PaperlayerView exact revision on arXiv

Source-linked AI summary

Show your work: An exploratory study of student experiences of Turnitin Clarity

J. Roe, M. Perkins

arXiv:2609.34618v1cs.HC

TL;DR

Product-oriented assessment can make it difficult to judge whether students’ digital work represents their abilities, while process-capture platforms bring student acceptability questions that remain little studied. This exploratory study compared seven students’ expectations of Turnitin Clarity with their accounts after a writing task. Expectations were mixedly confirmed, and participants’ accounts show that process capture can shape the writing it records.

  • Problem

    Student experience of integrated process-capture platforms designed for post-GenAI assessment remains little studied, despite their proposed role in making digital assessment processes visible.

    Source: Introduction · p. 2, Process Capturing Technologies in the… · p. 4, Methodology · p. 4

  • Method

    Seven students took part in pre- and post-use focus groups around a five-day writing task on Turnitin Clarity in an exploratory qualitative, longitudinal study.

    Source: Methodology · p. 5

  • Results

    Students’ functional and affective expectations were largely confirmed: the anticipated sense of being watched persisted, tempering perceived fairness for some and heightening self-monitoring for others.

    Source: Discussion · p. 13

  • Takeaways & Limitations

    Participants weighed fairness and evidentiary benefits against surveillance and other costs, and process capture reorganised the writing it set out to record.

    Source: Conclusion · pp. 15–16

    The small, single-institution, unassessed study may not reflect graded use, and a linear record may not represent varied writing practices.

    Source: Results · p. 8, Limitations · p. 15

Abstract

from arXiv · show

Process-capture platforms promise to make the writing behind an assessed artefact visible, and vendors increasingly market them as a fairness measure rather than a detection tool. Their acceptability to the students who must write within them, however, is largely untested. This exploratory qualitative study examines how students' expectations of one such platform, Turnitin Clarity, compared with their experience of using it. Seven students at a UK university, given no onboarding to the platform, took part in pre-use focus groups. Six then completed an unassessed 500-word writing task over five days, before all seven returned for a post-use focus group. Data was analysed using reflexive thematic analysis, with Expectation-Confirmation Theory as an orienting frame. The overall pattern was one of mixed confirmation. Participants' functional expectations were largely met, and the sense of being watched that they had anticipated persisted after use: for some it was offset by perceived fairness, while for others it heightened self-monitoring. The bounded AI assistant divided them: its limits were welcomed for marking out acceptable use, but some felt it weakened their ownership of the work and worried it would flatten what they produced. Across both phases, participants weighed the costs of observation against perceived gains in fairness and asked for transparency to run in both directions. Several described already writing defensively in anticipation of accusations of AI misuse. Acceptance of process technologies was reported as being conditional on practice time, two-way transparency, and explicit data governance. Participants also questioned whether a single linear document can represent a writing process they described as messy and multimodal. Across these accounts, process capture did not sit outside the writing it recorded but reorganised the practice it set out to observe.

Introduction

Process-capture platforms are proposed as a response to the difficulty of judging whether assessed digital artefacts represent students’ abilities, but they may also intensify surveillance and performative writing. This study compares seven students’ expectations of Turnitin Clarity with their accounts after using it, finding mixed confirmation and conditional acceptance.

Introduction

  • Product-focused assessment makes the process behind digital artefacts difficult to see, complicating judgments about whether work represents students’ knowledge or ability.Responses to this challenge have included both reducing technology and using detection software, alongside platforms intended to represent the learning process.Source: Introduction · p. 2
  • Process capture may promise fairer assessment and integrity assurance, but surveillance risks and performative writing may undermine validity.The platform’s ability to assure academic integrity remains unproven.Source: Introduction · p. 2
  • The study explores student perspectives because students’ willingness, perceptions, and experiences are key conditions for the viability of platforms that capture their writing.Seven students’ expectations and post-task accounts were examined through focus groups oriented around adapted Expectation-Confirmation Theory.Source: Introduction · p. 2
  • The research asks how students’ expectations compare with post-use views and what those views reveal about acceptability in higher-education assessment.Source: Introduction · p. 2
  • The anticipated sense of being watched persisted after use, offset for some by perceived fairness and heightened self-monitoring for others.Participants’ acceptance was conditional on practice time, two-way transparency, and explicit data governance, and they questioned whether a linear record could represent non-linear writing.Source: Introduction · p. 2

Literature Review

The literature links digital process data to insights about writing, proposes process-focused responses to GenAI assessment challenges, and raises concerns about surveillance. Yet student experience of integrated writing platforms designed for post-GenAI assessment remains little studied.

