Mateos et al, 2018). The effect of the rubric was analyzed after its use in two learning sessions focused on different stages of the writing process (planning and drafting and reviewing and revising) and delivered either individual (R+I) or collaborative (R+C) settings. A total of 101 undergraduates were assigned to three conditions (R+I, R+C, control) and wrote three argumentative syntheses, each based on two texts presenting opposing views on a topic (pretest-synthesis, intermedia-synthesis’ draft, revised intermediate-synthesis, and posttest-synthesis.). The rubric improved students’ learning, and these effects were already evident in the drafting phase and increasing marginally during the revision phase in individual settings. However, these effects were not greater under collaborative learning. Findings are discussed, and we conclude with recommendations for future research and educational implications.Keywords: argumentative synthesis, instructive rubric, collaborative setting, writing processes.ReferencesCuevas, I. Mateos, M., Casado-Ledesma, L., Olmos, R., Granado-Peinado, M., Luna, M., Núñez, J.A. y Martín, E. (2024). How to improve argumentative syntheses written by undergraduates using guides and instructional rubrics. European Journal of Psychology of Education, 39, 4573–4596. https://doi.org/10.1007/s10212-024-00890-xMateos, M., Martín, E., Cuevas, I., Villalón, R., Martínez, I., & González-Lamas, J. (2018). Improving written argumentative synthesis by teaching the integration of conflicting information from multiple sources. Cognition and Instruction, 36, 119–138. https://doi.org/10.1080/07370008.2018.1425300Vandermeulen, N., Van Steendam, E., De Maeyer, S., Lesterhuis, M & Rijlaarsdam, G (2024). Learning to write syntheses: the effect of process feedback and of observing models on performance and process behaviors. Reading dand Writing 37, 1375–1405. https://doi.org/10.1007/s11145-023-10483-7

Supporting peer feedback conversations during argumentative writing: rubric vs. conversation chart

Type: Single Paper
Presenting author: Landrieu, Yana , Ghent University, Belgium
Co-author: De Wever, Bram , Ghent University, Belgium
Co-author: Van Keer, Hilde , Ghent University, Belgium
Co-author: De Smedt, Fien , Ghent University, Belgium
Co-author: Bouwer, Renske , Utrecht University, Netherlands
Chair: Stukker, Ninke , Groningen University, Netherlands
Keywords: collaborative writing;intervention study;speaking and writing;writing tools and writing technology
Domain: Instructional Design; See Session Schedule

Abstract

Research topic/aim This dialogic writing study investigates how students’ peer feedback conversations can be supported during the revision phase of the collaborative writing process. Our research questions focus on whether providing students with a rubric or a conversation chart stimulates dialogic interaction and how these conversations relate to subsequent text revisions.Theoretical framework Grounded in Mercer and Wegerif’s (2002) and Bouwer’s and colleagues (2024) frameworks on exploratory talk, the study builds on research highlighting the collaborative potential of peer feedback during argumentative writing. While guidance is widely acknowledged as essential for effective peer feedback, little is known about which forms of support work best. This study examines the transition from oral peer feedback to written text revisions and explores whether provided peer feedback is (or is not) actually reflected in the subsequent text revisions.Methodology An intervention study was conducted with 102 students (aged 16–18) across eight lessons on argumentative writing. Using a pre-test post-test design, two conditions were compared: a rubric and a conversation chart condition. Data included peer feedback conversations analysed through content analysis and statistical tests: ANOVA, MANOVA, chi-square, and binary logistic regressions.Findings During peer feedback conversations, students primarily discussed quality of (counter)arguments and rebuttals. The conversation chart appeared to be most effective in fostering exploratory talk, particularly when combined with teacher intervention. However, transfer from dialogue to text revision was limited, indicating that peer feedback alone does not guarantee effective text revisions.Relevance This research addresses underexplored dimensions of writing: the collaborative nature of peer feedback and its connection to subsequent text revisions. Findings offer practical guidelines for integrating scaffolds and teacher support to enhance dialogic interaction and improve writing outcomes.ReferencesMercer, N., & Wegerif, R. (2002). Is exploratory talk productive talk? In K. Littleton & P. Light (Eds.), Learning with computers: analysing productive interaction (pp. 79–101). Bouwer, R., van Braak, M., & van der Veen, C. (2024). Dialogic writing in the upper grades of primary school: How to support peer feedback conversations that promote meaningful revisions. Learning and Instruction, 93. https://doi.org/10.1016/j.learninstruc.2024.101965

Text Features Associated with Students’ Generative AI Use: Norwegian Teachers’ Assessments

Type: Single Paper
Presenting author: Skar, Gustaf , Norwegian University of Science and Technology, Norway
Co-author: Jeffery, Jill , Leiden University Centre for Linguistics, Netherlands
Co-author: Kvistad, Anne Holten , Norwegian University of Science and Technology (NTNU), Norway
Co-author: Graham, Steve , Arizona State University, United States
Keywords: AI writing;assessment;Generative AI;writing instruction
Domain: -
Related Symposium: The future of writing education; See Session Schedule

Abstract

The release of generative AI (genAI) tools has changed the way that many educators interact with student writing, as they grapple with assessing how students use this technology for writing and how their uses may support or detract from learning. This paper draws from a survey of 530 Norwegian teachers designed to examine teachers’ perspectives on genAI, including their uses of AI to teach writing, their beliefs and ethical concerns about students’ AI use for writing, their preparedness to use AI, and, the focus of the current paper, the text features they associate with students’ AI use. GenAI presents new challenges for teachers’ writing assessment practice as it complicates their construction of the student author. Although written communication as academic assignment is skewed toward language performance to be assessed (Smagorinsky et al., 2010), a key aspect of the assessment process involves teachers’ interpretations of what a student is working to express in writing. Given that human communication is co-constructed, “it must follow that even when we don’t know the person who generated the language we are interpreting, we build a partial model of who they are and what common ground we think they share with us, and use this in interpreting their words” (Bender et al., 2021, p. 616). Many teachers are compelled to consider the extent to which their model of “the person who generated the language” is genAI-mediated. This paper focusses on a qualitative content analysis of an open survey item in which a subset of 129 teachers shared their perceptions of the text features that signal students’ use of generative AI and their stances toward these text features. We analysed teachers’ responses to investigate how they adapt their writing assessment practices in the context of students’ genAI use. We found that teachers viewed AI-associated text characteristics negatively, and they focused on language features indicative of voice and style when identifying aspects of student text that suggested AI use. Our results suggest that teachers’ individualized knowledge of students’ development vis-a-vis academic writing tasks and subject-matter learning factors into their judgments of whether a text is student-composed or AI-generated.

The future of writing education

Type: Symposium
Chair: De Smedt, Fien , Ghent University, Belgium
Organiser: Vereecke, Alix , Ghent University, Belgium
Organiser: Skar, Gustaf , Norwegian University of Science and Technology, Norway
Discussant: Bouwer, Renske , Utrecht University, Netherlands
Keywords: AI writing;assessment;Generative AI;writing instruction
Domain: Learning and Instructional Technology
Related Symposium: The future of writing education; See Session Schedule

Abstract

Writing has long been a cornerstone of education, serving both as a means of learning and as a key indicator of students’ understanding, reasoning, and communicative competence. Today, this foundational role is being challenged by the rapid emergence of generative artificial intelligence. From compulsory education to higher education, generative AI tools are increasingly influencing how learners engage with writing tasks, raising fundamental questions about authorship, originality, assessment, and the purposes of writing instruction itself. Rather than signaling the end of writing education, these developments invite a critical rethinking of writing education in an AI-rich educational landscape.This symposium brings together three research studies that collectively examine current developments in writing education in contexts where generative AI is increasingly embedded in educational practice. The first paper examines teachers’ detection of AI-generated text by exploring which textual features teachers associate with students’ use of generative AI. Drawing on survey data from Norwegian teachers, the study analyses how teachers interpret student writing and make judgments about authorship in contexts where generative AI is increasingly present. The second paper shifts attention from writing products to writing processes by examining how students’ interactions with generative AI can be used to inform the assessment of argumentative writing. It explores the potential of process data, such as prompts, revisions, and AI-mediated decision-making, as complementary evidence in writing assessment. The third paper focuses on higher education and investigates how generative AI can be integrated responsibly into students’ writing processes. It examines students’ existing uses of these tools and the role of instructional guidance in supporting critical, reflective, and autonomous writing practices.Taken together, the symposium offers a coherent and forward-looking view on the future of writing education, positioning generative AI not merely as a challenge, but as a resource that can inform and support writing processes.

THEtool: A software application for linguistic modeling of writing

Type: Demonstration Session
Presenting author: Ulasik, Margo , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Mahlow, Cerstin , ZHAW Zurich University of Applied Sciences, Switzerland
Chair: Pogner, Karl-Heinz , Copenhagen Business School, Denmark
Keywords: keystroke logging data processing and analysis;natural language processing (NLP);writing models;writing processes
Domain: -; See Session Schedule

Abstract

We present an open-source tool for analyzing writing process data in relation to linguistic structures: THEtool(https://github.com/mulasik/wta; Mahlow 2024; Ulasik and Miletic 2024; Ulasik et al. 2025). Although linguistic modeling of the writing process has gained importance in recent years, existing approaches, whether rooted in linguistic theory or writing research, remain insufficient to explain how writers actually produce and revise text at a linguistic level. THEtool enables writing researchers to investigate the writing process with a particular focus on sentences and their interaction with writing bursts and revisions. Because the software operates fully automatically and requires no manual intervention, it facilitates the efficient processing of large datasets. THEtool processes keystroke logging data in the XML-based IDFX format generated by Inputlog and ScriptLog, the de facto standard for storing and exchanging writing process data, thereby ensuring seamless integration with existing tools and workflows.To support a wide range of research applications, THEtool offers configurable key features, including language selection (currently German, Greek, French, and English, with straightforward extensibility to additional languages), the minimum pause duration that triggers the extraction of text and sentence versions within a writing burst, and relevance parameters for filtering text versions.THEtool is a fully functioning implementation of a model of text production based on the concept of layers: writing bursts, revisions, and sentence production are conceptualized as three distinct yet interacting layers that share a common timeline. Bursts may be interrupted by revision episodes or, in an abstract sense, by final punctuation marks signaling sentence completion. Revision processes can be interrupted by pauses or segmented by final punctuation. Likewise, sentence production may be interrupted by pauses or revisions. Projecting these layers onto one another enables new insights into the writing process from a linguistic perspective.We conducted exploratory studies in German, Greek, French, and English using THEtool. The results demonstrate both the feasibility and the analytical potential of the proposed approach.

Using Generative AI for Academic Writing: Students’ Practices and the Role of Explicit Instruction

Type: Single Paper
Presenting author: Vereecke, Alix , Ghent University, Belgium
Co-author: Bouwer, Renske , Utrecht University, Netherlands
Co-author: Van Keer, Hilde , Ghent University, Belgium
Co-author: De Wever, Bram , Ghent University, Belgium
Co-author: De Smedt, Fien , Ghent University, Belgium
Keywords: academic writing;AI writing;Generative AI;writing instruction
Domain: Higher Education
Related Symposium: The future of writing education; See Session Schedule

Abstract

Generative AI tools such as ChatGPT are now widespread in higher education and are often presented as promisingforms of support for academic writing, a complex skill that many students find challenging. While concerns about misuse and authorship persist, considerably less is known about how students actually use generative AI during the writing process, or about whether instructional guidance can support more responsible and effective use.This study adopts a two-phase design. In the first phase, a questionnaire study with 170 higher education students examined whether and how students use generative AI during writing, focusing on self-reported moments of use and the specific aspects of the writing process targeted. In the second phase, a pilot intervention study with 20 students explored students’ actual AI use in greater depth by comparing writing processes with and without explicit instruction on responsible AI use and prompting. Data sources included students’ prompts, generative AI conversations, and final texts, which were analysed using quantitative content analysis and comparative judgement to assess changes in prompt quality, revision practices, and overall text quality.Results from the questionnaire show that the vast majority of students (92%) report using generative AI during the writing process. However, students tend to use these tools in a limited manner, primarily for relatively straightforward tasks such as correcting language and spelling or reformulating existing text, rather than for more substantive support throughout the writing process. Moreover, self-report data provided only limited insight into students’ responsible use of AI. Findings from the pilot intervention study suggest that, in the absence of instruction, students do not consistently engage with generative AI in a responsible manner. Following explicit instruction, students formulated significantly higher-quality prompts and interacted more critically with AI-generated output. Although text quality improved for all students, no significant difference was found between students who did and did not receive instruction.Overall, the findings suggest that although generative AI is already widely used in academic writing, responsible and effective use cannot be assumed. Brief, targeted instruction on prompting and responsible AI use may therefore play a key role in supporting more meaningful integration of generative AI into students’ writing processes.

Writing on Paper or on Tablet? Error Patterns and Processing Time in Digital and Hybrid Formats

Type: Single Paper
Presenting author: Wörfel, Seda Yilmaz , Universität Hamburg, Germany
Co-author: Krelle, Michael , Universität Hamburg, Germany
Chair: Fitjar, Camilla , University of Stavanger, Norway
Keywords: evaluation and assessment of writing competence;spelling issues;writing at school;writing tools and writing technology
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Writing on Paper or on Tablet? Error Patterns and Processing Time in Digital and Hybrid FormatsRevised educational standards in Germany highlight the increasing relevance of digital competencies in school learning. The planned transition of standardized comparison tests to technology-based assessment (TBA) raises the question of how shifts from paper-and-pencil to digital formats affect orthographic performance. Given that handwriting and typing engage different cognitive and motor processes, digital formats may elicit distinct error types and correction strategies (Frahm, 2012; Jung et al., 2021). This underscores the need to examine how students adapt to these demands and how performance is influenced.To address this, two complementary studies were conducted. The first (HYBRID) investigated third- and fourth-grade students’ processing of orthographic tasks in a combined tablet–paper format. The second (DIGITAL) analyzed fully technology-based cloze tasks completed on tablets, with a focus on error patterns and processing time. Data from 100 primary school students were collected, drawing on synchronized screen and overhead video recordings to capture processing behavior.The comparison reveals systematic differences across formats. In the digital condition, students exhibited more comprehension-related hesitations and engaged in more orthographic correction attempts, whereas in the hybrid condition they more frequently undertook retrospective review of their written responses. Error frequency in the digital mode showed a positive correlation with processing time (rₛ = .33, p = .029), while no significant association emerged in the hybrid condition (rₛ = .14, p = .339). Quantitative analyses further indicate a higher overall error count in the hybrid mode.These findings underscore the need for closer examination of digital test formats. Beyond ensuring technological accessibility, schools must ensure didactic and diagnostic compatibility when integrating digital procedures into teaching and assessment.Literatur:Frahm, Sarah. 2012. Computerbasierte Testung der Rechtschreibleistung in Klasse Fünf - eine Empirische Studie Zu Mode-Effekten Im Kontext des Nationalen Bildungspanels. Berlin: Logos Verlag Berlin.Jung, Stefanie, Korbinian Moeller, Elise Klein, und Juergen Heller. 2021. «Mode Effect: An Issue of Perspective? Writing Mode Differences in a Spelling Assessment in German Children with and without Developmental Dyslexia». Dyslexia 27 (3): 373–410. https://doi.org/10.1002/dys.1675.

