Beyond passive rights-holders: toward student moral agency in assessment fairness

1 Introduction

In large-scale assessment, consequential validity has given the assessment field a decades-long literature on the moral responsibilities of test developers and institutions (Cronbach, 1988; Messick, 1998). Advances have been made in this area to promote student voice in examining validity and fairness of large-scale assessments (Cheng and DeLuca, 2011). Parallel advances in classroom assessment depict students as right-holders whose voices should be heard as agents navigating assessment environments (Adie et al., 2018; Willis et al., 2023). What rarely appears is a portrayal of students as moral agents: individuals capable of reasoning about what fairness actually requires and acting upon that reasoning in assessment practices, rather than solely reporting a perception of unfairness, or being rated on how their ethical normative judgements correspond to a teacher's or expert's (Fan et al., 2022; Rasooli et al., 2019).

This perspective article therefore advances a single claim: the advances made over the last decade in expanding student positioning in assessment have occurred largely without a corresponding recognition and development of student moral reasoning and conduct capacity. To address this gap, this perspective article will first review what advances have been made in classroom assessment research to reposition students’ roles in assessment. Then, theories of moral agency will be introduced and leveraged to trace how student fairness reasoning and conduct has been examined across five assessment domains: assessment literacy, self-assessment, peer and collaborative assessment, inclusive assessment, and AI-assisted assessment. This perspective article will close by arguing that moral agency in assessment is a developmental capacity requiring deliberate training, moving beyond giving students voice, rights, or an instrument to report their fairness perceptions.

2 The last decade: from object to partial subject

Fairness research in classroom assessment has gradually shifted from portraying students almost exclusively as objects of teachers’ fairness judgements toward being recognized as partial subjects within the process itself (Klenowski, 2009). This shift is motivated by the sociocultural theory of classroom assessment, which views learning as a process shaped by social, cultural, historical, and relational contexts, in which knowledge and identity develop together through participation in valued activities. In this theory, diversity, equity, belonging, and recognition of learners’ relational standing are central concerns (Elwood and Murphy, 2015; Shepard, 2019).

Several recurring principles inform the shifts in student repositioning. First, assessment is increasingly expected to recognize and build on the funds of knowledge and lived experiences students bring from their families and communities. This recognition supports students to participate in and even transform their communities (Fine and Furtak, 2020; Shepard et al., 2020). Second, students’ role in assessment has been advocated through a cluster of related constructs: student voice and rights (Sun et al., 2023), student agency (Adie et al., 2018), student assessment literacy (Hannigan et al., 2022), and student partnership in assessment (Smith et al., 2025). In particular, student agency in assessment has been further conceptualized to include ecological, authorial, sociomaterial, discursive, and digital framings (Nieminen et al., 2022; Nieminen et al., 2025). Ecological agency emerges from student actions situated in social and relational spaces. Authorial agency highlights students’ rights and responsibilities as authors of their own learning. Sociomaterial agency emerges from the interactions in the complex web of humans, objects, places, space, and time. Discursive agency relates to how agency is constituted via discourse. Digital agency concerns how the intersection of assessment in digital spaces augment student ownership of their learning. Missing from these five framings is an explicit discussion of student moral agency in assessment, which brings forth student fairness reasoning and conduct in assessment practices.

Third, recent research is increasingly interrogating diverse ways of knowing and how assessment structures represent them (Gipps and Stobart, 2009; Klenowski, 2014). Epistemic fairness offers a key conceptual anchor in assessment, concerning whose knowledge counts as legitimate and what forms of knowledge are represented or excluded (Fricker, 2007; Penuel and Watkins, 2019). Epistemic fairness comes closer to a moral-agentic move. While the development of this agency can be scaffolded by teachers, research is yet to show how such scaffolding moves students toward moral agency reasoning and conduct capacity for interrogating alternative knowledge structures represented and excluded in assessment (see Nieminen and Lahdenperä, 2024 for an example in mathematics assessment).

3 The missing dimension: moral agency

Questions of morality can be organized around three dimensions: (a) metaphysical questions about what moral good is, (b) epistemological questions about how we know it, and (c) axiological questions about why we should act on it. These three dimensions converge on a fourth: who the moral agent is (Boersema, 2014). Moral agency concerns personhood, responsibility, and the capacity to reason and act as a moral being. Extended to assessment, it is a capacity for students to develop fairness reasoning and conduct in relation to various assessment domains.