Process Data as Research

  • Keystroke logs can support inquiry into writing cognition through pauses, bursts, and revisions, while the paper’s process-data scope is device actions during digital textual assessment.The paper notes that this focus does not cover process data in other educational forms, such as art portfolios or physical prototypes.Source: Process Data as Research · p. 3

Process Data as an Answer for GenAI in Assessment

  • Responses to GenAI assessment include detection, assessment redesign, process-focused models, secured assessment lanes, and written or oral examinations, with no simple solution.Human-centred authorship checks may be difficult to scale in large cohorts.Source: Process Data as an Answer for GenAI in… · p. 3

Process Data as Surveillance

  • Literature frames process capture as surveillance that may discipline student-writers, invite resistance, and damage relational trust between students and teachers.Source: Process Data as Surveillance · p. 4

Process Capturing Technologies in the Age of GenAI

  • Early studies examine process-tracking tools and classroom source tracking, but student lived experience of integrated writing platforms remains unexamined in this literature.Turnitin Clarity combines process-tracking functions with a bounded AI assistant, unlike tools focused on tracking or authorship alone.Source: Process Capturing Technologies in the… · p. 4

Methodology

This exploratory qualitative study compared students’ pre-use expectations of Turnitin Clarity with their post-use accounts through focus groups before and after a five-day writing task. It used Expectation-Confirmation Theory as an orienting frame and reflexive thematic analysis to develop themes across the accounts.

Methodology

  • The study addresses limited research on students’ experiences of process-capture products in post-GenAI assessment through a qualitative, longitudinal design.Participants joined a pre-task focus group, used Turnitin Clarity for a structured task, then joined a post-task focus group.Source: Methodology · pp. 4–5

Theoretical Framework

  • Expectation-Confirmation Theory oriented comparison of prior expectations with experience, while interview questions covered functional, affective, normative, and identity categories.The post-use questions were written as counterparts to the pre-use questions, and the guide structure is shown in Table 1.Source: Theoretical Framework · p. 5, Table 1 · p. 5

Recruitment

  • Seven volunteer students at a UK university participated without structured demographic collection or platform induction; six completed the writing task.One participant could not access the platform, so their post-use contribution reflects attempted access rather than task completion.Source: Recruitment · p. 6

Procedure

  • Participants had five days to write 500 words using platform features as they wished, after receiving only a website link and brief description.The task was unassessed, used participants’ own devices and schedules, and submissions were not retained or used for model training.Source: Procedure · pp. 6–7
  • The post-use focus group combined the two pre-use groups in one session, a format the authors note may have influenced group dynamics.Focus groups lasted about 60 minutes, were held virtually, and were recorded with consent for transcription; recordings were deleted after transcription and anonymisation.Source: Procedure · pp. 6–7

Data Analysis and Reflexivity

  • Reflexive thematic analysis iteratively developed themes across the full dataset, combining ECT-oriented comparison with inductive coding and researcher reflexivity.The researchers did not use codebooks or inter-rater reliability measures and sought disconfirming accounts in light of their prior expectations about surveillance.Source: Data Analysis and Reflexivity · pp. 7–8

Results

Across six themes, participants weighed observation against fairness, described effects on writing and AI assistance, and identified conditions for acceptance.

The Desire for Two-Way Transparency

Contested Fairness as Justification

Writing Defensively

  • Fear of false accusation already led students to retain evidence and alter style; platform surveillance heightened self-monitoring and could lower academic register.Source: Writing Defensively · pp. 10–11

Assistance and Refused Co-Authorship

Linear Process Capture and the Realities of Writing

  • Participants' handwritten, iterative, multi-document, and offline practices exceeded the linear record, raising concerns about system interpretation and access interruptions.Source: Linear Process Capture and the… · p. 12

Conditional Acceptance and Preconditions for Trust

  • No participant rejected the platform outright, but acceptance remained conditional on practice time, transparency about assessment use, and opt-in governance for process data.Source: Conditional Acceptance and… · pp. 12–13

Discussion

Students reported mixed confirmation: expected monitoring largely matched experience, while perceived fairness offset surveillance for some and intensified self-monitoring for others. Acceptance was tied to reciprocal transparency, practice time, data governance, and concerns about whether process records represent writing fairly.

Discussion

  • Participants’ functional expectations were largely met, but anticipated surveillance persisted; perceived fairness tempered it for some, while others reported anxiety and workflow, learning, or agency costs.Participants also wanted reciprocal information about what the platform showed.Source: Discussion · pp. 13–14
  • Participants already altered writing to avoid GenAI accusations, and platform monitoring further prompted some to simplify their register or self-check more.A single captured document also seemed insufficient for representing multimodal, non-linear writing practices.Source: Discussion · p. 14
  • Boundaries on the AI assistant clarified acceptable use for some, while others worried assistance could weaken authorship and learning or homogenise student work.Participants also raised concerns about reduced diversity in work produced with shared prompts and a common assistant.Source: Discussion · p. 14
  • Participants questioned whether one process document could represent messy, multimodal writing, including planning across digital and physical materials.One participant suggested a bounded, timed online exam as a suitable setting, citing control of AI use.Source: Discussion · pp. 14–15

Implications

  • Participants identified practice time, two-way disclosure of captured data and marker-facing views, and explicit governance—including opt-in secondary use and deletion timelines—as conditions for acceptance.They also called for training and space to become familiar with the platform.Source: Implications · p. 15
  • Process evidence may add marking workload; the cited proposal is to align it with learning outcomes and use concise, structured traces rather than an unbounded second product.Suggested approaches include annotated drafts or brief reflections, with selective review supported by rubrics.Source: Implications · p. 15

Limitations

  • The findings come from seven self-selected students at one institution completing an unassessed task, so experience may differ under graded conditions.One participant also could not access the platform fully; marker perspectives remain unexamined, and the study was funded by Turnitin.Source: Limitations · p. 15

Conclusion

The study reports that process capture does not simply record student writing: participants brought defensive habits to the platform, and visibility added pressure for some. The authors conclude that process capture can reorganise writing and call for testing in assessed settings with broader samples and marker perspectives.

Conclusion

  • Participants weighed fairness, evidentiary protection, and bounded assistance against surveillance, workflow disruption, learning loss, and uncertainty about how process data would be judged.The exploratory longitudinal focus groups offered an account of expectations and experiences of a process-capturing writing platform with a bounded AI assistant.Source: Conclusion · p. 15
  • Participants described defensive writing before platform use, and some reported further self-monitoring, including lowering their register to avoid being read as copying.The conclusion notes that direct experience made some participants less sceptical, but did not remove defensive writing.Source: Conclusion · p. 15
  • The authors conclude that process capture reorganises the writing it records, so its record should be interpreted in that light and tested with larger, more diverse samples and markers.Further research is proposed in assessed settings and from marker perspectives.Source: Conclusion · p. 16

Disclaimers

The disclosures describe the authors’ use of generative AI tools and report Turnitin funding and platform access, while stating that Turnitin had no role in the study.

Disclaimers

  • The authors used ChatGPT and Claude for ideation, some draft text, editing, formatting, and refinement, then reviewed and substantially revised outputs.They state that the tools supported rather than replaced core author responsibilities.Source: Declaration of Generative AI and… · p. 16
  • Turnitin funded the study and provided free platform access, but had no input into its design, data work, interpretation, writing, or submission decision.Source: Funding Statement · p. 16
  • The authors declare no conflicts of interest beyond the Turnitin funding and platform access already disclosed.Source: Conflict of Interest Statement · p. 16

Appendix 1: Writing Task Instructions

The appendix gives participants an unassessed writing task to complete individually in Turnitin Clarity. It asks for an approximately 500-word academic discussion of social media’s effects on everyday communication, supported by sources and examples.

Appendix 1: Writing Task Instructions

  • The task was unassessed, and writing quality would not affect participants’ studies, modules, or academic records.Its stated purpose was hands-on experience with the platform.Source: Appendix 1: Writing Task Instructions · p. 21
  • Participants were asked to write approximately 500 words examining social media’s impact on everyday communication, using academic sources and specific examples.The response was to present a clear argument or discussion in an academic style.Source: Appendix 1: Writing Task Instructions · p. 21
  • Participants could take any evidence-supported perspective and were asked to work individually in Turnitin Clarity.Source: Appendix 1: Writing Task Instructions · p. 21
  • AI use was optional, and participants could use platform features; the activity was for reflecting on writing experience, not assessing writing quality or AI-use patterns.The task could be completed between focus groups without grading criteria.Source: Appendix 1: Writing Task Instructions · p. 21
Loading 2609.34618v1…