Self-feedback scaffolding through AI in online writing tasks

Type: Single Paper
Presenting author: Luna, Jose , University of Barcelona, Spain
Co-author: Guasch, Teresa , Open University of Catalonia (UOC), Spain
Co-author: Espasa, Anna , Universitat Oberta de Catalunya, Spain
Co-author: Mayordomo, Rosa M , Open University of Catalonia (UOC), Spain
Co-author: Melo, Montserrat Martinez , Fundació per a la Universitat Oberta de Catalunya (FUOC), Spain
Co-author: Martinez Melo, Montserrat , Fundació per a la Universitat Oberta de Catalunya (FUOC), Spain
Keywords: academic writing;feedback;Generative AI;writing tools and writing technology
Domain: Higher Education
Related Symposium: University students’ reflections on academic writing with genAI; See Session Schedule

Abstract

Students need to critically assess AI-generated feedback to avoid superficial learning (Bearman et al., 2024), particularly in writing processes where writing plays an epistemic role. A promising solution to enhance feedback practices with AI is to promote self-feedback processes. This is a process of cognitive change in which students generate new knowledge through comparing their current understanding or performance with external references, and its effectiveness relies on structured activities and scaffolding (Nicol, 2021). This study explores to what extent AI-supported self-feedback can effectively scaffold students’ writing in asynchronous environments. A total of 107 online students participated in a quasi-experiment. Students first completed an assignment. Immediately after submission, they accessed a timed online space. Following a reflective scaffolded process, students generated self-feedback while revising their initial assignment with AI insights. The quantitative analysis showed a significant improvement in students' scores from the first to the second submission (Z = -6.804; p < .001). Qualitative analyses of both students' interviews and writing reflections during the scaffolded process show that GenAI-mediated self-feedback is enacted through a set of recurrent actions. The reported self-feedback actions by students were: students primarily use GenAI to identify areas for improvement, revisit their understanding of key concepts, detect aspects they had overlooked, and connect their revisions to new knowledge. Interviews additionally reveal emergent topics that help to explain how students use GenAI. These include experimenting with prompting strategies to obtain more relevant feedback; directing corrections purposefully depending on their objectives; questioning GenAI’s reliability; experiencing uncertainty; and showing different levels of GenAI literacy. These results offer insights into the concrete mechanisms through which teachers can scaffold self-feedback process with GenIA in academic writing and contribute to the ongoing discussion on the potentials and dilemmas of GenAI in higher education. Bibliography Bearman, M., Tai, J., Dawson, P., Boud, D., & Ajjawi, R. (2024). Developing evaluative judgement for a time of generative artificial intelligence. Assessment & Evaluation in Higher Education, 49(6), 893–905. https://doi.org/10.1080/02602938.2024.2335321 Nicol, D. (2021). The power of internal feedback: Exploiting natural comparison processes. Assessment & Evaluation in higher education, 46(5), 756-778. https://doi.org/10.1080/02602938.2020.1823314

Students' reflections on using GenAI as a tool for cognition when writing an argumentative text

Type: Single Paper
Presenting author: Luna, Jose , University of Barcelona, Spain
Co-author: Fuertes-Alpiste, Marc , Universitat de Barcelona, Spain
Co-author: Castells, Núria , Faculty of Psychology, University of Barcelona, Spain
Co-author: Molas-Castells, Nuria , Universitat de Barcelona; Institut de Recerca en Educació (IREUB), Spain
Co-author: Alcaraz, Silvia , Universitat de Barcelona; Institut de Recerca en Educació (IREUB), Spain
Co-author: Lozano, Paula , Universitat de Barcelona; Institut de Recerca en Educació (IREUB), Spain
Co-author: Miño, Raquel , Universitat de Barcelona; Institut de Recerca en Educació (IREUB), Spain
Keywords: academic writing;AI writing;Generative AI;writing tools and writing technology
Domain: Higher Education
Related Symposium: University students’ reflections on academic writing with genAI; See Session Schedule

Abstract

This study aimed to analyze undergraduate students’ perceptions of the usefulness of Copilot as a tool for cognition (Fuertes-Alpiste, 2024) when writing argumentative texts. From this perspective, students are encouraged to use it as a mediational tool that supports problem solving in writing, to find new ideas, reviewing their texts in terms of content and language conventions, or helping them check citation formats when writing an argumentative text based on sources.A total of 152 undergraduate students from two education-related degree programs participated in a didactic sequence that included reading multiple texts, whole-group discussions, and the use of instructional guides with examples on how to write an argumentative text and how to employ different prompts with Copilot for this purpose. Students completed a questionnaire both before and after the didactic sequence.In the final questionnaire, students responded to Likert-scale items addressing the perceived usefulness and limitations of Copilot in supporting task completion, as well as items related to potential technical issues encountered when using the tool. Students were also asked open-ended questions about how using Copilot influenced their writing process, including ways in which it was helpful, unhelpful, or may have affected their autonomy, and were invited to provide examples.Preliminary results indicate that students value Copilot primarily as a tool for identifying ideas, revising their written texts, and including references. However, they also acknowledge the risk of becoming overly dependent on the tool when producing written documents, which they perceive as a potential threat to their creativity. These results can shed light on how generative AI tools can afford writing processes when used as tools for cognition and not as a substitute of students' cognition, eliciting their writing affordances and associated critical thinking skills. ReferenceFuertes-Alpiste, M. (2024). Framing Generative AI applications as tools for cognition in education. Pixel-Bit. Revista De Medios Y Educación, 71, 42–57. https://doi.org/10.12795/pixelbit.107697

Students’ reflections on academic writing in higher education: GenAI as sociomaterial actor

Type: Single Paper
Presenting author: Jusslin, Sofia , University of Helsinki, Finland
Co-author: Hort, Sofia , Mälardalen University; Åbo Akademi University, Sweden
Keywords: academic writing;AI writing;Generative AI;writing tools and writing technology
Domain: Higher Education
Related Symposium: University students’ reflections on academic writing with genAI; See Session Schedule

Abstract

This presentation addresses undergraduate students’ reflections on GenAI technologies and their role(s) in their academic writing, drawing from data from workshops with undergraduate students across scientific disciplines at a university in Finland. The study aims to explore how students conceptualize their academic writing in relation to GenAI technologies, drawing theoretically on sociomaterial frameworks using, for example, actor-network theory to understand writing as a process in which both human and non-human actors participate in shaping it (e.g., Clarke, 2002; Gourlay, 2015). The data encompasses audio-recorded conversations and mindmaps from four workshops (2,5 h each) with a total of 30 students in educational sciences, political science, and caring sciences. During the workshops, the students were tasked with mapping and discussing what they use in their academic writing, how, when, and why. No question was asked explicitly about GenAI. Nevertheless, the students discussed GenAI technologies in all workshops, sharing that they use various AI technologies, such as ChatGPT, co-pilot, and Gemini. Preliminary analyses indicate that the students use them, for example, as support when their writing processes become stalled, when needing to expand the amount of text or generate new perspectives, and to orient themselves in relevant literature. A prominent use of GenAI technologies is that they, in similar manners as for example dictionaries and thesauruses, can be used in the writing to adapt the text to the linguistic and stylistic norms that apply within their disciplines. As such, GenAI technologies often have, according to the students, other, more central functions than merely a text generator. This presentation will unfold the results of the study and discuss implications for writing with GenAI in higher education. ReferencesClarke, J. (2002). A new kind of symmetry: Actor-network theories and the new literacy studies. Studies in the Education of Adults, 34(2), 107–122. https://doi.org/10.1080/02660830.2002.11661465Gourlay, L. (2015). Posthuman texts: Nonhuman actors, mediators and the digital university. Social Semiotics, 25(4), 484–500. https://doi.org/10.1080/10350330.2015.1059578

University students’ reflections on academic writing with genAI

Type: Symposium
Chair: Jusslin, Sofia , University of Helsinki, Finland
Organiser: Jusslin, Sofia , University of Helsinki, Finland
Discussant: Van Steendam, Elke , KU Leuven, Belgium
Keywords: academic writing;AI writing;Generative AI;writing tools and writing technology
Domain: Higher Education
Related Symposium: University students’ reflections on academic writing with genAI; See Session Schedule

Abstract

The aim of this symposium is to address and discuss undergraduate students’ reflections on academic writing with generative artificial intelligence (GenAI). Academic writing is central to studies in higher education, and since OpenAI’s launch of ChatGPT in November 2022, the possible potentials and challenges of using generative artificial intelligence (GenAI) technologies in writing have been increasingly discussed and explored across scientific fields (e.g., Khalifa & Albadawy, 2024; Nguyen, 2024). Previous research has shown that GenAI has been described in different ways; in addition to a text generator, also as an assistant, tutor, teacher, and conversation partner, which makes a difference for students’ performance and constitutes an affective support (Kim et al., 2025; Ou et al., 2024). Several studies have explored undergraduate students’ perceptions on GenAI in writing, soliciting responses through interviews and surveys (e.g., Kim et al., 2025; Ou et al., 2024). Adding to this body of work, the presentations in this symposium offer other perspectives on undergraduate students’ academic writing with GenAI, using various theoretical perspectives, research designs, and methods. First, focus lies on students’ peer-reflections on academic writing, where they discussed GenAI as part of their academic writing without being specifically asked about GenAI. Second, focus lies on students’ reflections on engaging in academic writing tasks using GenAI, more specifically, self-feedback scaffolding through GenAI in online writing tasks and GenAI as a tool for cognition when writing argumentative texts. Thus, the symposium adds to ongoing discussions of potentials, challenges, and dilemmas that GenAI technologies present for academic writing in higher education. ReferencesKhalifa, M., & Albadawy, M. (2024). Using artificial intelligence in academic writing and research: An essential productivity tool. Computer Methods and Programs in Biomedicine Update, 5, 100145. https://doi.org/10.1016/j.cmpbup.2024.100145Kim, J., Yu, S., Detrick, R., & Li, N. (2025). Exploring students’ perspectives on Generative AI-assisted academic writing. Education and Information Technologies, 30(1), 1265–1300. https://doi.org/10.1007/s10639-024-12878-7Nguyen, A., Hong, Y., Dang, B., & Huang, X. (2024). Human-AI collaboration patterns in AI-assisted academic writing. Studies in Higher Education, 49(5), 847–864. https://doi.org/10.1080/03075079.2024.2323593Ou, A. W., Stöhr, C., & Malmström, H. (2024). Academic communication with AI-powered language tools in higher education: From a post-humanist perspective. System, 121, 103225. https://doi.org/10.1016/j.system.2024.103225

Integrating ChatGPT into EFL Writing Instruction: Effects of Teacher Modelling and Autonomous Use

Type: Single Paper
Presenting author: Van Steendam, Elke , KU Leuven, Belgium
Co-author: Le Thanh, Thao , KU LEUVEN, Belgium
Co-author: Sercu, Lies , Katholieke Universiteit Leuven, Belgium
Co-author: Hoang, Yen Phuong , Can Tho University Vietnam, Vietnam
Chair: Pogner, Karl-Heinz , Copenhagen Business School, Denmark
Keywords: academic writing;AI writing;intervention study;writing instruction
Domain: Instructional Design; See Session Schedule

Abstract

Artificial intelligence (AI) is no longer peripheral to writing education; it is embedded in learners’ everyday composing practices, yet a key question remains: how should AI be effectively integrated to support complex genres such as argumentative writing? While prior research highlights AI’s potential for localized feedback and revision, intervention studies comparing integration designs for producing full essays within established instructional frameworks are scarce. In EFL contexts, where linguistic and rhetorical demands compound cognitive load (Hyland, 2019), teacher modelling, making expert strategies visible across planning, drafting, revising, and self-regulation (Graham & Perin, 2007; Schunk & Zimmerman, 1998), offers a benchmark for evaluating AI-supported instruction. What remains unclear is whether AI can serve as a productive modelling partner, how it compares to modelling without AI, and whether autonomous AI use fosters sustained gains in text quality.To address this question, we set up a pretest-posttest experimental study with 130 Vietnamese EFL undergraduates completing a four-lesson sequence on argumentative writing aligned with Schunk and Zimmerman’s (1998) self-regulated skill acquisition model. Three conditions were implemented: (1) Teacher Modelling + ChatGPT (TM+GPT), where the teacher thought aloud while prompting and critiquing ChatGPT output; (2) Teacher Modelling only (TM), replicating strategy instruction without AI; and (3) Autonomous Learning + ChatGPT (AL+GPT), where students engaged ChatGPT independently as a writing coach. A mixed-method design captured (a) screen-capture and keystroke logs for processes, (b) writing products for text quality, and (c) questionnaires on perceptions. This paper focuses on the product-level question: What is the effect of ChatGPT-integrated instruction on text quality? Results show that TM+GPT produced the highest text-quality scores, outperforming both AL+GPT and TM. These findings suggest that AI yields the greatest benefit when embedded within explicit teacher modelling that scaffolds prompt design, critical evaluation of AI output, and alignment with rhetorical goals, rather than when students use AI autonomously or when instruction excludes AI. implications for integrating AI as a mediated modelling partner in EFL writing curricula will be discussed.

Prompt – write – revise – repeat: a writing-process study of AI-assisted writing in higher education

Type: Single Paper
Presenting author: Weder, Mirjam , University of Basel, Switzerland
Chair: Mahlow, Cerstin , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: academic writing;AI writing;cognitive processes;writing processes
Domain: Cognitive Science; See Session Schedule

Abstract

With the widespread adoption of generative AI for (academic) writing, established models of the writing process such as Hayes (2012) need to be re-conceptualized. It has been suggested that writing could be viewed as a “co-activity of humans and machines” (Steinhoff 2023, Brommer & Rezat in print).To date, extensive survey-based research documents students’ AI use in higher education based on self-reports (cf. Ravšelj et al. 2025), whereas observational studies examining how students shape and appropriate human-AI co-activity in writing processes remain scarce (cf., however, Jelson et al. 2025).This study aimed to investigates writing strategies students use in AI-assisted writing, in particular, how students adapt and combine sub-processes, such as prompting, treatment of the AI output, AI-assistant revision, and their own revisions, and how different strategies impact the characteristics and quality of texts. To this end, several data-collection instruments were used: screen capture (OBS Studio) and keystroke logging (Leijten & Van Waes 2013) to record text production processes and the interaction between human input and AI output; stimulated recall (Gass 2000) to capture (meta-)cognitive processes; and a short questionnaire on AI-supported writing strategies and participants’ self-efficacy beliefs.The paper reports on a study comprising 12 writing sessions with students of German studies who varied in their experience with academic writing and AI use, testing the combination of methods and exploring writing processes and strategies with the aim of developing a category system for their description and analysis. ReferencesBrommer, S., Rezat, S. (pre-print). Mensch-KI-Interaktion beim Schreiben – Theoretische Überlegungen zur Modellierung des Schreibprozesses. In: Weder, M., Bubenhofer, N. (eds.): Schreiben mit KI. transcipt.Hayes, J. R. (2012). Modeling and Remodeling Writing. In: Written Communication 29, 369–388. Leijten, M., Van Waes, L. (2013). Keystroke Logging in Writing Research: Using Inputlog to Analyze Writing Processes. Written Communication 30(3), 358-392. Ravšelj, D., et al. (2025). Higher education students’ perceptions of ChatGPT: A global study of early reactions. In: PLOS ONE, 20. https://doi.org/10.1371/journal.pone.0315011.Steinhoff, T. (2023): Der Computer schreibt (mit). Digitales Schreiben mit Word, Whatsapp, ChatGPT & Co. als Koaktivität von Mensch und Maschine. In: MiDU-Medien im Deutschunterricht, IDSL II. (1), 1–16.

A Direct Approach to the Study of Epistemic Decisions: Students Using AI for Thesis Writing

Type: Single Paper
Presenting author: Kruse, Otto , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Milićević, Dejana , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Rapp, Christian , ZHAW Zurich University of Applied Sciences, Germany
Keywords: academic writing;AI writing;cognitive processes;Generative AI
Domain: -
Related Symposium: Accessing the Epistemological Side of Writing: A Prolegomenon to the Era of AI; See Session Schedule

Abstract

Understanding how students make epistemic decisions when using AI technologies for academic writing requires methodological approaches that can capture the nuanced intellectual and rhetorical processes underlying their choices. While existing research has documented patterns of AI adoption and usage frequencies, there remains a significant gap in our understanding of the detailed thinking processes that guide students' decisions about when, how, and why to incorporate AI-generated content into their scholarly work. This study addresses this methodological challenge through a qualitative interview-based approach designed to access students' reflective accounts of their AI use experiences during thesis writing. As a contribution to get methodological access to AI use, this contribution reports from a larger study including three countries (Switzerland, Romania, Bulgaria) to interview students about their experiences with AI. The cross-national design allows for comparative insights into how different educational contexts and cultural backgrounds may shape students' approaches to AI integration in academic writing. The background problem of this is that we currently have many surface descriptions about AI use, but little understanding of the finer-grained thinking moves involved. Existing survey and usage data tell us what students do with AI, but not how they think through the complex decisions about knowledge construction, source integration, and authorial voice that AI use entails. Pilot interviews have been conducted with undergraduate and graduate students currently writing their theses. The interview protocol focuses on eliciting detailed narratives about specific instances of AI use, prompting students to articulate their decision-making processes, and exploring their conceptions of authorship, originality, and epistemic development in AI-assisted writing contexts. We will describe the questions that proved to be useful and summarize our experiences with this direct way of questioning students. Key results will be presented along with recommendations for interview strategies that successfully access students' epistemic reasoning in AI-assisted thesis writing.