The notion of moral agency is closely related to a conceptualization of ’self’. A self is a core of being that attempts to provide a unified and persistent personhood. This self orients moral reasoning and guides respective actions. There are two key conceptions of self: ‘humanist’ and ‘poststructuralist’ (Pesch, 2020). The humanist self presents a unified identity (a coherent autobiographical narrative of self) that makes (moral) choices and decisions. The humanist self is the dominant perspective in ethics (Abend, 2019), providing a normative schemata based on which individuals make moral decisions, evaluate actions of themselves and others, and hold moral doers accountable for the actions. This view is compatible with a conceptualization of moral agency from a socio-cognitive lens that is exercised through three self-regulatory subfunctions: (a) monitoring one's own conduct, (b) judging its culpability against personal standards and social circumstances, and (c) generating affective self-reactions (Bandura et al., 1996). Extended from this individual humanist self is a collective self that emerges from a shared narrative of the group biography. This collective identity enables drawing boundaries between ‘us’ vs. ‘them’, legitimizing group-based moral reasoning and actions.

In a poststructuralist conception, self is produced as the outcome of discursive practices, including categories and systems of language, which make certain ways of thinking about oneself possible. The poststructuralist conception further emphasizes boundaries that are created and maintained by the discursive categories in networks of relations that help to unify a multitude of impressions into a coherent order. Such boundaries support a unified understanding of what ‘I’ or ‘we’ is or is not that also guides moral reasoning and conduct.

Although humanist and poststructuralist models of self stem from different ontologies, they can be complementary (Pesch, 2020). Humans deal with the plurality and heterogeneity of discursive categories, and they engage in boundary work to unify these multiple categories in a network of relations that guides a coherent and singular action. The poststructuralist model captures these boundaries of categories in a network of relations that further informs the humanist model, which captures a unified, coherent, and singular action, for which humans are held individually and collectively accountable. Therefore, poststructuralist and humanist models of self work together to guide how human moral agency functions (Pesch, 2020). Either perspective supports an argument that moral agency is a skill that can be trained and developed through practice, judging between courses of action amid plurality of competing normative orders. This manuscript leverages both perspectives, where appropriate, to elucidate how students’ perceptions of fairness are captured across the five noted assessment domains and how student moral agency in assessment can be supported using both perspectives.

3.1 Moral agency in assessment literacy: modeled by teachers, not examined in students

Student moral agency can be developed by teachers’ modeling in assessment as well as creating spaces for students to learn, reason, negotiate, and behave based on their growing fairness capacity. Existing assessment literacy frameworks have recognized ethics as a foundation for teachers’ assessment practices (DeLuca et al., 2023; Pastore and Andrade, 2019; Xu and Brown, 2016). These frameworks, coupled with empirical work, have examined what teachers consider ethical and fair across various assessment domains (Green et al., 2007; Liu et al., 2016; Rasooli et al., 2023a; Tierney, 2014). Adopting critical assessment literacy, Tajeddin et al. (2022) have taken an additional step to investigate what teachers believe that assessment should do and be like. Using a moral agency lens, Campbell (2004) has explored how teachers reason and model ethical practices in their pedagogy and assessment as a basis to develop students’ moral agency. That said, data were not collected from students to examine the impact of such modeling. Collectively, these studies report on what teachers consider to be fair and in occasional cases the underpinning factors that drive or should drive teachers’ fairness practices. Recent work is gradually moving from identifying these factors toward fairness education in assessment that can promote teachers’ fairness reasoning and conduct. For example, Rasooli and Holden (2026) used the Classroom Assessment Fairness Inventory (CAFI) to promote fairness reasoning in preservice teachers in Ontario. The findings showed preservice teachers wrestling with various notions such as differentiation and equity, standardization and equality in treatment, consequences related to assessment impact on students, and the value of the CAFI and peer discussion in learning about assessment fairness (see also Rasooli, 2023 for an ITEMs module).