Accessing the Epistemological Side of Writing: A Prolegomenon to the Era of AI

Type: Symposium
Chair: Kruse, Otto , ZHAW Zurich University of Applied Sciences, Switzerland
Organiser: Milićević, Dejana , ZHAW Zurich University of Applied Sciences, Switzerland
Discussant: Schnatz, Nina Isabella , Zurich University of Applied Sciences (ZHAW), Switzerland
Keywords: academic writing;AI writing;corpus linguistics;Generative AI
Domain: Learning and Instructional Technology
Related Symposium: Accessing the Epistemological Side of Writing: A Prolegomenon to the Era of AI; See Session Schedule

Abstract

Traditionally, the study of writing has focused on rhetorical, linguistic, cultural, social, and process-related dimensions. The epistemological side of writing, however, has been left to the disciplines as they oversee their respective fields of knowledge. Rarely do we directly consider students’ conceptions of truth and their understanding of knowledge or knowing in the way William Perry (1970) has addressed it. Even if academic writing may be seen as the best way of developing epistemic consciousness, the term itself not often the focus of research, and the broad range of intellectual skills behind it remains hidden. We are aware, however, that every topical sentence demands complex judgements about its appropriateness and needs justification of its assumed truth. Such epistemic activities demand understanding of what is considered valid knowledge, how it is produced, what epistemic conventions exist, and how epistemic authority is built in a certain discipline. Beyond all this, the conception of truth is a nut that is hard to crack, not only for our students but also for philosophy. We are used to confusing our students by insisting that they rely on facts but should not believe in absolute truth. How do these two demands go together?With the inclusion of generative AI in writing processes, a new algorithmic “voice” enters the field that also requires epistemic framing. However, this voice has different qualities and shortcomings compared to human epistemological consciousness during writing. Writers must evaluate their own thoughts against the AI-generated content, which presents new challenges, particularly for beginners.This symposium introduces the concept of epistemic consciousness in writing. It presents several methodological approaches, manifested in four specific research projects, to bring to the surface epistemic processes involved in academic writing. Presenters will explain the logic of the enquiry in each project along with some initial results. We intend for the symposium to stimulate new avenues for research, contributing to the exploration of human–AI interaction in writing and thinking.Project 1: Qualitative InterviewsProject 2: How Expert and Novice Academics Write with GenAI: Think-Aloud ProtocolsProject 3: Corpus Linguistic Methods

Can Algorithm-based Feedback Help Students to Write Better? A Meta-analysis

Type: Single Paper
Presenting author: Scherer, Sina Verena , University of Münster, Germany
Co-author: Graham, Steve , Arizona State University, United States
Co-author: Busse, Vera , University of Münster, Germany
Chair: Weth, Constanze , Institute for Research on Multilingualism, University of Luxembourg, France
Keywords: AI writing;EFL and ESL writing;L1;meta-analysis
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Against the backdrop of rapid developments of algorithm-based feedback tools - from older tools mainly providing feedback on grammar and spelling to more advanced tools based on generative artificial intelligence offering more comprehensive writing support - our meta-analysis examines to what extent algorithm-based feedback improves not only surface- (e.g., grammar and spelling) but also deep-level (e.g., structure, content, coherence) writing outcomes for different (language) learners (first, second, and foreign language learners) at secondary school and university. Algorithm-based feedback tools may be very useful for language learners as they can provide timely feedback and help with revision (Escalante et al., 2023), which can be particularly relevant for foreign language (FL) learners who often have limited contact with first language (L1) speakers outside the language classroom, as opposed to second language (L2) learners.For this meta-analysis, we reviewed experimental and quasi-experimental studies published between 2011 and the end of 2024, covering five European languages in four different databases. Results from the 33 included studies indicated that algorithm-based feedback was beneficial for improving writing in general (g = 0.36). Specifically, positive effects were observed for surface-level outcomes at post-test (g = 0.31), though no lasting effects were found at maintenance (g = -0.02). In contrast, deep-level writing outcomes showed sustained improvement, with positive effects both at post-test (g = 0.31) and maintenance (g = 0.54). No significant differences between secondary and university students were observed. However, L2 learners, in general, seemed to profit most from algorithm-based feedback, showing gains in surface- (g = 0.77, bordering on significance), and deep-level outcomes (g = 0.46). While no significant differences were found between the effects of specific types of algorithm-based feedback tools in moderator analyses, feedback from Grammarly and Pigai statistically enhanced students’ writing but effects of ChatGPT feedback were non-significant. We discuss implications for future research and educational practice, also in light of the small transfer of learning from algorithm-based feedback to new writing tasks.ReferencesEscalante, J., Pack, A., & Barrett, A. (2023). AI-generated feedback on writing: insights into efficacy and ENL student preference. International Journal of Educational Technology in Higher Education, 20(1). https://doi.org/10.1186/s41239-023-00425-2

Encoding the Writing Process: TEI Between Research and Computational Use

Type: Roundtable
Presenting author: Ammari, Emna , Aix-Marseille Université, France
Presenting author: Bekius, Lamyk , University of Antwerp, Belgium
Presenting author: Darrak, Zahra Ed , Aix-Marseille University, France
Presenting author: Quaranta, Jean-Marc , Aix-Marseille University, France
Keywords: cognitive processes;Generative AI;keystroke logging data processing and analysis;writing process data collections
Domain: -; See Session Schedule

Abstract

The Text Encoding Initiative (TEI) has long been used in digital humanities to encode manuscripts and historical documents, primarily focusing on textual products. More recently, TEI has been applied to the encoding of the writing process itself (Bekius, 2023), opening new possibilities for integrating genetic criticism, writing studies, and process-oriented research.As an open and extensible XML-based markup language, TEI is a promising candidate for encoding not only manuscripts, but also born-digital writing processes, shifting the focus from documents to writing sessions and dynamic trajectories of text production. Such an approach enables new and potential applications, including the visualization of writing dynamics (e.g. through tools such as Keystroke Loxensis (Bekius 2024) as part of the eXtant toolkit) or the creation of structured datasets for computational analysis and artificial intelligence systems.Even though TEI could ensure interoperability across projects and disciplines, its complexity and verbosity raise concerns when applied to large-scale or fine-grained writing process data, such as keystroke logs. Encoding long writing sessions at a micro-level can present problems related to elements over-lapping, as well as being time-consuming and cognitively demanding.This roundtable explores this tension by asking whether TEI can realistically function as a standard for writing process research, and under what conditions. Key questions for discussion include: Is TEI suited to represent writing dynamics captured through log files? What alternatives or hybrid solutions might exist? Can parts of the encoding process be automated? A central focus will be the selection problem: which process data is actually relevant to encode, particularly when studying creativity in writing? An additional perspective from computer science will consider whether TEI-based representations of writing processes can function as inputs for artificial agents designed to reproduce an author’s writing style and creative dynamics.Bekius, Lamyk. (2023). Behind the Computer Screens: The use of keystroke logging for genetic criticism applied to born-digital works of literature. [PhD Dissertation Antwerp University & University of Amsterdam]. https://pure.uva.nl/ws/files/139150661/thesis.pdf.Bekius, Lamyk. (2024). ‘Nanogenetic econarratology : where narratology meets keystroke logging data’, in Van Hulle, Dirk (éd.), Genetic Narratology: Analysing Narrative Across Versions, Cambridge, Open book publishers, 2024.Workgroup on Genetic Editions. (2010). ‘An Encoding Model for Genetic Editions’. https://tei-c.org/Vault/TC/tcw19.html.

Evaluating Writing Quality of Engineering Student Reports using Natural Language Processing Tools

Type: Single Paper
Presenting author: Rybicki, Jan-Mikael , Aalto University, Finland
Co-author: Mendoza, Laura , University of Helsinki, Finland
Chair: Weder, Mirjam , University of Basel, Switzerland
Keywords: academic writing;assessment;EFL and ESL writing;natural language processing (NLP)
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Research topic, area of investigation and aimIn higher education, writing instructors evaluate the quality of student texts and provide formative feedback on their writing. This laborious work could be supported using automatic Natural Language Processing (NLP) tools. Much research on the indices produced by NLP tools and the quality of writing has focused on essay writing. However, little research has explored report writing in science and engineering domains. To address this gap, this study investigates the association between the NLP indices and holistic human ratings of academic reports written by English as a Second Language (ESL) students in a master’s level computer science course.Methodological designData consists of 100+ academic reports (average length approx. 2800 words, excluding references), which were evaluated by writing instructors. Multiple regression analyses were conducted to identify NLP indices that predict the holistic instructor ratings of student reports.FindingsThe preliminary findings indicate that a regression model combining TAACO (Crossley et al., 2019), TAALED (Kyle et al., 2021), TAALES (Kyle et al., 2018) and TAASSC (Kyle, 2016) indices predicts nearly 45% of variance in holistic ratings.Relevance to domain of writingThe findings of this study extend earlier writing research to a new context and genre, i.e., longer engineering texts, and offers insights into the usability of NLP tools in writing instruction.ReferencesCrossley, S. A., Kyle, K., & Dascalu, M. (2019). The Tool for the Automatic Analysis of Cohesion 2.0: Integrating Semantic Similarity and Text Overlap. Behavioral Research Methods 51(1), pp. 14-27. https://doi.org/10.3758/s13428-018-1142-4Kyle, Kristopher, “Measuring Syntactic Development in L2 Writing: Fine Grained Indices of Syntactic Complexity and Usage-Based Indices of Syntactic Sophistication.” Dissertation, Georgia State University, 2016. https://doi.org/10.57709/8501051Kyle, K., Crossley, S. A., & Berger, C. (2018). The Tool for the Analysis of Lexical Sophistication (TAALES): Version 2.0. Behavior Research Methods 50(3), pp. 1030-1046. https://doi.org/10.3758/s13428-017-0924-4Kyle, K., Crossley, S. A., & Jarvis, S. (2021). Assessing the Validity of Lexical Diversity using Direct Judgements. Language Assessment Quarterly 18(2), pp. 154-170. https://doi.org/10.1080/15434303.2020.1844205

How Expert and Novice Academics Write with GenAI: Think-Aloud Protocols

Type: Single Paper
Presenting author: Kessler, Karyn , University of California Santa Barbara, United States
Co-author: Rogers, Paul , University of Caifornia, Santa Barbara, United States
Co-author: Anson, Chris , North Carolina State University, United States
Co-author: Michiels, Paul , Carnegie Mellon University, United States
Keywords: cognitive processes;Generative AI;metacognition;protocol analysis
Domain: -
Related Symposium: Accessing the Epistemological Side of Writing: A Prolegomenon to the Era of AI; See Session Schedule

Abstract

Two related studies aim to track the infusion of GenAI into knowledge generation and diffusion processes among expert and novice academic writers across disciplines working on authentic revision tasks in writing. The first study examines experienced academic researchers and writers from diverse disciplinary backgrounds, including humanities, social sciences, and STEM fields. Using Zoom-based think-aloud methods along with keyboard tracking, the study captures real-time data on writers' cognitive processes and writing behaviors as they interact with GenAI systems. The think-aloud protocols highlight the ways in which and the degrees to which GenAI influences experienced writers' metacognitive and revision processes, epistemic development, and agency across domains of knowledge (Tardy, 2009; Kessler et al., 2026). By focusing on authentic revision tasks rather than artificial laboratory settings, the research ensures ecological validity and provides insights into actual scholarly practices. Results indicate the ways in which today's highly effective thinkers and knowledge producers incorporate (or don't) GenAI into their research and research writing practices. In the second study, undergraduate students used ChatGPT to assist them in writing 100-word literacy narratives focusing on a specific moment in their literate history. They then revised the output based on how effectively it captured their rhetorical, stylistic, and content-related intentions. Their entire process was recorded using screencast technology as they spoke their processes aloud. After finalizingtheir texts, they wrote a brief reflection on the experience. This contribution will present a thematic and code-based analysis of the epistemic decisions students made in their revisions of the outputs, with implications for reforming methods for supporting writing in the age of generative AI. Taken together, the two studies reveal differences between the epistemic processes of experienced and novice writers and suggest a developmental continuum for instruction in the use of generative AI in writing tasks.

Make Wraiting Great: Why Writing Still Matters in the Age of AI

Type: Roundtable
Presenting author: Baaijen, Veerle , CLCG University of Groningen, Netherlands
Presenting author: Johansson, Victoria , Kristianstad University, Sweden
Presenting author: Van Steendam, Elke , KU Leuven, Belgium
Presenting author: Wengelin, Åsa , University of Gothenburg, Sweden
Co-author: Young, Justin , University of Antwerp, Belgium
Keywords: AI writing;cognitive processes;Generative AI;writing theory
Domain: Motivational, Social and Affective Processes; See Session Schedule

Abstract

In an era dominated by artificial intelligence, the act of writing is often perceived as a skill that can be delegated to large language models (Pack & Maloney, 2023). Yet, writing remains essential for literacy development, cognitive development, and active participation in society. This roundtable invites researchers to explore why writing—also in times of generative AI – remains indispensable for fostering critical thinking, creativity, learning, and communication skills (Chang & Lee, 2025), while also discussing how writing with AI can shape our understanding of what writing is and can be, and how AI-supported writing may help struggling writers express their views in linguistic forms that would otherwise be inaccessible to them (Kasneci et al., 2023). We will discuss how writing cultivates deeper cognitive processes, such as reflection, revision, and synthesis – skills which are essential for participation in our complex literate societies. Writing empowers individuals to articulate ideas, engage in meaningful dialogue, and contribute to societal discourse. Hence, while AI tools can assist in generating and revising text, they should not replace the cognitive work of writing. Finally, this roundtable will examine the role of writing in promoting digital literacy and responsible use of AI, and suggest how writing with AI may change our theoretical descriptions of writing. Participants will share strategies for integrating writing into educational and professional settings, ensuring it remains a vital tool for cognitive and personal development. By highlighting the unique value of human writing, we aim to inspire a renewed commitment to nurturing writing as a fundamental skill in the AI age.

The linguistic impacts of generative AI on L2 writing output

Type: Single Paper
Presenting author: Meier, Ramona , Ghent University, Belgium
Co-author: Macken, Lieve , Ghent University, Belgium
Co-author: De Wilde, Vanessa , Ghent University, Belgium
Chair: Weth, Constanze , Institute for Research on Multilingualism, University of Luxembourg, France
Keywords: acquisition;EFL and ESL writing;Generative AI;writing and linguistics
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

In recent years, research on generative AI (GenAI) and its use for language learning has proliferated, highlighting affordances of the tools, while remaining conscious of potential limitations (Warschauer et al., 2023). Previous work on the use of GenAI tools for L2 English writing has explored the roles ChatGPT can fulfil by employing mainly (quasi-)experimental designs where AI training was provided (e.g. Fang & Han, 2025). However, there is a lack of work focusing on preexisting GenAI usage patterns in EFL students and their effect on L2 writing outcomes. While previous studies focus on the role of GenAI and its potentials, the impacts of such tools on linguistic factors, specifically in synthesis writing, remain underexplored (Yoo, 2025). This study aims to broaden our understanding of students’ preexisting GenAI practices and their impacts on synthesis writing. Participants in this cross-sectional study will complete a synthesis writing task twice (with and without GenAI). Screen recordings, semi-structured interviews, and measures of complexity, accuracy, and fluency (CAF) will be used to analyze their practices, engagement, and language. We expect to find improved performance on the GenAI-assisted task, potentially dependent on the methodical use of GenAI throughout the process, leading to more complex, accurate, and fluent texts. Theoretical and pedagogical implications of the study will also be discussed during the presentation. Keywords: GenAI, EFL learning, L2 writing development, CAF References Fang, S., & Han, Z. H. (2025). On the nascency of ChatGPT in foreign language teaching and learning. Annual Review of Applied Linguistics, 45, 253-273. Warschauer, M., Tseng, W., Yim, S., Webster, T., Jacob. S, Du, Q., Tate, T. (2023). The affordances and contradictions of AI-generated text for writers of English as a second or foreign language. Journal of Second Language Writing 62, Article 101071. Yoo J. (2025). Reading-Writing Connections: A Systematic Review of Second Language Synthesis Writing. L2 Journal: An Open Access refereed Journal for World Language Educators, 17(1), 1-55.

AI and Students' Academic Writing of Theses – Independent Work in Teacher Education

Type: Roundtable
Presenting author: Blomqvist, Per , Stockholm University, Sweden
Keywords: academic writing;assessment;Generative AI;writing processes
Domain: Higher Education; See Session Schedule

Abstract

AI and Students' Academic Writing of Theses – Independent Work in Teacher EducationGenerative AI is transforming the conditions for teaching and assessing students' academic writing. This is particularly relevant for various types of theses that are written over extended periods, where students are expected to develop independence as well as abilities in analytical, creative, and critical thinking (Magnusson & Zackariasson, 2019).Since the spring of 2025, a research and development project has been underway at Stockholm University within the primary teacher education program. The project aims to test and evaluate new methods and approaches for mentoring, teaching, and assessing students' academic writing in the course on Independent Work, with regard to the use of generative AI.The questions that the project seeks to answer are:How and in which parts of the writing process can AI tools be beneficial in developing students' independence and capacities for analytical, creative, and critical thinking?How and in which parts of the writing process can AI tools pose obstacles to developing these abilities?How does students' use of AI affect the ability of supervisors, teachers, and examiners to assess students' knowledge and skills in relation to the expected learning outcomes of the courses?The project involves five researchers from the Department of Teaching and Learning, along with approximately 120 students who are writing their theses in pairs over a ten-week period.In the project, teacher-produced educational materials, such as lesson plans and instructions, as well as students' formal and informal writing, including work logs, drafts, and evaluations, are documented. This documentation is utilized to illuminate changes in writing assignments, namely teachers' planning, implementation, and evaluation of teaching and assessment, in relation to students' opportunities to develop their academic writing, focusing on their ability for independent analytical and critical thinking in the context of generative AI use.During the roundtable discussions, I aim to explore these questions with other researchers and educators. The roundtable will begin with a presentation of the questions posed by the project and the actions taken in relation to them.

Digital writing and writing motivation

Type: Single Paper
Presenting author: Keller, Franziska , PH FHNW, Switzerland
Presenting author: Sturm, Afra , University of Applied Sciences and Arts Northwestern Switzerland PH (FHNW), Switzerland
Chair: Haake, Lisa , Institute of Educational Science, University of Münster, Germany
Keywords: fluency;motivational and emotional factors in writing;text composition;writing tools and writing technology
Domain: Learning and Social Interaction; See Session Schedule

Abstract

Writing is more than the ability to write a text: writing is embedded in a literacy practice, writers are part of a writing or literacy community (Graham, 2018). Digital writing platforms like myMoment (designed for grade 3 to 6) can provide students with a broader audience, strengthen their sense of ownership over their writing and increase their writing motivation. In our study, we examine how the communicative function as one form of writing motivation can be assessed, how this relates to writing competencies, and how writing motivation changes over the course of writing with myMoment.In our baseline survey with 157 students, we were able to replicate Graham et al.’s (2019) scale measuring students’ attitudes toward writing, as well as seven of the eight subscales of writing motivation from Graham et al. (2022). We complemented these scales with a communication-as-writing-motivation scale, as no such measurement has yet been suggested in the research literature. Our newly developed writing motivation subscale demonstrates an internal consistency of α = 0.78 (n = 148), and it correlates significantly and (predominantly) positively with writing fluency (p < 0.001, r = 0.272), as well as with narrative text quality (p = 0.032, non-linear relationship). The other writing motivation subscales we tested also correlate significantly with our writing performance data, but with either only writing fluency or only narrative text quality. Furthermore, we will present results on the development of this relationship between writing motivation and writing performance during the use of the digital writing platform myMoment, with a focus on struggling and advanced writers. Graham, S. (2018). A Revised Writer(s)-Within-Community Model of Writing. Educational Psychologist, 53(4), 258–279.Graham, S., Harris, K. R., Fishman, E. et al. (2019). Writing Skills, Knowledge, Motivation, and Strategic Behavior Predict Students’ Persuasive Writing Performance in the Context of Robust Writing Instruction. knowledge, 24.Graham, S., Harbaugh-Schattenkirk, A. G., Aitken, A. et al. (2022). Writing motivation questionnaire: validation and application as a formative assessment. Assessment in Education: Principles, Policy & Practice, 29(2), 238–261.

Effort, Agency, and Authorship in AI-Assisted Writing: Revisiting Flower & Hayes’ Model

Type: Single Paper
Presenting author: Büttgen, Katrin , Zuyd University of Applied Sciences, Netherlands
Co-author: Gruwel, Saskia Brand - , Maastricht University, Netherlands
Co-author: Pluymaekers, Mark , Zuyd University of Applied Sciences, Netherlands
Chair: Conijn, Rianne , Eindhoven University of Technology, Netherlands
Keywords: AI writing;cognitive processes;pedagogy and instruction;writing tools and writing technology
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Effort, Agency, and Authorship in AI-Assisted Writing: Revisiting Flower & Hayes’ ModelGenerative AI tools are reshaping the cognitive and rhetorical processes of writing.This study re-examines Flower and Hayes’ (1980) model of planning, translating, and revising through the lens of AI-assisted composition. Drawing on Cognitive Load Theory (Sweller et al., 2011) and frameworks of writer identity (Ivanič, 1998; Hyland, 2002), it investigates how AI intervention influences students’ perceived effort, agency, and authorship during academic writing. Unlike earlier work that conceptualised human–AI co-writing in general terms, this study provides phase-specific, empirical evidence of how effort, agency, and authorship shift across planning, translating, and revising – linking perceived ease to observed shifts in germane effort and agency.Eighty student reflections formed the primary dataset. From these, fifteen students were purposively sampled for semi-structured interviews, with a pre-specified saturation stopping rule. A small exploratory sub-sample will complete concurrent think-alouds to trace process-level decisions. This triangulation captured cognitive, experiential, and interpretive dimensions of the writing process. Thematic analysis traces how students negotiate agency and authorship across recursive phases of writing – delegating cognitive effort to the tool in some moments while reclaiming control over content and phrasing in others. Preliminary findings suggest that perceived ease may conceal a shift in cognitive engagement: when writing feels effortless, germane effort in idea development and revision is displaced to the tool. This cognitive offloading alters agency, shifting it from intentional decision-making to editorial supervision, while moments of reflective intervention reveal emerging co-agency and rhetorical awareness. The paper argues that AI does not erase authorship but redistributes it across human–machine collaboration, offering phase-specific insights to inform pedagogy that maintains germane effort and cultivates deliberate authorial agency.

Navigating the double bind: how AI reshapes financial analysts’ writing practices

Type: Single Paper
Presenting author: Whitehouse, Marlies , ZHAW Zurich University of Applied Sciences, Switzerland
Chair: Conijn, Rianne , Eindhoven University of Technology, Netherlands
Keywords: AI writing;professional writing;text analysis;writing in workplace settings
Domain: -; See Session Schedule

Abstract

Financial analysts are hired and paid to develop, explain and publish a point of view and a stance on matters in the financial markets. In doing so, financial analysts are in a double-bind situation: on the one hand, their forecast accuracy is factored into their financial compensation; on the other hand, reliable forecasts are never possible given the volatility and unpredictability of the financial markets (Arnold et al., 2025; Whitehouse, 2023). These circumstances encourage strategic recommendations that are written in such a way that they are always somehow true (Palmieri & Mazzali-Lurati, 2021). The double-bind situation of financial analysts is one of the main reasons why investment recommendations are difficult to understand by the addressees.With the emergence of AI, financial analysts are increasingly using AI tools to write their investment recommendations. This raises questions about the role of these emerging technologies in financial communication in general and, more specifically, how they affect the intelligibility of financial analysts' text products.In my presentation, I introduce the double-bind situation of financial analysts and its implications for financial communication (part 1). Based on interviews with financial analysts and a corpus of investment recommendations from Swiss banks (part 2), I use pragmatic text analysis (part 3) to examine how the use of AI writing tools in financial communication affects the strategic recommendations in financial analysts' text products (part 4). Finally, I discuss the implications of this development for the double-bind situation of financial analysts, for financial communication in general, and for society at large (part 5). Arnold, T., Roth, S., & Kleve, H. (2025). Double binds in dialogue: unraveling paradoxical communication in business families and family businesses. Management Review(36). https://doi.org/https://doi.org/10.31083/MRev39358Palmieri, R., & Mazzali-Lurati, S. (2021). Strategic communication with multiple audiences: polyphony, text stakeholders and argumentation. International Journal of Strategic Communication. https://doi.org/10.1080/1553118X.2021.1887873Whitehouse, M. (2023). Transdisciplinarity in Financial Communication. Palgrave McMillan. https://doi.org/10.1007/978-3-031-29115-9

Speech-to-Text for Students with Dyslexia - Implications from Studies in Sweden and Switzerland

Type: Single Paper
Presenting author: Bäck, Gunilla Almgren , Linnaeus University, Sweden
Presenting author: Keravec, Silvana Flütsch , University of Fribourg, Switzerland
Chair: Haake, Lisa , University of Münster, Germany
Keywords: speaking and writing;text composition;writing impairment;writing tools and writing technology
Domain: Learning and Special Education; See Session Schedule

Abstract

Research aimWriting is a key competence for academic and professional success. However, students with dyslexia face considerable barriers in text production, as their lower-order writing skills are insufficiently automated. This paper explores whether speech-to-text technology (STT) assists students with dyslexia in text production and whether there is a transfer to other modalities. Findings from complementary studies conducted in Sweden and Switzerland are synthesized to outline benefits and challenges for educational practice.Theoretical frameworkThe theoretical approach draws on Cognitive Load Theory (Sweller, 1994) and Bandura's (1997) concept of self-efficacy. STT may reduce cognitive load from lower-order writing processes, freeing resources for higher-order ones, and may strengthen self-efficacy compared to demanding writing tasks. Thus, STT may assist students with dyslexia in processes and products of text production. MethodsGunilla conducted a counterbalanced within-group study with typically developing middle school students and a multiple-baseline single-case study with students with dyslexia using STT. She also conducted a five-year follow-up interview study on experiences with assistive technologies used by students with dyslexia. Silvana conducted a quasi-experimental mixed-methods study with Grade 5 students with dyslexia. She investigated the effects of STT on text production and writing motivation and conducted interviews with teachers and specialists.FindingsThe present results confirm former mixed findings on the effectiveness of STT. While STT can be a helpful tool for students with dyslexia, co-morbidities may require additional adjustments. Monitoring progress and providing targeted scaffolding are essential and appreciated by students and professionals. The school environment also influences successful use. KeywordsSpeech-to-Text; Assistive Technology; dyslexia; text production1. ReferencesBandura, A. (1997). Self-efficacy: The Exercise of Control. W.H. Freeman/Times Books/Henry Holt & Co. Sweller, J. (1994). Cognitive Load Theory, Learning Difficulty, and Instructional Design. Learning and Instruction, 4, 295–312. PII: 0959-4752(94)90003-5

Who benefits from using speech-to-text as their writing tool?

Type: Single Paper
Presenting author: Wengelin, Åsa , University of Gothenburg, Sweden
Co-author: Kraft, Sanna , Linnaeus University, Sweden
Chair: Haake, Lisa , Institute of Educational Science, University of Münster, Germany
Keywords: spelling issues;text composition;writing impairment;writing tools and writing technology
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Writing presents significant challenges for many children, particularly those with reading and writing difficulties such as dyslexia. In addition to spelling problems, these children often produce texts of lower quality than their peers (Berninger et al., 2008; Connelly et al., 2006). These difficulties are commonly explained by cognitive bottlenecks during transcription, which place heavy demands on working memory and limit the resources available for higher-level writing processes (Berninger et al., 2002). One potential way to reduce transcription demands is the use of speech-to-text (STT) technology (Kraft, 2023; MacArthur & Cavalier, 2004; Quinlan, 2004). However, empirical knowledge of STT’s effects on children’s writing remains limited, particularly for languages other than English (Matre & Cameron, 2022), and it is still unclear for whom STT is most beneficial. This study examined the effects of built-in STT on writing among 57 children aged 10–12 and addressed two research questions: (a) which individual characteristics predict text quality in texts produced using STT, and (b) which children benefit most from using STT compared with typing. To address the first question, linear regression analyses examined whether working memory, reading skills, spelling skills, and expressive language skills predicted text quality in STT-produced texts. Although STT can reduce spelling demands, it may also introduce semantic inaccuracies due to misrecognition, placing additional demands on monitoring and revision. The results showed that neither working memory nor reading skills predicted text quality; only spelling and expressive language skills were significant predictors. To address the second question, participants were divided into three groups: children with both reading and spelling difficulties (n = 15), children with primarily spelling difficulties (n = 16), and children without reading and writing difficulties (n = 16). Texts produced using STT were compared with typed texts. Linear mixed models indicated that children with both decoding and spelling difficulties—but not those with only spelling difficulties—produced longer and higher-quality texts when using STT, even after minimal instruction. Overall, the findings suggest that STT, when combined with appropriate instructional support, can benefit some children with reading and writing difficulties, underscoring the need for further research investigating for whom it is most effective.

Writing with AI in Multilingual Classrooms: Translanguaging and Teacher–Student Perspectives

Type: Single Paper
Presenting author: Rakedzon, Tzipora , Technion - Israel Institute of Technology, Israel
Co-author: Jaber, Amani , Technion - Israel Institute of Technology, Israel
Chair: Conijn, Rianne , Eindhoven University of Technology, Netherlands
Keywords: AI writing;case study;pedagogy and instruction;social and cultural aspects of writing
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Writing with AI in Multilingual Classrooms: Translanguaging and Teacher–Student PerspectivesThe rapid integration of generative AI tools into classrooms is transforming how students search, learn, and write in the English as a foreign language (EFL) classroom, particularly in multilingual contexts where language choice shapes access and outcomes (Moorhouse et al., 2024; Yang & Lin, 2025). Yet little is known about how AI-mediated writing practices unfold in multilingual, multicultural school settings, or how such practices should inform writing pedagogy and assessment. This study investigates how Arab and Jewish Israeli secondary-school English teachers and their students use generative AI in English-language classroom writing tasks, and how multilingual language practices shape this use. We examine how learners draw on Hebrew, Arabic, and English when prompting AI, and how teachers and students perceive the usefulness and limitations of AI tools for writing. By analyzing language choice, perceptions, and writing in AI-mediated tasks, the study explores the intersection of translanguaging in EFL classrooms and critical digital literacy (Canagarajah, 2013; Pangrazio & Sefton-Green, 2021; Tzirides, 2024).Situated within a larger mixed-methods project in EFL classrooms in 6 Arab and Jewish high schools, the presentation reports on: (1) patterns of students’ translanguaging and multilingual prompting; (2) students’ AI-supported writing products, and (3) teachers’ and students’ perceptions of AI’s role and limitations in EFL learning and writing (Wang, 2024; Xiao, Yi, & Akhter, 2024). The research design includes the analysis of teacher and student surveys and semi-structured interviews; students’ AI-mediated writing tasks; students' reflection writing tasks on insights into AI-mediated writing; and the collection of prompts and writing artifacts. A central focus of the study is how generative AI reshapes learning and writing processes and influences students’ experiences, strategies, and language choice. The analysis also investigates teachers' perspectives and decisions regarding AI-mediated classroom use and identifies their professional development needs in integrating AI ethically and pedagogically. The study further explores how AI-supported writing tasks shift classroom norms of drafting, revision, and the use of multilingual resources, and offers recommendations for AI-integrated writing instruction and assessment.

AI and I: A rhizomatic analysis of writing processes with AI tools

Type: Poster Presentation
Presenting author: Silverdal, Sara , Umeå University, Sweden
Co-author: Hermansson, Carina , Stockholm University, Sweden
Chair: Grabowski, Joachim , Leibniz University, Germany
Keywords: creative writing;education;Generative AI;writing processes
Domain: Teaching and Teacher Education; See Session Schedule

Abstract

AI and I: A rhizomatic analysis of writing processes with AI toolsSara Silverdal, Umeå University and Carina Hermansson, Stockholm UniversityAs writing practices continually co-evolve with societal and technological change, the emergence of generative AI poses new challenges and opportunities for schools and students. This paper investigates how relationships between student writers and AI technologies are enacted during the writing process, and how these relationships reshape notions of authorship, agency, and textual production. Drawing on a socio-material framework and specifically employing a rhizomatic analytic approach (Mac Lure, 2013; Alvermann 2000), the study maps the assemblages that emerge when upper-secondary students in Sweden compose short stories with access to AI tools.The empirical material consists of 24 filmed writing sessions capturing students’ screens, facial expressions, and part of their intra-actions in the room. In addition, semi-structured interviews were conducted with six students - one group interview with four participants and two individual interviews. The paper focuses in depth on three exemplifying student cases to trace divergent pathways of becoming-with AI during writing.Initial findings reveal markedly different orientations toward AI: one student delegates much of the writing to the AI; another engages in iterative, reciprocal intra-actions with AI; and a third takes a critical stance refusing to use AI at all. Across these cases, authorship emerges as fluid and negotiable, shaped by the dynamic entanglements between students, tools, and texts. The analysis also highlights how critical thinking and reading of the AI generated products appears as a valuable asset to be able to work with these tools and interpret their responses.The study contributes to writing research by providing an empirically grounded account of how generative AI reshapes writing processes and writer identities. For pedagogy, the findings highlight the need to equip both teachers and students with critical, transparent, and equitable practices for working with AI tools. Such preparation is essential to ensure that AI becomes a resource for inclusive learning rather than a source of stratification in students’ writing development. Keywords:Writing process, creative writing, generative AI, upper secondary education

Assessing Digital Multimodal Composing in L2 Writing: A Scoping Review

Type: Poster Presentation
Presenting author: Yu, Nan , Beijing Normal University, China
Co-author: Lin, Dunlai , Beijing Normal University, China
Chair: Fürstenberg, Maurice , Ludwig-Maximilians-Universität (LMU), Germany
Keywords: assessment;evaluation and assessment of writing competence;multimodality;writing in a foreign or second language
Domain: Assessment and Evaluation; See Session Schedule

Abstract

AbstractThe continuous advancement of educational technologies has made digital multimodal composing (DMC) a burgeoning area of research in L2 writing. DMC refers to the design of a digital genre with the integration of multiple modes, such as text, image, sound, and gesture (Kessler, 2024). Instead of the traditional view of writing as monomodal written texts, DMC highlights the semiotic richness and technological affordance of contemporary writing practices. Despite growing pedagogical interest and positive evidence from L2 classrooms, appropriately assessing DMC products and composing processes remains a major challenge for writing teachers and researchers.While empirical and synthesis studies on DMC have proliferated within second language acquisition, the overall research landscape of DMC assessment remains underexplored. As a research synthesis approach, a scoping review can outline the status quo of an emergent topic and identify potential gaps for future research (Chong, 2025). Therefore, adopting the scoping review method and following the PRISMA guidelines, this paper selects and analyzes 30 research articles from 2005 to 2024 to map theoretical foundations, methodological approaches, and thematic trends in current DMC assessment research.Theoretically, current research mainly draws on three theories: systemic functional linguistics, multimodality theory, and multiliteracies theory. Methodologically, existing studies primarily employ the etic approach to explore key dimensions of DMC competence, as well as the data-driven approach to develop analytic rubrics for DMC products. Thematically, current scholarship focuses on construct definition and operationalization, teacher feedback literacy, and assessment tool development.Based on the identified limitations and gaps, corresponding directions for future research are put forward. This review contributes to a more comprehensive understanding of DMC assessment by synthesizing existing studies and offering practical implications for writing pedagogy and assessment.ReferencesChong, S. W. (2025). Synthesis Methods and Reporting Tool (SMART) for research syntheses in applied linguistics. Research Synthesis in Applied Linguistics, 1(1), 1-22.Kessler, M. (2024). Digital multimodal composing: Connecting theory, research and practice in second language acquisition. Multilingual Matters.