No parallel program of research yet exists to develop this fairness reasoning and conduct capacity in students. Existing assessment research has explored students’ perceptions of fairness and ethics in assessment. This research has used an ethical descriptive lens such as social psychology theory to describe what students perceive to be fair in assessment contexts (Murillo and Hidalgo, 2017; Rasooli et al., 2019). Further, researchers have used an ethical normative lens to judge students’ fairness ratings against experts’ or teachers’ (Fan et al., 2022; Gao et al., 2022). For example, the CAFI was used with students to elicit their perceptions of fairness across five dimensions spanning group work, cheating, exams, grading, and feedback (Rasooli et al., 2023b). The CAFI and related instruments offer research and pedagogical tools to support interrogating what students consider to be fair across classroom assessment domains. What empirical research has yet to do is engage students in reasoning about what underpins their fairness judgements and what mechanisms can support developing student moral agency in assessment.

3.2 Self-assessment: an unrealized route to moral agency

Self-assessment is typically modelled as planning, monitoring and evaluating one's work using external feedback (teacher, rubric) and internal reflection, and revising through iterative cycles (Panadero et al., 2016; Rickey et al., 2023; Yan and Brown, 2017). Agency in self-assessment research has been argued to be a developmental capacity dependent on student prior knowledge, exposure to standards, and structured scaffolding (Panadero and Brown, 2026). This body of research has focused largely on how student agency develops and broadens in relation to learning outcomes in subjects such as language, mathematics, and science. Specifically, research has shown that students may be biased in judging their own performance; higher-performing students may underestimate, while under-performing students may overestimate their performance (Brown et al., 2015). Some students were reported to be aware of such biases and attributed it to insecurity in sharing honest low self-ratings (Yan et al., 2023). By continued engagement and scaffolded guidance, self-assessment process can enhance student moral agency in their awareness of such implicit biases and guide students in taking further responsibility in honest sharing of their criteria mastery.

Apart from general academic learning, a missed opportunity in the self-assessment literature concerns how this assessment process supports student moral agency in relation to character and citizenship education; self-assessment could be used to support students’ moral agency by having them write reflections on their own civic and ethical learning and then evaluate the breadth and depth of their engagement across a set of dimensions. For example, de Groot and Lo (2021) propose a model of democratic citizenship including eight dimensions that can be supported by student self-assessment. These dimensions encompass (a) critical community participation, (b) critical community education, (c) critical personal participation, (d) critical personal education, (e) basic community participation, (f) basic community education, (g) basic personal participation, and (h) basic personal education. These dimensions span democratic education and participation at both individual and collective levels addressing diverse issues such as systemic injustice, historical social movements, navigation of value conflicts, voting, volunteering, and social-emotional competencies at basic and critical levels. Students and teachers can use this framework via self-assessment to develop moral agency in fairness reasoning and conduct for character and citizenship education.

3.3 Peer and collaborative assessment: reporting without reasoning

Collaborative assessment is typically examined across three dimensions: contributions, workload, and interactional patterns. These dimensions support both group interdependence and individual accountability (Meijer et al., 2020; Strijbos, 2016). Empirical findings map closely onto these dimensions; students report unfairness when they cannot choose peers or are grouped with unmotivated members, when workload is uneven, or when free-riding goes unaddressed, while grade stakes push higher performers to take over tasks to protect their own outcomes (Forsell et al., 2020; Forsell et al., 2025; Rasooli et al., 2025b). Positive perceptions cluster around equal workload distribution, and the demand for an individual grade (Rasooli et al., 2025b). At the heart of students’ perceptions of unfairness are two central concerns: equal distribution of workload and the allocation of group grades. Students may view group work as unfair when they perceive that a member has contributed less than others, such as in cases involving free-riding or concerns about selecting group members, or when a higher-performing student assumes responsibility for most of the work. Similarly, awarding the same grade to all group members may be perceived as unfair when students believe that individual contributions are not adequately recognized. This concern about grading is closely linked to expectations of equal participation. When students perceive that these expectations have been violated for reasons they consider illegitimate, they may place greater emphasis on receiving grades that reflect their individual contributions (Rasooli et al., 2018).

These perceptions of fairness can be examined from a poststructuralist moral perspective (Pesch, 2020). Categories such as equality, individualism, and meritocracy form a network of relations underpinning assessment and grading structures, which reinforces a particular conception of fairness as equality in groupwork assessment. Proposed fixes to make group work fairer, such as individual marks and peer grading to monitor internal group contributions reproduce this logic rather than question it. To rethink fairness in groupwork assessment, the network of relations shaped by equality-individualism-meritocracy categories needs alternatives. An alternative network of relations can include categories such as participation, representation, and relational recognition, supported by students’ own reflection on how they worked together.