Emergent Literacy Development: A Socio-Constructivist Program in Preschool

Type: Poster Presentation
Presenting author: Salvador, Liliana , ISPA - University Institute, Portugal
Co-author: Martins, Margarida Alves , ISPA-Instituto Universitário EDUNOVA.ISPA, Portugal
Co-author: Maruta, Filipa , ISPA - University Institute, Portugal
Co-author: Laranjeira, Ricardo , ISPA - University Institute, Portugal
Chair: Whitehouse, Marlies , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: collaborative writing;experimental approaches;intervention study;learning to write
Domain: Learning and Social Interaction; See Session Schedule

Abstract

Emergent literacy refers to the foundational skills, knowledge, and behaviours that precede formal reading and writing instruction. It encompasses the natural development of literacy as children interact with their environment. These early literacy skills, as letter knowledge, phonological awareness invented spelling and early reading are crucial for successful reading and writing development, influencing long-term academic outcomes. The socio-constructive approach to literacy development considers that children build knowledge through meaningful interactions with peers and educators which role is to provide guidance, scaffolding and minimal intervention to support children’s discoveries. In this context our aim was to evaluate the effectiveness of a socio-constructivist emergent literacy programme in preschool designated to develop key literacy skills. Four classes from 2 schools in the Lisbon area attended by 88 5-year-olds participated in this study. The emergent literacy program was developed with 49 children attending two of these classes (experimental group). In the other 2 classes comprising 39 children, traditional literacy activities were developed (control group). In both classes the activities were developed by the educators during their classes. Children’s phonological awareness, letter knowledge, reading and spelling were assessed at the beginning and end of the school year. The emergent literacy program comprised 12 sessions, each beginning with contextualized activities (e.g., storytelling, singing a song, watching a short film) that provided a framework for subsequent learning. This was followed by activities addressing several emergent literacy skills (e.g., phonological awareness, letter knowledge, vocabulary, early interaction with print, invented spelling). All sessions began with a large group activity, followed by a small group activity, and finished with an individual activity. The control group activities consisted, mainly, of traditional tasks such as rhyming, singing songs, storytelling, and copying letters and words, in groups or individually, with low levels of interaction between the children. The study results demonstrated statistically significant differences between the groups, with the experimental group showing substantial improvements in letter knowledge, phonological awareness, spelling, and reading compared to the control group. These findings suggest that emergent literacy programs incorporating socio-constructivist and naturalistic practices can be highly effective in developing fundamental skills in preschool children.

From Ratings to Formative Feedback: An AI-Based System for Automated Essay Scoring

Type: Demonstration Session
Presenting author: König, Nina , Institut für Bildungsevaluation, Switzerland
Co-author: Berger, Stéphanie , Institut für Bildungsevaluation, Switzerland
Co-author: Stahlhut, Laura , Institut für Bildungsevaluation, Switzerland
Co-author: Debelak, Rudolf , Universität Zürich, Switzerland
Chair: Rybicki, Jan-Mikael , Aalto University, Finland
Keywords: evaluation and assessment of writing competence;feedback;Generative AI;natural language processing (NLP)
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Feedback is widely recognised as one of the most powerful influences on learning, particularly in the development of writing competence. However, in everyday classroom practice, the provision of detailed and timely feedback on student texts is constrained by limited time resources. Automated essay scoring (AES) has the potential to mitigate this tension, provided that it is pedagogically sound and sensitive to the complexity of writing.This demonstration introduces an AI-based AES system developed for primary and lower secondary education. The system generates structured feedback within seconds, addressing four core dimensions of writing: content quality, coherence and cohesion, language accuracy, and stylistic appropriateness. In addition to score-based ratings across eight criteria, the system provides qualitative, dimension-specific feedback designed to support formative learning processes.The development of the system builds on a large empirical foundation of 36,739 digitised student essays that were evaluated by trained human raters. By combining large language models with targeted natural language processing techniques and educational assessment frameworks, the system aims to produce automated feedback that is more consistent, transparent, and pedagogically grounded than that of general-purpose AI applications. The demonstration briefly outlines these design principles and explains the rationale underlying the selected feedback dimensions.The demonstration then focuses on how these principles are operationalised in practice. Participants are shown how the system structures multi-dimensional feedback, generates qualitative comments from textual features, and presents feedback in an interpretable manner for educational use. Particular attention is given to interface and feedback design choices that support formative use in the classroom and clearly differentiate the system from generic AI-based writing tools.Overall, the demonstration contributes to current discussions on AI in writing education by illustrating how automated feedback systems can be designed to augment instructional practice and support learning in classroom contexts.

How generative AI reshapes students' writing practices at a French university writing center

Type: Poster Presentation
Presenting author: Miftah, Hamza , Université Clermont Auvergne, France
Chair: Whitehouse, Marlies , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: academic writing;Generative AI;writing processes;writing tools and writing technology
Domain: Higher Education; See Session Schedule

Abstract

This paper examines how academic writing in higher education is transformed when writing is learned, regulated, and evaluated in interaction with generative artificial intelligence (GAI) tools. Focusing on master’s students’ learning of academic writing at a French university writing center, the study considers academic writing as an activity system (Engeström, 2014) and as a situated literacy practice (Lea & Street, 1998), in a context where students are typically confronted with the task of writing a master’s thesis without prior instruction in academic writing.From an activity theory standpoint, academic writing is seen as a goal-directed activity in which subjects, tools, and communities interact over time (Russell, 1997). From an academic literacies perspective, what is considered a valued text is embedded in broader relations of position and identity (Lillis & Tuck, 2016). This double lens allows us to examine not only what students do with GAI, but also how it positions them within communities of practice. The data combine a survey on rhetorical awareness, self-regulation, and GAI-related practices with semi-structured interviews conducted with master’s students attending the writing center. This work is drawn from an ongoing doctoral project on students’ learning of academic writing. Expected findings include differentiated profiles of learners according to how they mobilize GAI, genre knowledge and self-regulatory strategies to align their texts with perceived expectations. These profiles are expected to support the view that academic writing increasingly involves the use of GAI tools, not simply to offload writing tasks, but to mediate academic genres for novice writers whose disciplinary identities are still under construction, by making certain norms and expectations more explicit to them.This paper argues that studying writing through the lens of activity theory and academic literacies offers an understanding of GAI as a structuring component in the broader system of writing, as it reconfigures access to norms, resources, and legitimate participation in academic communities (Lave & Wenger, 1991). The needs revealed by students’ use of GAI tools offer research-informed directions for writing support that focuses on agency and rhetorical awareness in the use of tools, rather than on the technical regulation of GAI use.

Institutional policies on generative AI in BA thesis writing: Evidence from Romanian universities

Type: Poster Presentation
Presenting author: Baniceru, Ana Cristina , West University of Timisoara, Romania
Presenting author: Doroholschi, Claudia Ioana , West University of Timisoara, Romania, Romania
Co-author: Dinca, Andreea , West University of Timisoara, Romania
Co-author: Rogobete, Roxana , West University of Timisoara, Romania
Co-author: Chitez, Madalina , West University of Timisoara, Romania
Chair: Grabowski, Joachim , Leibniz University, Germany
Keywords: academic writing;education;Generative AI;pedagogy and instruction
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

LLMs have changed educational practices in universities across the world. This impact might be even greater in the case of bachelor theses, often written by less experienced students who might need more support with writing and might resort to LLMs to provide it. Universities have responded by creating policy frameworks that set the limits of permitted and disallowed uses of generative AI (e.g., Jin et al., 2025). Not all institutions, however, have been equally quick to respond to these challenges. In Romania, many universities have preferred to wait for models of action to become available from leading international institutions or official boards and have not yet articulated clear AI-related policies, which often leaves students and supervisors without clear guidance about how AI may be used in bachelor theses.In the present study, we analyse data from questionnaires and interviews, as well as publicly available policy documents from universities in the country to answer the following questions: Do Romanian higher education institutions have AI-related policies, and when were they implemented? Do these policies include specific provisions regarding Bachelor theses? Are students and supervisors aware of the existence of such policies, and do they integrate them in their work? Our findings show that, to date, not many Romanian universities have explicit policies regarding the use of AI. When they do exist, these often include only limited sections dedicated to the use of AI and few offer practical guidance on how to use AI in an ethical manner. By contrast, respondents to the interviews and questionnaires emphasize the need for institutional policies and for a consistent approach to the use of AI tools. Finally, we problematize this tension between the expressed needs of the academic communities and what the educational system currently provides, and make recommendations for the development of practical, discipline-sensitive guidance to support students’responsible use of AI in university contexts.Jin, Y., Yan, L., Echeverria, V., Gašević, D., & Martinez-Maldonado, R. (2025). Generative AI in higher education: A global perspective of institutional adoption policies and guidelines. Computers and Education: Artificial Intelligence, 8. https://doi.org/10.1016/j.caeai.2024.100348

Retrieval-Augmented Generation for Formative Thesis Writing Assessment

Type: Demonstration Session
Presenting author: Gautschi, Curtis , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Milićević, Dejana , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Rapp, Christian , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Ott, Jakob , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Heierli, Jasmin , ZHAW Zurich University of Applied Sciences, Switzerland
Chair: Rybicki, Jan-Mikael , Aalto University, Finland
Keywords: academic writing;AI writing;assessment;evaluation and assessment of writing competence
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

The development of academic writing skills in higher education presents numerous challenges. Instructors face high workloads alongside the ongoing need to provide focused and pedagogically sound feedback. While AI tools can assist in this area, current solutions show limitations. Automated writing assessment tools tend to focus on surface‑level language features (Zhao, 2025), and generative AI feedback may suffer from hallucinations, fail to address specific criteria, or lack alignment with teaching content (Gautschi, 2025). In addition, fine‑tuning large language models for specialized purposes—and many paid solutions—may be cost‑prohibitive.Recent developments in GenAI, particularly retrieval‑augmented generation (RAG) systems, offer a promising alternative (Li et al., 2025; Swacha & Gracel, 2025). Although RAG‑based architectures have been applied to academic writing support, to date they have not been applied to the specific problem of academic writing assessment for thesis writing. Existing tools such as CorpusChat (see Cheung & Crosthwaite, 2025) demonstrate RAG‑based support for student writers but do not include an assessment component. To our knowledge, no existing tool integrates RAG for assessment with the goal of providing feedback aligned with instructor or writing program specifications.To address these issues, we have developed a RAG‑based system for generating formative feedback. This approach allows for reduced hallucination, greater focus, and improved flexibility and control over generated feedback. Our SaaS‑independent multi‑chat, multi‑context RAG application (Node.js server, React frontend) incorporates user and persistence management, full handling of multiple RAG document contexts, recursive splitting, vector storage (Qdrant), and query rewriting to optimize similarity searches. Local context folders include target structures for feedback, target criteria, and samples of evaluated texts. The system provides feedback based on a modified IMRD‑based structure model for thesis writing. This demonstration session showcases the system’s potential to promote academic writing skills in higher education, benefiting both students and lecturers through a flexible, pedagogically grounded formative feedback ecosystem.

The challenge of creating a coherent text: A Coherence-Focused AI Chatbot for Academic Writing

Type: Demonstration Session
Presenting author: Rakedzon, Tzipora , Technion - Israel Institute of Technology, Israel
Co-author: Haddad-Najjar, Laila , Technion - Israel Institute of Technology, Israel
Chair: Rybicki, Jan-Mikael , Aalto University, Finland
Keywords: academic writing;AI writing;EFL and ESL writing;writing tools and writing technology
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

The challenge of creating a coherent text: A Coherence-Focused AI Chatbot for Academic Writing Despite growing interest in AI-supported writing instruction, relatively little is known about how interactive AI tools affect higher-order writing skills, such as textual coherence. To address this gap, we present the development of a coherence-focused feedback chatbot designed explicitly for research-based writing. The tool aims to translate writing research on cohesion and coherence into practical, real-time guidance in academic writing for students and early career academics.The tool is grounded in cohesion theory, which explains how linguistic features create unity and continuity across a text (Crossley et al., 2016; Halliday & Hasan, 2014; Morris et al., 2025). The coherence-focused chatbot was developed focusing on these cohesion theory frameworks, through iterative prompt engineering, and integrated into the free, online De-jargonizer platform (Rakedzon et al., 2017). The chatbot provides individualized, question-driven prompts that guide students in identifying aspects of coherence, such as lexical overlap across sentences/paragraphs, semantic similarity between ideas, and use of transitions. At each stage, they receive AI-generated suggestions and revise their texts across iterations. A reflection and a questionnaire on the process follow this. During the demonstration, attendees will receive an overview of the pedagogical rationale, tool development, and use in research design, followed by a demonstration of the chatbot. Participants will be invited to test the tool on their own writing samples and explore how coherence indices are employed to generate tailored feedback. The project demonstrates how research can be translated into classroom-ready practice, advancing understanding of AI’s potential to support higher-order writing skills in multilingual contexts. More broadly, it highlights how coherence-aware AI tools can contribute to more inclusive, adaptive, and discipline-responsive academic writing instruction.

Writing Fluency Always Matters, No Matter the Writing Technology

Type: Poster Presentation
Presenting author: Furer, Yves , Zurich University of Teacher Education, Switzerland
Co-author: Kolovou, Dimitra , University of Teacher Education of the Grisons, Switzerland
Co-author: Berger, Stéphanie , Institute for Educational Evaluation, Switzerland
Co-author: Philipp, Maik , Zurich University of Teacher Education, Switzerland
Chair: Fürstenberg, Maurice , Ludwig-Maximilians-Universität (LMU), Germany
Keywords: errors and error correction;fluency;handwriting;writing tools and writing technology
Domain: Assessment and Evaluation; See Session Schedule

Abstract

According to the Not-so-Simple View of Writing, transcription is a central component of writing (Ahmed et al., 2022). While the mechanical component of transcription (handwriting/typing) must be learned separately, the spelling component can be transferred from one writing technology to another. Additionally, computers offer additional support with spell checkers. However, there is a lack of studies that have examined the influence of different writing technologies in connection with spell checkers on secondary school students using a large sample size (Feng et al., 2019).The present study therefore investigates: Does the quality and fluency of students' texts differ when they write using different writing technology (handwriting / typing with and without spell checking)? Does the fluency of students' writing with different writing technologies explain differences in the quality of their texts?To answer these questions, 912 students (M = 14 years; 51% female) completed three writing tasks. The first writing task measured writing fluency. Text quality was measured with the second and third tasks (two different text types). Roughly one-third of the students wrote by hand (364), one-third wrote on a computer without spell check (301), and one-third wrote on a computer with spell check (277). Human raters and GPT-4o were used to determine text quality based on a rating scheme with four dimensions: content, coherence & consistency, language, and style.By running analyses of variance, groups differed significantly in writing fluency with less text produced by both computer groups, but not in their text quality (RQ 1). We employed regression analysis and found that writing fluency was a strong predictor of text quality irrespective of writing technology (RQ 2). Overall, our results emphasize the importance of writing fluency for writing practice in schools. Students need sufficient exercise with all writing technologies.References: Ahmed, Y., Kent, S., Cirino, P. T., & Keller-Margulis, M. (2022). The Not-So-Simple View of Writing in Struggling Readers/Writers. Reading & Writing Quarterly, 38(3), 272–296. Feng, L., Lindner, A., Ji, X. R., & Malatesha Joshi, R. (2019). The roles of handwriting and keyboarding in writing: a meta-analytic review. Reading and Writing, 32(1), 33–63.

Writing to learn in the new A(I)ge

Type: Poster Presentation
Presenting author: Staats, Julia , University Utrecht, Netherlands
Chair: Grabowski, Joachim , Leibniz University, Germany
Keywords: academic writing;education;Generative AI;writing to learn
Domain: Higher Education; See Session Schedule

Abstract

Writing-to-learn (WTL) can increase students’ understanding of disciplinary content (Armstrong et al., 2008; Bangert-Drowns et al., 2004). However, since generative artificial intelligence (genAI) was made freely available, we observe in our courses (Utrecht University, Bachelor Biology) that students use genAI during their thinking and writing processes for writing assignments. When genAI helps with or even takes over processes in the students’ writing process, the question arises whether WTL is still relevant as learning approach in future education. This study aimed to gain insight into how third-year university students perceive writing assignments and use genAI. Students from the final bachelor thesis course were invited to fill in an anonymous survey on genAI-use, self-efficacy for academic writing, writing beliefs, and how they experience academic writing in general. Results from close ended questions showed that students (n = 29) generally report that writing helps them understand content better, with no significant difference between students who do use genAI (n = 19) and students who do not use genAI (n = 10), t(26,637) = 1,75, p = 0,093. We also found no significant difference in how they experience academic writing (t(19,056) = -0,29, p = 0,774) and writing self-efficacy (H(1) = 0,544, p = 0,461) between these groups. Students who use genAI for writing assignments mostly use it as a brainstorm partner and to improve self-written texts and spelling. They least use it to generate texts, to compare literature, and to verify if their own text corresponds to the content of the source. Students give different reasons not to use genAI. Mainly low reliability and quality of AI-generated texts and it’s negative influence on learning were mentioned. This study forms a basis for a follow-up study across our whole student population to see if and how genAI-use poses a risk on the WTL-process throughout the bachelor.

Beyond Text-Focused Feedback: The Added Value of Keystroke Logging Feedback & Dialogic Peer Feedback

Type: Single Paper
Presenting author: Leijten, Mariëlle , Universiteit Antwerpen, Netherlands
Presenting author: Vandermeulen, Nina , University of Antwerp, Belgium
Keywords: feedback;Generative AI;keystroke logging;writing processes
Domain: Learning and Instructional Technology
Related Symposium: Writing process feedback; See Session Schedule

Abstract

Master’s students in Professional Communication & Management revise their texts several times before submitting a final version, guided by feedback. In addition to traditional, text-focused feedback, we introduced a combination of technologically supported process feedback (based on keystroke logging data) and a human-centred approach in which teachers supported students in reflecting on their writing processes. This process-oriented feedback was complemented by dialogic peer feedback, prompting students to engage in dialogue about their texts and underlying writing strategies.A total of 126 students wrote a bad-news email. Their writing processes were logged with Inputlog. After submitting a first draft, 57 students received an individual process report based on KSL data (Vandermeulen et al., 2020). Reflection was stimulated through comparisons with exemplar processes, some of which illustrated diverse ways of integrating GenAI tools into the writing process. A new KSL-based visualisation, the dynamic source network graph, was also piloted, mapping all consulted sources and their interconnections. Students subsequently clustered these sources into meaningful categories (e.g., GenAI tools, theory on bad-news emails, internet searches on content or formulation).All students then received text-focused feedback and revised their texts. Results showed that students exposed to both process- and text-focused feedback achieved significantly higher scores on their second drafts than those receiving text-focused feedback only.Subsequently, 53 students attended a session on requesting, giving, and processing feedback (De Kleijn, 2022; Tielemans et al., 2021), and were provided with tools to foster peer feedback dialogue (Bouwer et al., 2024; Landrieu et al., 2024). Analyses of third and final versions are underway to assess the added value of this dialogic peer exchange.Questionnaires and focus group discussions showed that students found the process reports clear and the exemplar comparisons insightful. Students emphasised, however, the need for teacher support in interpreting process data. Overall, 75% considered dialogic peer feedback useful, with more than half rating it more valuable than traditional peer feedback.Future research should further explore how combining KSL-based insights with teacher-guided reflection and dialogic peer feedback might foster students’ writing development and help them navigate GenAI tools more deliberately.

ChatGPT as a writing coach: A mixed-methods study in higher education

Type: Single Paper
Presenting author: Baaijen, Veerle , CLCG University of Groningen, Netherlands
Co-author: Hrib, Iulia , Faculty of Arts, University of Groningen, Netherlands
Chair: Galbraith, David , University of Southampton, United Kingdom
Keywords: academic writing;AI writing;education;feedback
Domain: Higher Education; See Session Schedule

Abstract

The role of ChatGPT in education has been a widely discussed topic, considering its ability to provide immediate feedback and personalised guidance to users (Lo, 2023). This mixed-methods study investigates ChatGPT’s role in enhancing text quality through feedback in higher education, focusing on its potential to support argumentative writing. The research comprises two within-participant design studies (N=16) and a qualitative analysis of student interactions with ChatGPT.Study 1 examined the impact of structured, task-level ChatGPT feedback on text revisions, with participants revising their drafts without direct interaction with the chatbot. Study 2 allowed free interaction with ChatGPT, supplemented by stimulated recall interviews to explore students’ perceptions of its utility. In both studies, text quality was assessed across organization, understanding, argument quality, and mechanics, while qualitative data, including chatbot interactions and revisions, were analyzed using Strobl et al.’s (2024) adapted framework and inductive coding.Results revealed significant improvements in text quality in both studies (Study 1: t(7)=-3.69, p

From Expert Habits to Student Support: Using Process-Tracing to Build GenAI Writing Guidance

Type: Single Paper
Presenting author: Speltz, Emily Dux , Embry-Riddle Aeronautical University, United States
Co-author: Chamberlain, Darryl , Embry-Riddle Aeronautical University, United States
Co-author: Aik, Maissane , Embry-Riddle Aeronautical University, United States
Co-author: Blood, Beata , Embry-Riddle Aeronautical University, United States
Co-author: Zaldivar, Zoe , Embry-Riddle Aeronautical University, United States
Keywords: AI writing;feedback;Generative AI;writing processes
Domain: Learning and Instructional Technology
Related Symposium: Writing process feedback; See Session Schedule

Abstract

This presentation introduces an exploratory study that investigates how expert and student writers integrate generative AI (GenAI) into their writing processes and how we might translate expert strategies into meaningful process-focused guidance for students. Motivated by the need for more situated support for GenAI-assisted writing, this research combines qualitative case studies with process-tracing technologies to uncover patterns in writers’ use of GenAI tools.The study proceeded in three phases. First, we observed eight self-identified “expert” users: professionals across industry and academia who use GenAI regularly in communication-centric work. These participants engaged in an authentic writing task while using GenAI tools. Through screen capture, keystroke logging (via Grammarly Authorship), think-aloud and stimulated recall protocols, and retrospective interviews, we documented how these experts strategically incorporated AI assistance into drafting, revising, and decision-making processes. Second, we conducted parallel sessions with ten novice student writers to capture how less-experienced users navigated similar GenAI-supported tasks. In both of these phases, we extracted observable patterns across sessions by inductively developing a codebook of actions throughout the writing process.In the third phase, we compared expert and student process behaviors to identify key differences in GenAI usage, such as when writers pause to reflect, reject, or revise AI-generated suggestions, or engage in iterative prompting. Using our codebook of process actions, we developed a set of process-focused GenAI writing strategies based on expert behaviors, which we then used to systematically develop feedback for students displaying certain patterns of actions. This phase of data collection is ongoing but will be completed prior to the conference; we will describe how students responded to the scaffolded feedback provided to them on the basis of their process behaviors. This presentation will highlight preliminary findings from both expert and student process behaviors, share insights on integrating consumer-facing tools like Grammarly Authorship into writing research, and discuss the process-focused feedback developed for GenAI-integrated writing. We argue that pairing process-tracing data with qualitative case study methods enables more nuanced, scalable observations of GenAI-integrated writing, which can advance both writing process research and pedagogical design for AI-assisted composition.

From Higher Education to Secondary Schools: Developing an OER for genAI-Supported Scientific Writing

Type: Single Paper
Presenting author: Schnatz, Nina Isabella , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Bieri, Aline , ZHAW Zurich University of Applied Sciences, Switzerland
Chair: Maurer, Michelle N. , University of Stavanger, Norway
Keywords: AI writing;design-based research;text genre;writing instruction
Domain: Higher Education; See Session Schedule

Abstract

Writing is widely recognised as an epistemic tool in higher education: it structures inquiry, supports knowledge creation, and enables students to participate in disciplinary discourse. These epistemic demands also shape Swiss secondary education, where learners in Berufs-/Maturitätsschulen must produce a propaedeutic research paper as part of their final examinations. The increasing presence of generative AI (genAI) in academic writing introduces challenges across educational levels. While genAI can support idea generation, structuring, and revision, research shows that students often struggle to integrate AI outputs into coherent, genre-appropriate, and epistemically responsible writing processes. This highlights the need for pedagogical designs that scaffold reflective and transparent genAI use throughout the writing process. This paper presents the development of an open educational resource (OER) designed to support genAI-assisted scientific writing in Swiss secondary schools. The OER is part of a broader design-based research (DBR) programme on genAI-integrated writing in higher education but is not itself an iterative DBR cycle. Instead, it represents a transfer of design principles and scaffolding mechanisms from two higher-education DBR iterations of a genAI-supported scientific writing course at the Zurich University of Applied Sciences (ZHAW). The resulting OER includes prompting activities, genre-focused self-study units, and reflective tasks adapted to the BM-/Matura-Arbeit context. It will be introduced to teachers in May 2026 to support implementation in the 2026/27 school year. The theoretical framework draws on writing-process models and genre approaches, conceptualising genAI as a tool to be critically evaluated within the epistemic aims of scientific writing. Methodologically, the OER design draws on analysis of course artifacts (prompting journals, student texts, writing tasks, scaffolds), student surveys from FS24 and FS25, and instructor feedback. Additional insights stem from workshops in 2025, which indicated strong demand for guidance on genAI use, authorship, and academic integrity. Expected outcomes include a modular OER that supports key writing stages while fostering genre knowledge, reflective practice, and epistemic responsibility. The paper contributes to writing research by showing how DBR-informed cross-level transfer can strengthen scientific writing pedagogy and support a smoother transition from secondary to tertiary education.Keywords: genAI-supported writing, scientific writing, writing pedagogy, epistemic practices

Writing process feedback

Type: Symposium
Chair: Vandermeulen, Nina , University of Antwerp, Belgium
Organiser: Vandermeulen, Nina , University of Antwerp, Belgium
Discussant: Speltz, Emily Dux , Embry-Riddle Aeronautical University, United States
Keywords: feedback;Generative AI;writing processes;writing tools and writing technology
Domain: Learning and Instructional Technology
Related Symposium: Writing process feedback; See Session Schedule

Abstract

This symposium continues the growing conversation on process-focused writing feedback, extending work presented at SIG Writing 2024 (Paris). Building on earlier work using process data and real-time analytics to inform pedagogy, the 2026 session turns toward the next frontier: advancing writing process feedback through AI-integrated and other technology-rich environments that foreground writers’ intentions and decision-making. Across three empirical projects, contributors examine how fine-grained writing-process data —from keystroke logs to GenAI interaction data— can be translated into actionable feedback for both researchers and educators.Together, the presentations explore how writers at different levels of expertise use and reflect on their writing processes: from expert and student integration of GenAI tools, to students’ alignment of intentions and actions during complex source-based writing, to the pedagogical value of process reports and exemplars (grounded in keystroke logging data) combined with dialogic peer feedback. We consider how process-focused feedback can foster aspects of learning such as self-awareness and reflection, regulation, and agency across learning contexts. By bringing these strands together, the symposium invites discussion on methodological innovation, data ethics, and pedagogical design in the next generation of process-focused writing research. It also aims to bridge insights from different methodologies (such as qualitative case studies, process-tracing technologies, and classroom interventions) to envision how process-focused feedback can most effectively be provided to student writers.The symposium on writing process feedback will consist of three paper presentations followed by the discussant’s response, with time for Q&A among presenters and an open, structured discussion with participants to identify future directions for process-focused feedback research.

Cooperative Writing: Perspectives from Three Intervention Studies

Type: Symposium
Chair: Philipp, Maik , Zurich University of Teacher Education, Switzerland
Organiser: Philipp, Maik , Zurich University of Teacher Education, Switzerland
Discussant: De Smedt, Fien , Ghent University, Belgium
Keywords: collaborative writing;education;intervention study;writing processes
Domain: Learning and Social Interaction
Related Symposium: Cooperative Writing: Perspectives from Three Intervention Studies; See Session Schedule

Abstract

Writing, as a cognitively demanding skill, can be improved through various intervention approaches (Graham, 2025). One of these is cooperative writing, in which peers carry out various cognitive processes together in social contexts. Cooperative writing can be conceptualized as an umbrella term describing a process in which peers work together and serve different roles in the three main processes of writing: planning, drafting, and revising (Alamargot & Chanquoy, 2001; Svenlin & Sørhaug, 2023). The contributions of the symposium focus on these three main processes from three current intervention studies in primary and secondary schools. They show how writing research contributes to the improvement and better understanding of school writing practices. Contribution 1 combines generative artificial intelligence with cooperative planning dialogues among 8th grade students. The students write arguments, with AI supporting content generation and the students being responsible for selection and organization. The dependent measures concern writing motivation.Contribution 2 focuses on the interactive negotiation processes involved in science learning within a writing-to-learn setting. It supports cooperative formulation of 5th grade students with scaffolds and shifts the focus of analysis and evaluation to both writing and learning aspects.Contribution 3 deals with the effectiveness of three different revision approaches that are compared against each other with secondary school students. The effects of the interventions are scrutinized with a new task that captures evaluation with special emphasis on higher order concerns. References Alamargot, D. & Chanquoy, L. (2001). Through the Models of Writing. Springer. Graham, S. (2025). What Do Meta‑Analyses Tell Us about the Teaching of Writing? In C. A. MacArthur, S. Graham & J. Fitzgerald (Eds.), Handbook of Writing Research (3. Aufl., S. 181–202). Guilford. Svenlin, M. & Sørhaug, J. O. (2023). Collaborative Writing in L1 School Contexts: A Scoping Review. Scandinavian Journal of Educational Research, 67(6), 980–996. https://doi.org/10.1080/00313831.2022.2115128

Does hybrid feedback foster L2 writing development?

Type: Poster Presentation
Presenting author: De Clercq, Orphee , Ghent University, Belgium
Presenting author: Özturan, Tuba , Erzincan Binali Yıldırım University, Türkiye
Chair: Weder, Mirjam , University of Basel, Switzerland
Keywords: EFL and ESL writing;feedback;Generative AI;natural language processing (NLP)
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Feedback is a pivotal component of both L1 and L2 students’ writing development (McCarthy et al., 2022), but providing in-depth feedback is a labour-intensive process (Godwin-Jones, 2022). Recent developments in generative artificial intelligence (GenAI) have increased interest in its use for providing personalized and real-time feedback in second language (L2) writing instruction. However, there is limited research on how GenAI-feedback combined with teacher mediation/control may support L2 writers’ development over time. Therefore, this study aims to investigate whether such hybrid feedback triggers the development of linguistic complexity in L2 writing.The study was conducted in a 15-week undergraduate Writing Skills course at a medium-sized university in Türkiye. Participants were 19 native Turkish students from the Department of English Translation and Interpretation with A2-level English proficiency. During the course, they completed eight timed, paragraph-level writing tasks across multiple genres, such as opinion, definition, process, and narrative, without technological support. After each task, students typed their drafts into shared Google Docs. They then received hybrid feedback: First, the course lecturer used GenAI (ChatGPT) to receive structured feedback focusing on the quality of the topic sentence, three common linguistic errors, three common global errors, and a fully revised version of the paragraph. Second, the course lecturer reviewed the GenAI-generated feedback and selected only accurate and appropriate responses, which were then shared with the students. Also, students wrote short reflection reports explaining how they engaged with the feedback and which suggestions they focused on. The dataset includes students’ original writing tasks, the hybrid feedback, and the reflection reports.The data analysis is still ongoing and focuses on analysing the linguistic complexity, considering both lexical and grammatical aspects (Bulté & Housen, 2012). To this purpose all text versions have been processed with the NLP tools for the Social Sciences (https://www.linguisticanalysistools.org/) and by selecting only those measures which are theoretically relevant (Bulté et al., 2025). By adopting a longitudinal perspective, this study aims to examine patterns of development rather than one-time improvements. Overall, this study contributes to discussions on the pedagogical efficiency of hybrid feedback in L2 writing instruction.