Peer assessment is also positioned as the mechanism that resolves fairness tensions by tracking individual contribution; this positioning, however, makes it a tool operating inside the same individualist network, not one that questions it. Empirical peer assessment research has shown that bias enters via intrapersonal and interpersonal variables — friendship, trust, psychological safety (Panadero et al., 2023): peers over-score friends (Panadero et al., 2016), and acquainted peers grade less strictly than unfamiliar ones (Domínguez et al., 2016), though findings are not fully consistent (Azarnoosh, 2013). Reading these findings through a humanist moral perspective (Pesch, 2020) suggests that collective group identity motivates friendship bias. Students may experience psychological risks when exercising moral agency against peer norms; students may not report peers due to loyalty and preservation of friendship as well as fear of consequences related to peer retaliation. Creating structures such as individual reflection and peer feedback against explicit standards would support students in developing fair representation, recognition, and participation of peer members. Explicit conversations with students about the psychological risks they may experience during peer assessment (e.g., friendship loyalty and related consequences) can raise their awareness of these risks. Further, students can be invited to reflect on how honest reporting supports true friendship and protects important values, such as taking responsibility for honesty, that promotes peer learning for individual and social development. Teachers can explicitly recognize and celebrate acts of student honesty in peer assessment reporting that signals directions for positive moral agency development. Such cognitive reframing (e.g., the meaning of true friendship, taking responsibility for honesty, recognizing these values in classrooms) can support a culture of psychological safety and moral agency (fairness reasoning and conduct) in supporting peers’ learning.

Across both literatures, students have mostly been positioned as reporters of fairness perceptions. For example, Rasooli et al. (2025a) provide a 20-item Peer Assessment Fairness Instrument (PAFI) that gives an account of dimensions based on which students perceive fairness in peer assessment including grading outcomes, procedural dynamics, and interactions. Such efforts yet remain, by design, a report of perceptions rather than an exercise in developing student moral agency in collaborative and peer assessment. Moral agency requires something more explicit: discussing with students why they consider equal workload or individual grading fair, how assessment itself reinforces those conceptions, whether peer assessment really functions as the panacea for tracking individual contribution, and what alternative network of categories such as representation, recognition, and participation might offer avenues for rethinking fairness in collaborative and peer assessment.

3.4 Inclusive assessment: a model worth building on

Inclusive assessment for students with disabilities has largely been framed around accommodation. Nieminen et al.’s (2024) synthesis showed that across a large number of studies, students with disabilities experienced assessment itself as a site of exclusion. Similar studies have demonstrated that some students reported reluctance to use the accommodations they were entitled to, citing stigma, embarrassment, a sense of being different from peers, being seen as unable to meet expectations without assistance, and a perceived unfair advantage over classmates (Rasooli et al., 2021; Tai et al., 2023).

These findings echo the same network of category relations (i.e., individualism-equality-meritocracy) operating in collaborative and peer assessment, reinforcing an exclusionary experience of assessment for students with disabilities. Nieminen (2025) argues that the accommodation model itself legitimizes a binary divide between disabled and non-disabled students; normalcy becomes a category assessment actively stabilizes, which is precisely why some students avoid accommodations for fear of being marked abnormal. Thus, such categories discourage some students from using accommodation to which they are legally and rightfully entitled. A parallel dynamic runs through individualism: assessment makes students responsible for their own inclusion, requiring disclosure and documentation, which reinforces disability as an individual property to be proven (Nieminen, 2025). Nieminen (2025) goes further to argue that assessment's core functions, such as equal standardization and individualism, sit in genuine paradox with inclusion's commitments to diversity and interdependence, a tension to be lived with, mitigated, or knowingly embraced rather than fully solved.

Yet this body of work, like the perception research reviewed throughout this perspective article, goes only as far as reporting what students with disabilities perceive about fairness in accommodation; it does not position them as moral agents examining what they consider fair, and why and how their fairness reasoning and conduct can be further developed. One recent study points more clearly toward what a moral-agentic research agenda in inclusive assessment could look like. Nieminen (2026) asked 139 students with disabilities to generate their own suggestions for more inclusive assessment. Students proposed design changes such as multiple assessment formats, choice over how, when, and why they are assessed, flexible procedures, and removing closed-book examinations. Further, students suggested procedural changes to the adjustment system itself, including simpler applications, clearer information, and normalizing adjustments as routine rather than exceptional.

Additional research could examine critically how fairness principles like equality are propagated and legitimized in society, applied within accommodation contexts, and internalized by students with disabilities. Such an examination requires working with both students with and without disabilities to build familiarity with the medical and social models of disability and how each model shapes how they think about assessment. The medical model locates disability within the individual as a deficit to be remediated, against the social model, which relocates the problem to attitudinal, structural, and environmental barriers (Graham and Willis, 2026). Such research will support students with and without disabilities as moral agents in evaluating their existing fairness conceptions as well as how their fairness reasoning can be developed as a basis to generate proposals for fairer and accessible assessment practices grounded in their own experiences.

3.5 AI-assisted assessment: compliance without reasoning

Fairness in AI-assisted assessment can be understood in its network of relations with interdependent categories. In their systematic review, Ali and Maroulis (2026) identified these categories that function together in network of relations to ensure that a work is attributable to an individual student and accurate representation of student ability. These categories include validity, integrity, originality, academic misconduct, cheating, unauthorized assistance, and misrepresentation of authorship. The network of relations based on these categories shapes and maintains a discursive boundary of what fairness means in AI-assisted assessment. Three dominant discourses were found within this literature. Risk-oriented discourses express moral panic about the erosion of students’ critical thinking and problem-solving capacities, and about undermining the value of student work and university degrees. Rule-based, compliance-driven discourses safeguard adherence to institutional policy of which AI declaration forms are the clearest example. Assessment design-focused discourses highlight assessment design that should generate evidence of individual student's learning. Across these discourses, students are positioned as compliant subjects rather than active participants. Fairness across these discursive categories is underpinned by the nexus of individualism-equality-meritocracy that was also in operation within collaborative and inclusive assessments. Individualism highlights individual student ownership of the work; equality emphasizes standardized conditions for AI-assisted assessment design and administration; meritocracy supports ranking and rewarding students’ AI-assisted work based on its representation of student ability levels. This nexus prevents higher education from losing its social mandate amid fears of large-scale academic misconduct, warranting university guidelines and instructions for AI-assisted assessment to be built around permitted use and compliance (Nieminen et al., 2026). What is missing in these AI-assisted assessment discourses and university guidelines is student positioning as moral agents, whose fairness reasoning and conduct need to be developed for myriad cases concerning their learning in AI-related contexts. This extends beyond questions of legitimate assistance to whether students should disengage from a given AI tool in instances where the tool itself operates unethically or when its use raises privacy or intellectual-property concerns (Bearman et al., 2024).

Developing student moral agency is timely because AI-assisted assessment has been described as a wicked problem, in which security, learning, fairness, authenticity, and workload cannot be optimized through any single, stable solution (Corbin et al., 2026). Student moral agency therefore needs to examine the plurality of competing normative orders, working with various categories of fairness that can be applied within a context (Pesch, 2020). Further, developing student moral agency echoes postplagiarism thinking for AI-assisted work. Eaton (2025) argues for a shift from a punitive to an educative mindset in academic integrity; from a project of catching cheaters to one of developing ethical citizens who understand the value of honest work and learning. Ying’s (2026) scoping review of student voice in AI-related academic integrity points to what such an educative approach could concretely involve: teaching students how generative AI actually works and where its limitations and biases lie, and engaging them with the underlying philosophical questions of authorship, originality, and intellectual labor that AI use raises, rather than settling these questions on students’ behalf in advance. The same review found institutional guidelines, and students’ involvement in constructing rather than simply receiving them, to be a further condition supporting this kind of development. A small but growing literature beyond assessment specifically favors cultivating student agency in academic integrity alongside compliance-based approaches (Brickhill et al., 2025; Morari et al., 2026). Both understandings of AI as a wicked problem and postplagiarism warrant moral agency support as a complementary, educative layer alongside institutional compliance mechanisms as a guarantee of fairness. Further, student understanding of academic misconduct and how to avoid it, peer disapproval of academic misconduct, clear institutional messaging, and structural assessment redesign have been reported to support a culture of academic integrity (Ali and Maroulis, 2026; Parson, 2026; Sharma, 2026; Ying, 2026). This perspective article adds to these understandings a more explicit focus on developing student moral reasoning and conduct in AI-assisted assessments.

Current empirical research on students’ perceived fairness in AI-assisted assessment treats them as reporters, without engaging further with why they perceive as such and what value sources underpin such perceptions. Students’ own accounts of AI-assisted assessment reveal ambivalence rather than settled moral reasoning. Chai et al. (2024) found students judged AI-based grading fairer than human grading, while Kizilcec et al. (2024) documented that some students regarded using AI to improve an already AI-generated piece of writing as no longer legitimate research, while others considered improving an AI-generated draft entirely acceptable. Yusuf et al. (2024) reported that a substantial majority of 1,217 survey respondents regarded incorporating AI into academic work as outright cheating. Luo (2025) found students fearful of using AI tools, given how unclear the boundary between acceptable and unacceptable use felt; a state Chan (2024) terms AI guilt, a moral discomfort at using AI for tasks traditionally carried out by the student alone. Chan’s (2025) survey of students’ perceptions of “AI-giarism” (academic misconduct involving AI) found a full standard deviation of variability across items, suggesting these judgements are genuinely subjective rather than settled: different students draw the line between legitimate and illegitimate AI use in materially different places.

A smaller body of work goes further than reporting confusion, and is instructive for what moral agency in this space would actually involve. Bearman et al. (2026) found that students’ moral judgements about AI use were bound up with emotions such as fear, guilt, and relief confirming that moral agency carries an affective component. In this study, some students reasoned that their work “should come from me,” locating fairness in whether the work represented their own self, not merely whether it could be attributed to them, a different concern from validity, which asks whether an assessment accurately captures what a student knows, not whether the work reflects who they are. In some students’ fairness accounts, work is meaningful when it represents them, which involves ontologies of who they are, emotional processes of what they value, and moral self of what they see as right for themselves. Such an account differs from the technical nature of validity that concerns collecting and evaluating evidence often by external experts showing what students accurately know and can do. This may explain why students reach for “fairness” rather than “validity” to describe what is at stake for them. The use of validity term with its technical nature further marginalizes student active positioning. The same study found that students exercised moral agency by weighing AI's usefulness for improving their work against a fear of becoming dependent on it. Some students deferred entirely to an institutional policy for what counted as acceptable, while others held moral positions of their own that diverged from that policy, demonstrating a need for guided support for developing student moral agency.

A further gap concerns the distance between reasoning and conduct. Ying’s (2026) scoping review of 38 studies on student voice in AI-related academic integrity found that students who judged a practice unethical nonetheless sometimes engaged in it, citing moral justification, unclear institutional guidance, workload and time pressure, and the normalizing effect of peers doing the same. These results echo several of the very disengagement mechanisms already discussed in relation to Bandura et al. (1996). The same review found students describing markedly different postures toward AI including resistant, skeptical, enthusiastic, pragmatic and consistently perceiving its ethical boundaries as blurred. Importantly, Ying (2026) also found that students themselves expressed interest in developing their own moral agency in this space, with empirical research on how to do so simply is gradually burgeoning (Parson, 2026).

All this is evidence of both moral confusion and a reasoning-conduct gap in AI-assisted assessment, emphasizing the need for students to have sustained, longitudinally studied, opportunities to reason through fairness issues in academic integrity cases as they arise. Such an effort can support students in noticing where what they believe and what they do diverge, and seeing how their own moral reasoning develops as a result. Developing student moral agency would extend to helping students see the network of category relations underpinning their own reasoning — validity, integrity, originality, individualism, equality, meritocracy — and offering alternative ways of thinking about fairness beyond that network.

3.6 Synthesis: from reporting and compliance to moral agency

Across the five domains, the same pattern recurs. Students are positioned either as reporters of fairness perceptions or as compliant subjects of procedures designed by others (Table 1). Teachers’ ethical capacity is modeled but its effect on students goes unexamined; self-assessment bias is calibrated against a teacher's judgement; unfairness in group work and accommodation is surveyed; and declaration forms are administered in AI-assisted assessment.

Table 1. Student roles in reporting/compliance versus moral-agentic positioning across five assessment domains.

Underlying this pattern is the individualism-equality-meritocracy nexus: assessment attributes work and responsibility to individuals, standardizes conditions and treatment, and ranks and rewards demonstrated ability. The nexus is explicit in assessment literacy, self-assessment, collaborative, inclusive, and AI-assisted assessment.

A moral-agentic positioning changes the student's role in each domain (Table 1), but the underlying moves are shared. Students are asked why they hold their fairness views, not only what those views are. They are helped to see the network of categories underpinning that reasoning, and they are offered alternatives such as participation, representation, and relational recognition. They then practice weighing these alternatives in concrete dilemmas, with feedback, over time. This is what treating fairness reasoning and conduct as a developmental capacity requires across domains.

4 Conclusion: cultivating moral agency as a developmental achievement

This perspective article has traced how fairness has been framed in classroom assessment research in order to outline a research agenda for the next decade. Fairness research in classroom assessment has gradually shifted toward recognizing students as partial subjects. Theoretical research on student voice, agency, assessment literacy, and partnership in assessment has begun to position students more actively in the assessment process. Empirical research has focused on what students report to be fair or unfair in classroom assessment. Scanning across the assessment domains (Table 1) reveals that students are asked what they perceive to be fair, rarely why they believe it, and almost never invited to reason toward a position of their own or to reflect on how they have developed their fairness reasoning and conduct across their assessment dilemmas. Yet recognizing students as agents does not mean their agency will always serve academic learning or fairness. Students can exercise their agency in ways that undermine assessment validity and academic integrity, even in formative and low-stakes contexts (Harris et al., 2018). Students may resist engaging in assessment processes (e.g., not completing tasks, purposefully underperforming), subvert assessment through cheating, or contest assessment practices by rejecting feedback. Students justify these actions via protecting their egos or valued relations with peers and teachers, prioritizing efforts towards tasks that are tied to marks, and investing minimum effort to achieve maximum results (Harris et al., 2018). These findings, however, need to be further examined once they receive targeted opportunities to develop their moral reasoning and conduct.

Thus, the research agenda for the next decade should explore how students reason and behave in relation to fairness across various assessment domains, as well as how they can be supported to develop their moral agency. Such research has begun in preservice teacher education using the CAFI (Rasooli and Holden, 2026). However, a parallel program of research for students needs further empirical attention (Parson, 2026). This research can develop and examine structured deliberation exercises inviting students to construct and defend positions on contested fairness questions in self, peer, and collaborative assessment, inclusive assessment, and AI-assisted assessment, rather than only reporting perceptions of them. By working through moral assessment dilemmas, students can train their moral agency to decide which courses of action are the right ones amid competing normative choices.

In practice, students’ moral agency can be developed by using instruments such as the CAFI that include scenarios with dilemmas spanning group work, cheating, exams, grading, and feedback. Students can be invited to respond to these scenarios individually, share their responses with peers, reflect with peers on what principles of fairness drive similarities and differences in their responses, and engage in class reflection with their instructors. The purpose of such classroom deliberations is to provide space for students to examine their tacit fairness assumptions and develop awareness of alternative ways of thinking about fairness. Students need additional time to face their fairness values critically by evaluating what sources have contributed to shaping them. This process is ontological, emotional, and moral as also shown by findings from Bearman et al. (2026) because students engage with values that shape their moral being and self. Instructors need to be aware of the strong emotions that may surface in this process of moral agency development and remind students that such emotional processes are part of the journey. Students may seek definitive answers about what fairness is in specific scenarios, reflecting the natural tendency toward unity and coherence in the humanist model of the moral self (Pesch, 2020). However, instructors can support students in moving toward appreciating the plurality of discursive categories such as equality, individualism, meritocracy (among others such as those previously discussed for AI-assisted assessment), that shape a boundary for what fairness means in a particular context. By showing how this network of categories works together to shape fairness across domains of fairness, students can come to appreciate how assessment is structured in our societies and how it inculcates values that shape perceptions of teachers, students, and parents. This work is needed to move toward alternative ways of thinking about fairness and how else assessment practices can be arranged in society, with students as co-reasoners rather than only reporters of fairness.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. I used GenAI for editing the text; the ideas are the author’s and he takes full responsibility for them.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Gavin T. L. Brown, The University of Auckland, New Zealand

Reviewed by: Duen-Huang Huang, Chaoyang University of Technology, Taiwan