Enhancing Automated Essay Scoring by Integrating Rule-Based Language Checking with Generative Models

Type: Single Paper
Presenting author: Berger, Stéphanie , Institut für Bildungsevaluation, Switzerland
Presenting author: Stahlhut, Laura , Institut für Bildungsevaluation, Switzerland
Co-author: König, Nina , Institut für Bildungsevaluation, Switzerland
Co-author: Debelak, Rudolf , Institut für Erziehungswissenschaften, Universität Zürich, Switzerland
Chair: Perrin, Daniel , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: evaluation and assessment of writing competence;feedback;Generative AI;writing at school
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Recent advances in generative artificial intelligence (AI) have enabled automated feedback systems that offer scalable support for writing instruction in classroom settings. While large language models (LLMs) can generate formative feedback efficiently, prior research indicates that such feedback often contains hallucinations or lacks linguistic precision, thereby limiting its pedagogical usefulness (Jia et al., 2024; Cheng & Amiri, 2025). This study investigates whether integrating rule-based language-checking methods into a generative AI feedback system improves the accuracy and instructional value of automated feedback for student essays in primary and lower secondary education.To this end, we developed an AI-based feedback system that generates (1) ratings of spelling and grammar on separate four-point scales and (2) written feedback summarizing linguistic quality and listing detected errors with suggested corrections. Using this system, feedback was generated for 100 student essays under two conditions: generative AI augmented with rule-based methods and generative AI only.To evaluate the quality of both the ratings and the written feedback, linguistic experts independently scored the essays and reviewed the AI-generated feedback regarding hallucinations and inaccurate corrections. Preliminary results show that the correlation between human and AI spelling ratings increases from r = 0.608 to r = 0.713 when rule-based methods are integrated, while the correlation for grammar remained comparable (r = 0.607 vs. r = 0.576). To contextualize these findings, we present qualitative examples illustrating how the integration of rule-based checks corrected specific linguistic inaccuracies in the generative output. These findings suggest that hybrid systems can improve the accuracy of automated writing feedback, particularly for spelling.References Cheng, J., & Amiri, H. (2025). Linguistic blind spots of large language models. In NAACL 2025 Cognitive Modeling and Computational Linguistics Workshop. arXiv. https://doi.org/10.48550/arXiv.2503.19260 Jia, Q., Cui, J., Du, H., Rashid, P., Xi, R., Li, R., & Gehringer, E. (2024). LLM-generated feedback in real classes and beyond: Perspectives from students and instructors. In D. A. Joyner, B. Paaßen, & C. Demmans Epp (Eds.), Proceedings of the 17th International Conference on Educational Data Mining (pp. 862–867). International Educational Data Mining Society. https://doi.org/10.5281/zenodo.12729974

From Ratings to Formative Feedback: An AI-Based System for Automated Essay Scoring

Type: Poster Presentation
Presenting author: König, Nina , Institut für Bildungsevaluation, Switzerland
Co-author: Berger, Stéphanie , Institut für Bildungsevaluation, Switzerland
Co-author: Stahlhut, Laura , Institut für Bildungsevaluation, Switzerland
Co-author: Debelak, Rudolf , Universität Zürich, Switzerland
Chair: Weder, Mirjam , University of Basel, Switzerland
Keywords: evaluation and assessment of writing competence;feedback;Generative AI;natural language processing (NLP)
Domain: Assessment and Evaluation; See Session Schedule

Abstract

Feedback is widely recognised as one of the most powerful influences on learning, particularly in the development of writing competence. However, in everyday classroom practice, the provision of detailed and timely feedback on student texts is constrained by limited time resources. Automated essay scoring (AES) has the potential to mitigate this tension, provided that it is pedagogically sound and sensitive to the complexity of writing.This poster presents the design and underlying architecture of an AI-based AES system developed for primary and lower secondary education. The system generates structured feedback within seconds, addressing four core dimensions of writing: content quality, coherence and cohesion, language accuracy, and stylistic appropriateness. In addition to score-based ratings across eight criteria, the system provides qualitative, dimension-specific feedback designed to support formative learning processes.The development of the system builds on a large empirical foundation of 36,739 digitised student essays that were evaluated by trained human raters. By combining large language models with targeted natural language processing techniques and educational assessment frameworks, the system aims to produce automated feedback that is more consistent, transparent, and pedagogically grounded than that of general-purpose AI applications. The poster outlines these design principles and explains the rationale underlying the selected feedback dimensions.The poster then focuses on how these principles are operationalised in practice. It is shown how the system structures multi-dimensional feedback, generates qualitative comments from textual features, and presents feedback in an interpretable manner for educational use. Particular attention is given to interface and feedback design choices that support formative use in the classroom and clearly differentiate the system from generic AI-based writing tools.Overall, the poster contributes to current discussions on AI in writing education by illustrating how automated feedback systems can be designed to augment instructional practice and support learning in classroom contexts.

Literary Writing Process Modeling: across manuscript drafts and digital traces

Type: Single Paper
Presenting author: Ammari, Emna , Aix-Marseille Université, France
Co-author: Bellot, Patrice , Aix-Marseille University/CNRS, France
Co-author: Leblay, Christophe , University of Turku, School of Languages and Translation studies, Finland
Chair: Chanquoy, Lucile , Université Nice Sophia Antipolis, France
Keywords: AI writing;creative writing;keystroke logging data processing and analysis;writing models
Domain: Learning and Instructional Technology; See Session Schedule

Abstract

Investigating literary writing dynamics and authors’ revision signatures is increasingly recognized as a crucial field, drawing on both genetic criticism and psycholinguistics, as well as advanced generative AI systems. Despite this growing interest, a combined analysis of heritage manuscripts alongside contemporary keystroke logging data remains largely uncharted. Therefore, this proposal aims to bridge this gap by proposing a fine-grained modeling of literary writing and revision processes, developed within the Cré@LAME project (Literary Cre@tion and Author Manuscript Analysis), supporting an interactive assisted rewriting system, attuned to the author’s profile and revision strategies.The approach relies on a set of LLM-based agents specialized in context-aware rewriting, each performing a specific editorial role aligned with distinct revision intentions. These agents are coordinated by a multi-layer, multi-view Graph Neural Network (GNN) that models the evolution of textual states across heterogeneous materials, from linear manuscript transcriptions to digital writing traces.This network captures both linguistic (lexical, syntactic, semantic) and revision-oriented dimensions, reflecting editing operations and authorial intentions, across multiple levels, while guiding the agents’ rewriting operations according to learned patterns of textual evolution. This GNN thereby maintains coherence in editing operations while tracking author-specific revision practices.Accordingly, this work introduces a novel computational framework for textual genesis that addresses key aspects, including multi-granular data heterogeneity across manuscripts and digital log files, the inference of relevant indicators of authorial revision trajectories, and unified hierarchical representations formats of revision processes, integrating cross-source materials, suitable for multi-level graph modeling.Overall, this contribution advances research on textual genesis by highlighting how the integrated modeling of manuscript materials and digital traces provides deeper insights into authorial practices and the dynamics of literary creation.

Peer Feedback and Text Evaluation

Type: Single Paper
Presenting author: Sturm, Afra , University of Applied Sciences and Arts Northwestern Switzerland PH (FHNW), Switzerland
Co-author: Nyffeler, Angela , Fachhochschule Nordwestschweiz, Switzerland
Co-author: Morre, Gabriel , University of Applied Sciences and Arts Northwestern Switzerland PH (FHNW), Switzerland
Co-author: Hirt, Carmen Nadja , University of Zurich, Switzerland
Co-author: Karlen, Yves , University of Zurich, Switzerland
Keywords: cognitive processes;collaborative writing;intervention study;writing instruction
Domain: -
Related Symposium: Cooperative Writing: Perspectives from Three Intervention Studies; See Session Schedule

Abstract

The process of text revision is understood by cognitively oriented approaches as a sequence of activities that include reading through, evaluating, and revising the text (MacArthur, 2012). Effective peer feedback approaches address all three activities and support learners in different ways. The following questions can be used to guide these three activities: a) Reading through: How do I understand a text written by someone else? What is its overall idea? b) Evaluating: Can the author achieve the intended effect with the text? Does the text correspond to the respective genre? What should they change? c) Revising: How could the author implement these changes?In our intervention study regarding writing argumentative texts in grade 7 (N = 363), three peer feedback approaches are examined in comparison to a control group:· LUPA: an adaptation of CDO (De La Paz, Swanson & Graham, 1998).· SMABUSCH: explicit instruction of a revision strategy combined with teaching an argumentative text structure (Sturm, 2022).· REDIT: an editorial group discusses several texts and is observed by the audience (Amir, Atkin & Rijlaarsdam, 2021).While LUPA and SMABUSCH were implemented in pairs, REDIT was implemented in groups of up to eight students.Among other instruments, we used a task for evaluating a foreign text (analogous to López et al., 2021, but with authentic student texts), an argumentative writing task, and a reading comprehension test (Schneider, Schlagmüller & Ennemoser, 2017). Initial results of the evaluation task at t0 show that 33% of students failed to identify any higher order concerns (HOC), while another 29% identified only one out of six HOC passages. Students experienced even greater difficulty identifying underlying problems or proposing solutions.We will present first results on how evaluation skills develop across five measurement points, whether differences emerge depending on the peer feedback procedure, and the role of reading skills.

Reciprocal peer feedback with argumentative text structure

Type: Poster Presentation
Presenting author: Morre, Gabriel , University of Applied Sciences and Arts Northwestern Switzerland PH (FHNW), Switzerland
Chair: Fürstenberg, Maurice , Ludwig-Maximilians-Universität (LMU), Germany
Keywords: collaborative writing;evaluation and assessment of writing competence;intervention study;writing instruction
Domain: -; See Session Schedule

Abstract

Reciprocal peer feedback with argumentative text structureText revision is understood as a sub-competence that enables students to distance themselves from their own text, allowing them to identify inconsistencies and develop alternatives (Baurmann & Pohl, 2009). Cognitively oriented approaches consider revision as a sequence of activities that involve reading, evaluating, and revising the text (MacArthur, 2012).As part of an intervention study on revising argumentative texts in 7th grade, one of three experimental groups used the peer feedback approach Smabusch (N = 106 students). This approach combines the explicit teaching of a text-pattern-based revision strategy (Sturm, 2022) with reciprocal feedback (according to MacArthur, Schwartz & Graham, 1991). The acronym Smabusch focuses on an argumentative text structure (situation, opinion, argument, reasoning and examples to support it, and smash as the “winning argument”).Initial analyses indicate that Smabusch results in a positive change in strategy efficiency. This raises the question of how students in the experimental group use Smabusch to evaluate texts and how they proceed when implementing the strategy. The poster presentation will present further results also focusing on setting a writing goal and evaluating a text. Baurmann, Jürgen; Pohl, Thorsten (2009): Schreiben – Texte verfassen. [Writing – Composing Texts] In: Bremerich-Vos, Albert; Granzer, Dietlinde; Behrens, Ulrike und Köller, Olaf (Hrsg.): Bildungsstandards für die Grundschule. Deutsch konkret. [Educational standards for elementary school. German in concrete terms] Berlin: Cornelsen Verlag Scriptor. S. 75–103.MacArthur, Charles A.; Graham, Steve; Schwartz, Shirley (1991): Knowledge of Revision and Revising Behavior among Students with Learning Disabilities. In: Learning Disability Quarterly 14/1. S. 61–73.MacArthur, C. A. (2012). Evaluation and Revision. In V.W. Berninger (Ed.), Past, present, and future contributions of cognitive writing research to cognitiv psychology (pp. 461–483). Psychology Press.Sturm, A. (2022). Prozess- und produktorientierte Schreibförderung in Kombination [Process- and product-oriented writing instruction combined]. In V. Busse, N. Müller & L. Siekmann (Hrsg.), Schreiben fachübergreifend fördern. Theoretische Grundlagen und Praxisanregungen für Schule, Unterricht und Lehrerinnen- und Lehrerbildung [Promoting interdisciplinary writing. Theoretical foundations and practical recommendations for schools, instruction and teacher education] (S. 96–113). Klett Kallmeyer.

Scaffolding Multilingual Writers in Source-Based Argumentative Writing: An Intervention Study

Type: Poster Presentation
Presenting author: Debrah, Cecilia Owusu , University of Energy and Natural Resources, Ghana
Co-author: Yegblemenawo, Stella Afi Makafui , Kwame Nkrumah University of Science and Technology, Kumasi, Ghana
Chair: Keller, Stefan , Zurich University of Teacher Education, Switzerland
Keywords: academic writing;AI writing;intervention study;L2
Domain: Developmental Aspects of Instruction; See Session Schedule

Abstract

Source-based argumentative writing remains a demanding task, especially for multilingual writers in higher education, as they are expected to interpret diverse texts, synthesize multiple perspectives, and develop coherent arguments in a second language (Chuang & Yan, 2023). This study draws on a Vygotskian sociocultural perspective on mediated learning to explore how a scaffolded instructional intervention enhances students' engagement with sources in their argumentative writing, particularly given the growing influence of digitally mediated tools on students' academic literacy skills. Conducted over fourteen weeks, the qualitative study involved 60 undergraduate civil engineering students enrolled in the second part of a two-semester academic writing course. The intervention was based on five scaffolded phases: analyzing sources, summarizing, synthesizing, planning, and drafting, designed to make the process manageable and transparent. To reflect authentic writing practices in digitally mediated contexts, students recorded any AI tools they used during task completion. Data sources included 10 semi-structured interviews, classroom observations, instructional materials, and drafts of students’ writing assignments. Braun and Clark’s (2019) reflexive thematic analysis was used to examine how students navigated each stage, the challenges encountered, and the strategies employed to integrate sources into coherent written arguments. Findings show that scaffolded sequences helped students break down complex tasks, identify connections between texts, and build confidence in developing arguments. While AI-assisted tools provided localized support, the scaffolded activities remained the primary guide for deeper interpretive and rhetorical choices essential for effective academic writing. This research offers valuable insights into how structured scaffolding can aid L2 writers’ growth in source-based argumentative writing.

Secondary Students’ Decision-Making Processes Underlying L1 Writing Processes with GenAI

Type: Single Paper
Presenting author: Sliepenbeek, Leonie , Tilburg University, Netherlands
Co-author: van der Loo, Janneke , Tilburg University, Netherlands
Chair: Álvarez-Moreno, Carmen , Universidad de Leon, Spain
Keywords: Generative AI;L1;writing beliefs;writing processes
Domain: -; See Session Schedule

Abstract

Since the emergence of ChatGPT, generative artificial intelligence (GenAI) has been widely adopted by students in secondary and higher education for different tasks, such as writing. Yet empirical evidence how usage of GenAI affects writing processes has been scarce. In this qualitative pilot study we investigated how (Dutch) secondary school students’ L1 writing processes unfold when allowed to write with unguided support of GenAI when taking individual factors (self-efficacy and writing beliefs) into account.Three participants from grade 10 of pre-university secondary education were selected upon their scores on a Self-Efficacy for Writing Scale with statements regarding both writing with pen and paper and with support of GenAI. They were asked to write a synthesis text based on three sources, which meant they had to select relevant information, organize this and integrate these ideas into a new argumentative text. They were instructed to use GenAI as seen fit and their writing process was captured with both screen recording and keylogging software. To understand their decision-making process an additional questionnaire about their writing beliefs was filled out and semi-structured interviews were held afterwards.During our presentation we will demonstrate our findings about the interplay between individual factors and participants’ writing behaviour, as seen in the following example. One participant scored relatively high on both dimensions of self-efficacy, indicating they felt rather confident about their writing. Accordingly, this participant used GenAI only once (to ask for a definition) and wrote his text without returning to this output. The assessment of their own decision-making process during the interview showed that they explicitly refrained from using GenAI due to their beliefs about the value of learning to write for themselves. Early analyses of the other participants’ decision-making processes also suggest that the degree and type of GenAI usage may be closely linked to both self-efficacy and writing beliefs. We believe this study contributes to our understanding of how LLMs may be situated within theoretical models of writing and may provide a valuable starting point for effective writing interventions, as findings show which challenges and opportunities GenAI brings to writing classrooms.

Social Regulation in AI-Supported Feedback Ecologies: Disciplinary vs Non-Disciplinary Peers

Type: Single Paper
Presenting author: Khuder, Baraa , Chalmers University of Technology, Sweden
Chair: Perrin, Daniel , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: AI writing;feedback;Generative AI;writing processes
Domain: Learning and Social Interaction; See Session Schedule

Abstract

Research on feedback literacy and social regulation of learning increasingly acknowledges the improtance of multiple feedback sources; however, we still know relatively little about how regulation unfolds across different feedback ecologies, particularly in varied human–AI configurations. Drawing on models of self-, co-, and socially shared regulation of learning, this study examines how doctoral students regulate their writing when revising with (a) AI plus disciplinary peers and (b) AI plus non-disciplinary peers. Fifty-five PhD students were allocated to two conditions: one in which they received AI feedback and discussed their texts with disciplinary peers in groups of four, and another in which they received AI feedback and discussed their texts with non-disciplinary peers in groups of four. Data comprised (1) AI interaction histories, (2) 14 audio-recorded “listening room” discussions, and (3) ~300-word individual reflections comparing AI and peer feedback. Transcripts were segmented into episodes and coded for forms of regulation (self-, co-, and socially shared regulation) and functions of regulation (planning, monitoring, evaluating, adapting). Across ecologies, AI never participated in genuinely socially shared regulation; episodes of shared regulation emerged only in human–human negotiation. In AI + disciplinary peer groups, AI most often functioned as a co-regulator: students tended to follow AI suggestions when a disciplinary peer could “watch over”, with regulation distributed between AI guidance and expert peer oversight. In AI + non-disciplinary peer groups, AI was more often recruited as a resource for self-regulation: students critically evaluated and selectively adapted AI feedback in the absence of disciplinary authority. The study offers a nuanced account of how different actors in feedback ecologies shape regulatory processes, and the presentation will discuss pedagogical implications for designing feedback from multiple resources in doctoral writing courses.

Teenagers writing expository texts with and without gen-AI

Type: Single Paper
Presenting author: Johansson, Victoria , Kristianstad University, Sweden
Co-author: Wengelin, Åsa , University of Gothenburg, Sweden
Co-author: Lindgren, Eva , Umeå University, Sweden
Co-author: Sturk, Erika , Umeå University, Sweden
Co-author: Hermansson, Carina , Umeå University, Sweden
Co-author: Silverdal, Sara , Umeå University, Sweden
Chair: Chanquoy, Lucile , Université Nice Sophia Antipolis, France
Keywords: AI writing;writing models;writing process data collections;written language competence
Domain: Teaching and Teacher Education; See Session Schedule

Abstract

Writing with generative AI (gen-AI) introduces new affordances and constraints that invite reconsideration of long-standing writing models, such as Hayes and Flower’s (1980) framework and Kellogg’s (1996) model of working memory in writing. This study examines how key writing processes—planning, translating, and revising—and related subprocesses such as goal setting, audience adaptation, reading, and evaluation unfold when students write with gen-AI. Adopting a developmental perspective, it qualitatively compares writing with and without AI support in a cross-sectional design involving students (n = 52) aged 13, 15, and 18. In a classroom setting, the students produced comparable expository texts first without and then with a gen-AI tool of their choice, while their writing was captured through screen recordings. Post-task interviews probed their strategies and reflections on differences between the modes.These questions guide the study: (1) How do writing processes unfold with and without using gen-AI, and are there age-related differences? (2) How do writers interact with the gen-AI tool (e.g., prompting), and how do they make use of the generated text?Initial results show that all students produce coherent and linguistically appropriate expository texts without AI, consistent with earlier descriptions of developmental writing (e.g., Johansson, 2009; Wengelin et al., 2014). In contrast, age-related patterns emerge when AI is introduced. The 13-year-olds often use gen-AI to produce full texts based on task prompts and report valuing the tool’s ability to generate lengthy responses. The 15-year-olds tend to use gen-AI primarily for idea generation, rewriting the AI-generated material to align with their own voice. The 18-year-olds more often use gen-AI to refine their existing ideas and strengthen the logic of their texts.This developmental trend demonstrate that the youngest writers rely on gen-AI mainly to support translating processes, the middle group for planning processes, and the oldest group for revising processes. The findings are discussed in relation to how gen-AI may differentially support components of the writing process depending on writers’ developmental needs and strategic awareness, and how the use of gen-AI during writing can reshape existing writing models. Understanding these evolving practices is also essential for instructional approaches including assessment.

The Effects of ChatGPT Feedback on Student Engagement: A Longitudinal Study

Type: Single Paper
Presenting author: Katanneck, Stanislav , Université catholique de Louvain (UCL), Belgium
Chair: Perrin, Daniel , ZHAW Zurich University of Applied Sciences, Switzerland
Keywords: AI writing;feedback;Generative AI;L2
Domain: Motivational, Social and Affective Processes; See Session Schedule

Abstract

ChatGPT can provide timely, personalized and informative feedback to improve text quality and learning success. It can thus mitigate teachers’ workload, particularly in writing-intensive courses. Despite these advantages, it remains unclear to what extent L2 learners engage with and incorporate feedback in the revision process for the improvement of text quality, as feedback uptake depends on several external and internal factors (Liu & Storch 2010). Furthermore, recent studies emphasize that students’ engagement with written corrective feedback changes over time, and that these dynamics of students’ engagement with feedback have not been explored yet (Mao & Icy 2024: 815). Therefore, the present study analyzes the impact of GenAI-assisted feedback (exemplified by ChatGPT-4) in combination with teacher feedback in extensive university German as a foreign language courses (CEFR, B2/ C1). The study focuses on the following research questions: RQ1: To what extent can the combination of GenAI-assisted feedback and teacher feedback support the revision phase in the writing process?RQ2: Which dynamics can be identified in the learner profiles based on the engagement with ChatGPT-based feedback? This longitudinal study with international students of German as a foreign language adopts a sequential explanatory mixed-methods research design (QUAN ® qual) to answer the research questions. For the quantitative analysis (QUAN) learners’ engagement (including all subtypes: behavioral, emotional, cognitive and social) is measured by using a standardized questionnaire with closed items in 13-week courses. This data (n=74) is used to carry out a hierarchical cluster analysis with Ward-Linkage to identify latent learner profiles and to assess the dynamics of engagement over time. The qualitative component (qual) of the study comprises the analysis of open-ended questions in reflection sheets as well as interviews in focus groups to get a holistic view of the feedback uptake and students’ engagement. Preliminary findings indicate that ChatGPT feedback on syntactic complexity is effective in improving linguistic accuracy and syntactic range, while teacher feedback is beneficial for fostering self-reflection, strategic revision, and writing motivation. The results are transferable to other L2 contexts, in particular for general language courses and academic writing and thus offers a replicable framework for integrating GenAI feedback into writing pedagogy.

Thesis Writing with Generative AI: A Multi-Session Process Analysis

Type: Single Paper
Presenting author: Conijn, Rianne , Eindhoven University of Technology, Netherlands
Co-author: Van Waes, Luuk , University of Antwerp, Belgium
Chair: Álvarez-Moreno, Carmen , Universidad de Leon, Spain
Keywords: academic writing;Generative AI;keystroke logging;writing processes
Domain: Higher Education; See Session Schedule

Abstract

The use of Generative Artificial Intelligence (genAI) in education has had a substantial influence on the way students write. Given the rapid adoption genAI across higher education, it is important to ensure that its use does not compromise learning. However, to make informed pedagogical decisions on how to (or not to) use genAI in academic writing, teaching and assessment, we must first understand how students - and in the next stage also experts - interact with these tools.Previous studies have shown that genAI affects students’ writing processes in different ways. For example, some students use genAI more instrumentally, whereas others use it more reflectively, leading to distinct patterns in how their writing develops. However, prior studies have primarily relied on single-session writing processes. In the present paper, we extend this line of research by analyzing multi-session writing processes in the context of writing a master's thesis. Specifically, we followed the writing process of three master theses students in Cognitive Psychology and Social Sciences over a period of 20 weeks. The number of writing sessions varied substantially among the three students, with totals ranging from 42 to 78 and 110 sessions. Their writing processes were collected using keystroke logging and complemented with students’ interactions with genAI. Inspired by recent writing research, we analyze the keystroke and genAI-interaction data from three perspectives: (1) macro level: examining overarching process management and identifying the intensity of genAI use throughout the full thesis trajectory; (2) meso level: characterizing the individual writing sessions based on revision strategies, writing fluency, and interactions with external sources, including genAI; (3) micro level: identifying how specific genAI interactions influenced moment-to-moment revising and pausing behavior. Preliminary results show that the participants’ use of genAI differed considerably: one participant relied heavily on genAI in the early stages for searching and summarizing sources; another used it moderately in the middle stages to gain an understanding of theories, methodologies, and analytical approaches; and the third interacted with genAI primarily towards the end, using it as a conversational partner to discuss results. Further macro-, meso- and micro-level analyses are currently underway.

Typing Instruction: Teachers’ Professional Competence and Instructional Practices

Type: Single Paper
Presenting author: Lindauer, Nadja , School for Teacher Education, University of Applied Sciences and Arts Northwestern Switzerland, Switzerland
Presenting author: Schüler, Lisa , Universität Bielefeld, Germany
Chair: Álvarez-Moreno, Carmen , Universidad de Leon, Spain
Keywords: proficiency;writing across disciplines or across the curriculum;writing instruction;writing tools and writing technology
Domain: Teaching and Teacher Education; See Session Schedule

Abstract

Typing is a fundamental skill for producing written texts and participating in digital communication. For these reasons, many countries have included typing in their curricula, thereby assigning schools an important role in developing these skills (e.g., KMK, 2022). However, because the curricular integration often remains unspecific, typing is rarely taught systematically in schools (Pinet et al., 2025). In addition, there is a lack of basic training in teacher education. As a result, teachers feel inadequately prepared to teach typing (Donne, 2012). Research on the teachers’ professional competences in typing instruction is limited (Schüler & Lindauer, 2025). The project TasDi (Didaktik des Tastaturschreibens und der Textverarbeitung) addresses this research gap: In one sub-study, the teachers’ knowledge, beliefs, and teaching practices were examined in order to derive implications for teacher training and the development of teaching materials. Expert interviews were conducted with 23 teachers involved in typing instruction in the German-speaking countries (Germany, Switzerland, Austria), including, for example, German and computer science teachers. The interviews were semi-structured, audio-recorded, transcribed, and analyzed using content analysis (Kuckartz & Rädiker, 2024).The presentation provides insight into selected findings on teachers’ prerequisites and teaching routines. The interviews show, for example, that teachers enter the profession via significantly different training paths. With regard to teaching practices, it becomes clear that typing instruction is not uniformly integrated into specific subjects and that different approaches are used for guiding learners (e.g., collaborative vs. individual work). Further differences can be seen in the role of teachers when working with digital learning programs.Donne, V. (2012). Keyboard Instruction for Students with a Disability. The Clearing House: A Journal of Educational Strategies, Issues and Ideas, 85(5), 201–206.KMK. (2022). Bildungsstandards für das Fach Deutsch. Primarbereich i.d.F.v. 23.06.2022.Kuckartz, U., & Rädiker, S. (2024). Qualitative Inhaltsanalyse. Methoden, Praxis, Umsetzung mit Software und künstlicher Intelligenz. Beltz Juventa.Pinet, S., Zielinski, C., Alario, F.-X., & Longcamp, M. (2025). On the acquisition of typing skills without formal training by school-aged children. Reading and Writing. Schüler, L., & Lindauer, N. (2025). Die Rolle der Lehrperson im (digitalen) Tastaturschreibunterricht. In L. Schüler & N. Lindauer (Hrsg.), Didaktik des Tastaturschreibens (S. 147–182). Ruhr-Universität Bochum.

We, Myself and AI: On the Benefits of Combining AI and Cooperative Planning for Writing Motivation

Type: Single Paper
Presenting author: Philipp, Maik , Zurich University of Teacher Education, Switzerland
Co-author: Wolfangel, Mirjam , Pädagogische Hochschule Zürich, Switzerland
Co-author: Unger, Valentin , St.Gallen University of Teacher Education, Switzerland
Co-author: Glaser, Cornelia , Pädagogische Hochschule Heidelberg, Germany
Keywords: collaborative writing;Generative AI;writing beliefs;writing instruction
Domain: -
Related Symposium: Cooperative Writing: Perspectives from Three Intervention Studies; See Session Schedule

Abstract

Background: Generative artificial intelligence (genAI) is currently disrupting writing practices in schools and raises the question of how writing can be used meaningfully in the classroom. Against this background, we designed an intervention with adolescents that uses ChatGPT to generate arguments, which are then further developed during collaborative planning discussions. Many of the intervention features directly address motivational mechanisms from self-determination theory (Ryan & Deci, 2017) and social cognitive learning theory (Bandura, 1997) but require empirical testing and generally remain in need of further research in the field of writing.Methodological design: We expect to see increases in autonomous writing motivation (H1) and declines in controlled writing motivation (H2). We will measure these changes using validated scales (Smedt et al., 2022). We also hypothesize improvements in self-efficacy in planning arguments (H3; scale by Smedt et al., 2022) and in self-regulated argumentative writing (H4, scale by Wang et al., in press). We will test the hypotheses using repeated-measurement variance analysis in a pre-post design with three randomly assigned groups of 389 eighth-grade students stemming from 23 intact classes: a genAI group, an alternative treatment, and a pure control groupExpected results: At the time of submission, the intervention study is still ongoing. We will present preliminary results at the conference and provide a more detailed introduction to the intervention. From an instructional design perspective, the project with its focus on the motivation and use of genAI for planning represents important work in writing research for the further development of writing practice.ReferencesBandura, A. (1997). Self-Efficacy: The Exercise of Control. Freeman. Ryan, R. M. & Deci, E. L. (2017). Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness. Guilford. Smedt, F. de, Landrieu, Y., Wever, B. de & van Keer, H. (2022). Do Cognitive Processes and Motives for Argumentative Writing Converge in Writer Profiles? Journal of Educational Research, 115(4), 258–270. https://doi.org/10.1080/00220671.2022.2122020 Wang, J., Graham, S., Kim, Y.‑S. G. & Steiss, J. (in press). Zooming into Two Measurement Issues in Writing Self-Efficacy: Revision as a Distinct Dimension and the Generality Hypothesis in Argumentative Writing. Reading and Writing. https://doi.org/10.1007/s11145-025-10679-z

What Can Sentence-Centric Writing Models Reveal about the Writing Process?

Type: Single Paper
Presenting author: Ulasik, Margo , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Mahlow, Cerstin , ZHAW Zurich University of Applied Sciences, Switzerland
Co-author: Piotrowski, Michael , University of Lausanne, Switzerland
Chair: Chanquoy, Lucile , Université Nice Sophia Antipolis, France
Keywords: keystroke logging data processing and analysis;natural language processing (NLP);writing and linguistics;writing models
Domain: -; See Session Schedule

Abstract

Sentences are fundamental communicative units (Bühler 1918), and written texts are generally understood to consist of these units, but research on how writers produce sentences remains limited. Although linguistic modeling of the writing process has gathered interest in recent years, existing approaches, whether grounded in linguistic theory or in writing research, remain insufficient to explain how writers actually produce and revise text at a linguistic level. Prior work has investigated correspondences between writing bursts and linguistic structure (e.g., Kaufer et al. 1986; Cislaru and Olive 2018; Feltgen et al. 2023), examined revisions from a linguistic perspective (e.g., Manseri and Jouvenel 2025), proposed methods for transforming writing-process data into linguistic units (e.g., Leijten et al. 2019), and provided initial contributions to sentence-level analyses of the writing process (Miletic et al. 2022; Mahlow et al. 2024; Ulasik and Miletic 2024).We advance the state of the art by building on these developments and on the theoretical framework for sentence-centric modeling introduced by Ulasik et al. (2025). Our approach enables detailed tracking of sentence production through the analysis of sentence transformations and the detection of pauses at sentence boundaries. It supports systematic identification of bursts within sentence production and offers a method for characterizing the scope of transformations and bursts with respect to individual sentences.To investigate the potential of the model, we apply our software tool for sentence-centric modeling of writing, THEtool (https://github.com/mulasik/wta), to real-world data from the KLiCKe corpus (Yu Tian et al. 2025). This demonstrates the potential insights that emerge when shifting the analytical perspective from bursts or revisions to a sentence-centric view.

Digital Writing Technologies & AI

Presentations

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract