Sage Journals: Discover world-class research

Abstract

This study investigates experts' discourses of children in the context of the datafication of education in Finland. Drawing on interviews with 25 experts from education policy, educational technology, and data justice sectors, the present research identifies four key discourses: the “agentic child” as an active participant in data practices; the “juridical child” requiring protection under data laws; the “citizen-child” prepared for future societal roles through data literacy; and the “learning child” benefiting from personalized, data-driven pedagogy. These discourses reveal a complex, multifaceted conceptualization of children, shaped by ethical, legal, and pedagogical considerations, highlighting the interplay between childhood agency, rights, and education in a data-driven society.

Keywords

Children datafication discourse education expert

Introduction

Youth, growth, and childhood and all such things have many things that are formless and somewhat confusing. And things, which have remained somewhat unnamed (Data justice J).

The excerpt cited above originates from one of the experts interviewed for this study. Describing childhood as “formless,” “confusing,” and “unnamed” could hardly be further removed from the precision-oriented language around data and datafication, in which “artifacts in the world are turned into data through extraction, measurement, labeling, and ordering” (Crawford, 2021: 127). Nevertheless, these two contrasting discursive spaces—child and data—are regularly brought together in public and scholarly discussions regarding the datafication of education. In these discussions, precise and organized data is typically promised to help teachers better understand the messy and unorganized child (e.g., Couldry and Yu, 2018; Garner, n.d; O’Neill et al., 2022; Perttola, 2018; Mertala, 2024).

Even though datafication of childhood has been a subject of a growing research interest during the 2000s, the discursive formation of “the child” in the context of datafication remains to be unstudied. Instead, the focus has been, for instance, on the kinds of “data doubles” that are created of children (e.g., Bradbury, 2020) or the politics and processes of datafication as micro and macro level phenomena (e.g., Palsa et al., 2025; Pangrazio and Maova, 2025). This study contributes to following this gap in knowledge by examining experts’ discourses of children in the context of the datafication of education. Discourse is a fluid concept that can be used in a particular as well as a general, abstract way (Fairclough, 2003). In this paper, discourses are understood as texts, which may represent “the same area of the world [here children and datafication] from different perspectives” (Fairclough, 2003: 26), which locates within the more particular end of the continuum. Our approach to datafication of education grounds on Jarke and Breiter’s (2019: 1) definition in which it “comprises the collection of data on all levels of educational systems (individual, classroom, school, region, state, international), potentially about all processes of teaching, learning, and school management”. That said, the definition of the datafication of education is expanded to include teaching children about data¹: one objective of contemporary education is argued to be preparing students to “live with datafication” (Pangrazio and Sefton-Green, 2022), typically operationalized as different forms of data literacy education (see Fagerlund et al., 2025).

Discourses about children and childhood matter because they “represent a wide range of ideas about who children are and can be, as well as how they should and could live their lives” (Smith, 2015: 21). As the use of the words “are” and “should” suggests, discourses possess normative and performative power (Fairclough, 2011). This makes the study of experts’ discourses particularly significant in the educational sector. Experts represent a powerful group whose views are often considered in decision-making: “a person is considered an expert if she or he possesses an institutionalized authority to construct reality,” as Meuser and Nagel (2008: 9) neatly put.

Research on experts’ discourses is especially relevant to emergent phenomena, which are under constant negotiation. In many countries, the datafication of education is in an emerging stage and in search of best practices and policies. Take Finland, the empirical context of this study, for example. School lockdowns in 2020 resulted in widespread concerns about learning loss (Mertala, 2021), a discussion further accelerated by declining scores in cross-national assessments such as PISA (Hiltunen et al., 2023), ICILS (Fagerlund et al., 2024), and TIMSS (Hiltunen et al., 2024). The use of data-generating educational technologies, such as learning analytics (potentially combined with children’s register data), has been proposed as a solution to these issues, with large-scale development projects funded by the government and municipalities (EDUCA, n.d.; DigiOne, n.d). At the same time, concerns about children’s privacy and their right to data literacy education are regularly raised by public administrators (Salomaa and Palsa, 2019) and researchers (Silvennoinen et al., 2024). This suggests that the discourses of children and the datafication of education are more multivocal than unified. To better understand the variety and nature of the discourses of children and the datafication of education, this paper asks the following research question: What kinds of discourses do experts express about children in the context of the datafication of primary and secondary education in Finland?

Discourses of children and childhood: Theoretical underpinnings

Outside of fields like pediatric medical science, positivistic developmental psychology, or econometrics, childhood is often approached as discourse—a body of representational practice (Ryan, 2016). Taking a discursive stance means that “child” and “childhood” are considered matters of continuous negotiation and reproduction in a particular society at a specific historical time (Alasuutari and Karila, 2010). An illustrative way to exemplify the “negotiable” nature of a child is to consider the division of the child either as a being or a becoming. Social scientist Uprichard (2008: 304), drawing on a plethora of literature, summarizes their differences as follows:

The ‘being’ child is seen as a social actor in his or her own right, who is actively constructing their own ‘childhood,’ and who has views and experiences about being a child; the ‘becoming’ child is seen as an ‘adult in the making,’ who is lacking universal skills and features of the ‘adult’ that they will become.

The view of a child as becoming was (and is) common in developmental psychology, especially in the phase and stage theories by figures like Jean Piaget, as well as in “traditional” sociology (see Johnsson, 1995; Matthews, 2007). In the 1980s and 1990s, the idea of childhood as (only) a natural phase of life (Alasuutari and Karila, 2010) was contested by proponents of the new sociology of childhood (e.g., James et al., 1998). They argued that children are social actors capable of making sense of and affecting their societies (Matthews, 2007). However, the introduction of new ways of seeing the child did not lead to a “Kuhnian paradigm shift,” and the two views still coexist (see Ryan, 2016). In fact, Uprichard (2008; see also Matthews, 2007) states that neither approach is satisfactory in itself, but that they can be used together in complementary ways.

Discourses about the child are not restricted to those of being and becoming. Another well-known example involves the discourses of “the evil Dionysian child,” “the innocent Apollonian child,” and “the responsible Athenian child” (Jenks, 2005; Smith, 2012). These general discourses are typically refined when children’s discourses are explored in more specific contexts. Research on discourses of children and digital technology, for instance, has identified various detailed imaginaries, ranging from a “Dionysian” dangerous technology user to an “Apollonian” vulnerable victim and a digitally disadvantaged needy child, to name only a few examples (e.g., Mertala, 2019; Selwyn, 2003). According to Selwyn (2003), the dangerous child is actively and aggressively using technology, potentially harming both themselves and others. Examples of the vulnerable victim discourse, in turn, include concerns about how technology use may detrimentally affect children’s physical (e.g., myopia, poor posture) and mental health (Dong and Mertala, 2021). The digitally disadvantaged needy child, who lacks access to digital technologies (Mertala, 2019), closely resembles the “child in need” discourse common in social policy (Moss et al., 2000).

There is evidence suggesting variation within the discourses of children in the context of the datafication of education. In their analysis of edtech corporate discourses, media scholars Couldry and Yu (2018) noted the existence of a discourse of a “responsible student,” which aligns with the image of the Athenian child. Responsibility, in the context of datafication, refers to the child’s obligation to generate data by using digital technology like learning analytics: if data is the key to better learning, then data generation becomes a responsible action.

The responsible child discourse provides insights into how the datafication of childhood has been naturalized within commercial discourses, where surveillance is constructed as inherent to childhood itself and even necessary for the health, development, and learning of the child (Couldry and Yu, 2018; Macheroni, 2020). However, at the same time, the intensifying dataveillance during childhood is also seen as highly problematic. Savirimuthu (2020: 310), a legal scholar, has asked “whether the political act of integrating the lifeworlds of children into the technological infrastructures of the personal data economy and framing of legal responsibilities to be owed by data controllers through data protection rules and principles is truly empowering.” In other words, for Savirimuthu (2020), the datafied child appears disempowered and in need of juridical protection, thus representing another kind of discourse.

The present study

The research reported in this paper is part of the broader Research Council of Finland funded research project Movement for Data Literacy, which examines the datafication of primary and secondary education in Finland. The data for the study was collected as part of the project’s first work package, the objective of which was to map the macro-level discourses surrounding the datafication of Finnish primary and secondary education.

Data and participants

The data consists of research interviews with 25 experts representing the fields of educational policy (n = 7; e.g., members of public administration), educational technology (n = 8; e.g., technology company representatives), and data justice² (n = 10; including researchers and non-governmental organizations addressing data from a societal perspective). The combination of these domains was believed to provide multivocal data offering a holistic view of the macro-level discourses of the datafication of education. The interviewees were selected via preliminary screening (e.g., organizational websites, news, white papers) as well as snowball sampling.

The interviews took place between November 2023 and February 2024. Four interviews were held face-to-face, while 21 interviews were conducted remotely via Zoom. Author 1 conducted seven interviews, and Authors 2 and 3 conducted nine each. All interviews were recorded and transcribed. The length of the interviews varied between 47 and 122 minutes, and the total amount of data is 31 hours and 18 minutes/227427 words. The data was pseudonymized and stored securely in an online space (Nextcloud) on the university’s servers. To protect the anonymity of the experts, they are referred to with a code that combines the domain (educational policy, educational technology, or data justice) with an alphabet.

The interview process applied the principles of semi-structured expert interviews (Meuser and Nagel, 2008). Instead of asking each participant the same questions in the same order, a detailed manual consisting of themes to be addressed during the interviews was created (DiCicco-Bloom and Crabtree, 2006). The main themes, introduced in Table 1, were provided to the participants in advance in a summarized form.

Table 1.

Main themes of the interview.

Main theme	Description
General perspective on datafication	Exploring datafication, including the production, collection, and use of digital data as part of an ordinary school day, as well as broader administrative, educational technological, and policy perspectives. Participants’ views on related developments were also sought
Skill needs	Perspectives on the knowledge and skills required for datafication. Participants were asked whether specific competencies are necessary for students, education professionals, or others
Current situation in schools	Views on the current state of datafication in schools, examined at various levels (e.g., classroom practices, administration, education policy, or the educational technology sector)
Background of the current situation	Participants’ perspectives on how the current situation arose: What factors or elements contributed to the present state of datafication
Evaluation of the current state	Assessments of the benefits or challenges associated with school datafication, analyzed across different levels (e.g., classroom practices, administration, policy, or the technology sector)
Future	Participants’ views on potential future scenarios for datafication and the factors likely to influence them
Open floor	Participants were invited to share additional insights or perspectives not covered by the structured themes

Additionally, we prepared a figure (see Figure 1) as a narrative artifact (Uitto et al., 2023) to facilitate the interviews. The manual and the figure were tested and refined through two rounds of pilot interviews. In four interviews, additional materials from the participants’ organizations’ websites or other resources were used to support more concrete and contextualized discussions.

Figure 1.

The narrative artifact used in the interviews.

Interviews began with a warm-up chat, which was not recorded. Although research notifications and privacy notices were provided to participants upon first contact, the recorded interview started by confirming that the participants had read, understood, and accepted these documents. If a participant had not read the documents, they were reviewed and accepted before proceeding. Before addressing the main themes, we asked participants to discuss their current and past work and its relevance to the datafication of education.

Analysis

An abductive approach was employed for data analysis. This approach rejects the idea that the researcher’s observations and interpretations could be purely inductive and acknowledges the inclusion of guiding theoretical threads in the analysis process (Grönfors, 2011). The primary theoretical threads in this study were the discourses of children as beings and becomings (Uprichard, 2008), evil, innocent, and thoughtful (Smith, 2012), and needy or rich (Moss et al., 2000). Additional threads included the child as a subject entitled to data protection/regulation (Savirimuthu, 2020) and the child as a responsible student (Couldry and Yu, 2018). However, unlike in deductive analysis, following a theoretical thread does not mean taking the theory as given or using the analysis process solely to test it. Instead, theory-informed codes and categories were refined, reformed, and expanded through inductive interpretations.

The qualitative analysis software ATLAS.ti (version 23) was used for open coding. All parts of the interviews discussing children and data were marked, accompanied by short memos outlining initial interpretations. After coding, the excerpts were analyzed horizontally (reading all thematically similar excerpts) and vertically (reading all coded excerpts from an individual interviewee) to gain an overall understanding of the discourses. The interpretative process is illustrated through one example excerpt³:

Well, especially, something like school dropout prediction, it requires quite a few different data sources to combine, starting from social services, so it is of course a bit challenging in terms of data protection [...] that if it gives the teacher a better idea of how the teaching group is progressing, where perhaps [they] need to intervene (Data justice C).

From this excerpt, the following interpretation was made: A child with a specific background—which can be deduced from data sources, including registry data from social services—is profiled as a potential school dropout. Thus, the child is seen as a needy (Moss et al., 2000) becoming (Uprichard, 2008), whose future societal opportunities depend on data-informed interventions by the teacher. This vignette was coded as representing a citizen-child discourse, where the goal of datafication in education is to prepare children for the future by addressing disadvantages arising from socioeconomic background. The participant also reflected on children’s juridical rights regarding data collection, sharing, and profiling (Savirimuthu, 2020), suggesting the presence of a juridical child discourse. In this view, childhood is seen as valuable in its current state (being, Uprichard, 2008) and deserving of protection (innocence, Smith, 2012). Additionally, the excerpt discussed children collectively rather than as unique individuals. Through this process, we identified four discourses related to children and datafication: the agentic child, the juridical child, the citizen-child, and the learning child. These will be explored in detail in the following sections, along with the discursive space they form.

Findings

We will introduce each of the discourses summarized in Table 2 in individual sections, beginning with the Agentic child. Before proceeding, we wish to emphasize that the identified discourses are not mutually exclusive categories. As the reader will observe—through the pseudonym codes in the data excerpts—individual experts often described children and the datafication of education using multiple discourses. This observation is revisited and further explored in the discussion section.

Table 2.

Summary of the discourses.

Discourse	Agentic child	Juridical child	Citizen-child	Learning child
Main temporal dimension of the discursive space	Being	Being	Becoming	Becoming
Main ontological dimension of the discursive space	Individual	Collective	Collective	Individual
Relationship between children and datafication	Participation	Protection	Preparation	Personalization
Summary of the discourse	A child as a unique subject who is an active participant in school’s data policies and practices, either by utilizing data for personal or communal benefits or by resisting data generation	Children as an age-based collective who, due to their legal rights as minors, are entitled to be protected by adults against unnecessary and potentially harmful data generation and use	Children as an age-based collective of “future adults in-the-making” for whom formal education should teach data literacy and provide data-informed interventions if the child is at risk of societal exclusion through school dropout	A child as a unique subject whose learning processes and pathways must be personalized via data-informed pedagogy to maximize their learning potential

The agentic child

In the agentic child discourse, children are seen as active subjects in relation to the datafication of education. Agency is described as participation in negotiations about the school’s data practices, rather than mere compliance. Thus, the agentic child is dominantly an individual being—a social actor in their own right (Uprichard, 2008). Examples of agency are divided into two strands: resistance and utilization.

Agency in resistance frames the child as someone who can opt out of using data-generating technologies or manipulate the data generation process so that the resulting data becomes unusable or meaningless. One participant described how even a single student’s refusal to participate could disrupt the system, highlighting children’s ability to undermine the utility of data through intentional non-compliance or manipulation:

Let's just say that I would predict that in a classroom environment where there are a lot of other things, so if in principle even one student is of the opinion that she does not want to use this, and then you have to come up with an alternative [...] I could also believe that when you go to the [data] produced by the children themselves, so when adults are quite good at thinking and manipulating and messing around, so children can equally think and manipulate and mess around about [data]. (Data justice G).

Agency in data utilization describes goal-oriented, self-initiated choices and actions involving data, through which the child seeks personal or communal benefits. One participant envisioned children using self-tracking technologies, such as sports watches, to bring their own health and activity data to annual meetings with the school nurse. This would enable the nurse to provide more personalized feedback based on the child’s unique data: “Could it be that the student has her own data, and she can show the nurse the annual report, that this is my set for this year?” (Technology U). Another example was provided by Technology Z who proposed that children could use digital portfolios to present teachers with a more comprehensive view of their skills and accomplishments. This approach would allow children to showcase abilities developed outside the school setting, offering a fuller picture of their competencies:

I myself have sometimes thought that when a pupil is assessed at school, she could probably give evidence of her own competence in other settings than school. Because they may have abilities and skills, but for one reason or another, they are not able to show them at school. Then if it could, let's think of it this way: if the pupil is a captain of a sports club, for example, when she has been selected as a captain, she has certain qualities. But if she cannot show those qualities at school, then could she somehow demonstrate her own competence? Or even in 4H club activities or scouting certain things. That would be useful at school too.

The juridical child

In the juridical child discourse, childhood was primarily described as an age-based collective where all children are subject to the same legislative regulations. References to documents such as the General Data Protection Regulation (GDPR) were common. While children were often described as beings who possess rights regarding their data—such as the right to understand their data rights or to provide consent for certain systems (“to make a person, a child, a learner understand what my rights are”; Data justice C; “Then, with the consent of the learners, they can also be asked if they want to introduce such a system”; Policy M)—the ultimate objective of regulation was to protect children.

Protection was approached from two perspectives: data minimization and data management. In the first perspective, a foundational concern prompting the need for protection was that technology companies and other agencies might generate and collect more data than necessary for supporting children’s learning and well-being. Therefore, it is the responsibility of adults, such as educators and policymakers, to monitor and control the generation, handling, and sharing of children’s data. In the data excerpt below, the interviewee discusses this topic with an implicit reference to the principle of “data minimization,” which states that “a data controller should limit the collection of personal information to what is directly relevant and necessary to accomplish a specified purpose” (European Data Protection Supervisor):

“In my opinion, we must monitor what is in the best interests of the learner and what is really relevant to their well-being in the long term. And also to protect children. In a certain way, even if a company or an agent thinks it is great that data are collected from a wide range of sources and everywhere, we should always consider the child’s right to privacy and the right not to collect any data other than what is necessary. And after that, how that data is handled, what rights there are to give the data away. (Policy M emphasis added).

The imperative for regulation was not restricted to commercial actors such as technology companies and data brokers. Stakeholders also questioned how children’s data was managed within schools themselves. For instance, one participant pondered whether it is always in the best interest of the child to have their data transferred from one teacher to another when starting a new grade or changing schools. This participant implicitly referred to GDPR’s “right to be forgotten,” which gives individuals the right to request that organizations delete their personal data:

...whether it would be better for pupils to have the right to start with a clean slate, so to speak. And in some cases, it might be good to have a fresh start in a new environment. There would be no baggage, but the interpretations are made from ‘who I am now’ and not from ‘who I have been’ in the past. In other words, who owns the data, and do the student or students have the right to be forgotten? (Technology Z; emphasis added).

On the other hand, children were also seen to have limited capacity to make informed decisions regarding the lifespan of their data. Thus, they require protection not only from external actors but also from their own actions, such as opting out of data generation and collection—a choice they might later regret. The tension between protecting children’s right to express their views and safeguarding them from potentially harmful decisions is neatly illustrated in the following excerpt. Here, the participant first frames children as beings, emphasizing the importance of respecting their right to be heard, referencing the United Nations Convention on the Rights of the Child (particularly Article 12). They then contemplate the appropriate age at which children can make sustainable and responsible decisions, thereby shifting the perspective to view children as becomings:

However, we should think about hearing the child. If you think about the Convention on the Rights of the Child and all the things that should take the child and her opinion and views into account. So, at what point, when, is the child old enough to decide in a certain way? [...] And then, if the data is destroyed for some reason, and then you realize 10 years later that it would have been good to have this data, because now I understand that I would have needed it. (Data justice I).

The citizen-child

In the citizen-child discourse, children were predominantly described as a collective of becomings, future adults in-the-making. The purpose of datafication in education was framed as preparing these children for the future by providing them with essential skills and addressing societal disadvantages. Regarding skill development, children were described as needing to learn about data and datafication—what data is, how data can be used, and how datafication affects people—to become fully functioning members of society in the future. Experts frequently commented that data literacy should be recognized as a set of “civic skills” (Policy L) and “core competences” (Policy Q) that all citizens should possess. As Data justice G explained:

If we think about Finnish politics, everything is aimed at ensuring that we have good and qualified workers who produce tax money, so I can see that digitalisation is being thought of and hoped that we will have qualified and competent people who can cope in today's world, or who will cope in the world of the future.

As the reader may recall, a different quote from Data justice G was used to exemplify the resisting form of the agentic child discourse (“I could also believe that when you go to the [data] produced by the children themselves, so when adults are quite good at thinking and manipulating and messing around, so children can equally think and manipulate and mess around about [data]”). In other words, Data justice G, like many other participants, did not approach children and datafication in education through a single discourse only.

Furthermore, comparing the excerpts above highlights how experts either aligned with or distanced themselves from certain discourses. In the first quote, Data justice G takes a direct and personal stance toward the agentic child discourse by stating, “I could also believe” in the agency of children. On the other hand, in the second quote, by saying, “if we think about Finnish politics, everything is aimed at” and “I can see that digitalisation is being thought of and hoped that”, the interviewee distances themself from the competence-focused discourse of the citizen-child, identifying it as being driven by politicians, for example.

In tackling disadvantages, the citizen-child discourse also drew on the view that some children, due to disadvantaged backgrounds (e.g., low socio-economic status, immigrant background), were generically at risk for school dropout and future social exclusion. In this framing, citizenship—understood as full membership in society—was seen as something these children would not achieve unless proactive, corrective, data-informed measures were taken in schools. As Technology R put it: “It is of high importance that we identify at-risk students already before problems arise.” Technology V elaborated on this idea, envisioning a “national data program” to proactively screen for potential school dropouts:

[The data] would certainly be used to identify drop-outs. So we could sort of see if we should put a little bit more money into potential drop-outs, because then it will come back many times over if we get them on the right side of the fence, as it were, and everything should happen in early stages.

The learning child

Lastly, the child was also represented as a learner whose educational path needs to be personalized to improve the efficiency of education. The learning child was predominantly seen as an individual whose unique needs, strengths, and desires must be considered in formal education. Since data generated in the present was perceived as especially beneficial for children’s future, children were primarily described as becomings. Occasionally, children’s needs were explicitly framed as learning difficulties, as illustrated by Policy K:

In a way, I just think that it [data] should guide the provision of support. If we really simplify things, then quite a lot of pupils who are now in some way symptomatic may have dyslexia or mathematical learning difficulties...

More commonly, however, personalization focused on tailoring the content, pace, and forms of learning individually for each child, regardless of any formal diagnoses. For instance, the following excerpt suggests that each child possesses a unique, temperament-based learning style that educational data technologies, referred to as “intelligent agents,” could identify and adapt to, thereby offering personalized tasks and content:

If I as a learner have a profile accumulated over a lifetime of my learning style, how hard a challenge I like to get, how I deal with disappointments and failures. So when I go to high school or university, I can directly put it [the learning profile] there, so that the intelligent agent who offers me learning tasks, based on it [data], directly knows how to act in the way that is most useful to me. That would be quite reasonable. (Data justice C).

Some experts criticized the traditional, curriculum-based approach to teaching, which allows little or no flexibility for children. According to Technology R the idea that all children must learn the same material at the same time and pace is unsustainable: “But do I want to learn on Wednesday, 14 January 2024, when the teacher opens the book on photosynthesis?” Instead, they argued that data should be used to provide children with their own personalized learning assistants:

The lesson begins. Each student takes their own ‘Netflix of learning,’ welcoming you with ‘Hey, welcome [student’s name], last time you learned about photosynthesis. Do you have any questions about it? Do you want me to help you with that? Do you feel you understand what photosynthesis means?’ And I say, ‘I don’t really understand it yet; can you tell me a little bit more about it and what I need photosynthesis for?’ Whereas another student in the class might ask, ‘This is quite simple for me, tell me more,’ in a kind of way. And a third will ask, ‘What does photosynthesis mean in Finnish? I don’t understand the word flower; what is a flower in Finnish?’ (Technology R).

Discussion and conclusions

The objective of this study was to explore and describe experts’ discourses of children in the context of the datafication of education. To gather multivocal data, we interviewed 25 expert stakeholders from the fields of education policy, educational technology, and data justice. Through an abductive analysis process, we identified four discourses: the agentic child, the juridical child, the citizen-child, and the learning child. Together, these discourses form a discursive space—illustrated in Figure 2—which describes and structures the variety of ways a given phenomenon is discussed; in this case, children and the datafication of education (Kalalahti and Varjo, 2012).

Figure 2.

Discursive space of children in the context of datafication of education.

As showcased in the Findings section, the experts rarely described children and the datafication of education through a single discourse. Instead, they—metaphorically speaking—moved within the discursive space, drawing on two or more discourses. In some cases, participants’ movement within the discursive space highlighted the complexity and multifaceted nature of the subject—children and the datafication of education—and demonstrated that their views could not be captured by a single discourse alone. In other cases, participants acknowledged the variety of discourses they were aware of and (more or less explicitly) aligned with or distanced themselves from certain ones.

As a result, the discursive space of children in the context of the datafication of education is simultaneously complementary and contradictory. This arguably reflects broader discourses of both children and data—considered together and as separate entities. In the following subsections, we provide insights into the foundations of these discourses and the discursive space they constitute. More specifically, we reflect on the contextual roots of the discourses and compare them with the archetypes of the Apollonian, Athenian, and Dionysian child, as well as discuss the ways in which data are discursively framed.

Contextual reflections

While discourses shape social reality, they are also shaped by it. In other words, they are always bounded by historical and contemporary contexts. For instance, explicit and implicit references to official guidelines and regulations—such as the GDPR and the United Nations Convention on the Rights of the Child—were evident in the juridical child discourse. Similarly, the discourse emphasizing the need for children to learn about data in order to become responsible and employable citizens resonates with discussions around media literacy (Uusitalo, 2015) and computing education (Mertala et al., 2020), as well as with the global objectives of data literacy education (Fagerlund et al., 2025).

The idea that data can be used to identify and support children from disadvantaged backgrounds reflects Nordic welfare ideology, which prioritizes social equality and aims to reduce disparities through universal measures such as free education (Klette, 2018). The disadvantaged citizen-child discourse may also have been amplified by recent findings from large-scale assessment studies such as PISA (Hiltunen et al., 2023), which show that students from higher socio-economic backgrounds significantly outperform those from lower ones. These results, which have received substantial attention in Finnish media and public discussions (e.g., Vashko, 2023), likely contributed to shaping this discourse.

Furthermore, discourses are also influenced by the immediate social context in which they are produced. Research interviews are a specific genre of interaction (Fairclough, 2003), and participants may have limited their responses to what they perceived as socially desirable—a common issue in qualitative research (Bergen and Labonté, 2020). Additionally, while our participants represented diverse fields (education policy, educational technology, and data justice), this list is not exhaustive. These factors should be acknowledged as limitations of the present study.

Reflections on archetypal discourses of children and childhood

The four discourses constructed in this study also draw from the “general” discourses of children and childhood. The citizen-child discourse, for example, aligns with the “child in need” discourse (Moss et al., 2000), portraying disadvantaged children as requiring data-informed interventions to prevent social exclusion. The agentic child discourse, particularly when discussing resistance through data manipulation, resonates with the “Dionysian evil child” discourse (Jenks, 2005), as it frames the child as acting against established norms and practices. Conversely, agency as self-initiated utilization of data aligns with the “Athenian responsible child” discourse (Smith, 2015), portraying children as responsible actors leveraging data for personal benefit. This image mirrors the “responsible student” described by Couldry and Yu (2018).

The Athenian child was also present in the learning child discourse. While children frequently use digital devices in schools for purposes beyond learning, such as gaming or watching YouTube (Högrström et al., 2024; Lakka, 2024; McCoy, 2020; Paakkari, 2020), the learning child discourse emphasized a wholehearted orientation towards learning. In contrast, the juridical child discourse closely resembles the “innocent Apollonian child” discourse (Jenks, 2005), portraying children as requiring adult protection from intense data collection and its unintended consequences (see also Savirimuthu, 2020). However, the view that children also need protection from their own potentially harmful decisions—such as opting out of beneficial data generation—introduces elements of the Dionysian child into the juridical child discourse.

The portrayal of children through different and even contrasting archetypes within and across discourses underscores the complex and “unstable” (Olson and Rampaul, 2013) nature of the discursive construction of children and childhood. As Uprichard (2008, p. 303; emphasis original) notes, “children and childhood are always and necessarily ‘being and becoming.’” Similarly, they are always and necessarily Apollonian, Dionysian, and Athenian. Therefore, while contradictions within a single discourse may initially seem counterintuitive, they arguably provide a more truthful representation of children—and people in general—as creatures of paradox.

Reflections on data discourses

Speaking of paradoxes, the discursive framing of data—often conveyed through explicit or implicit metaphors—seems to imbue the concept with somewhat contradictory qualities. Two of the most common metaphors depict data either as “a force of nature to be controlled” to prevent disasters, or as “fuel to be consumed” to achieve desired outcomes (Puschmann and Burgess, 2014; see also Nolin, 2020). Both of these framings were present in the discourses identified in this study. For instance, within the juridical child discourse, data were described as something that must be controlled; if left unchecked, it could have harmful consequences for children’s wellbeing and future. Conversely, the metaphor of data as fuel was particularly evident in the citizen-child and learning child discourses. In the former, (big) data were framed as fuel for identifying and rescuing pupils risk of dropping out. In the latter, data were seen as fuelling the learning process of each individual pupil by enabling tailored instruction.

Discourses as constructors of reality: Aspirations for future research

At the outset of this article, we argued that the normative and performative nature of discourses—particularly those of experts (Meuser and Nagel, 2008)—contributes to shaping social reality (Fairclough, 2011). Here, we reflect on what the identified discourses might mean for the future. Negotiations often involve seeking consensus and making compromises. Therefore, discourses that are harmonious or mutually supportive are more likely to be approved by different stakeholders and, consequently, represent the most probable future scenarios. Indeed, some of the identified discourses formed a rather cohesive whole, drawing on the same “general” childhood discourse. For instance, the (Athenian) agentic child who generates and utilizes data during both school and leisure time aligns with the (Athenian) learning child benefitting from personalized learning analytics powered by data accumulated throughout the course of their lives.

Discourses that emphasize the importance of ubiquitous data generation are likely to gain traction in the current zeitgeist. Not only is the cross-sectional datafication of education (i.e., the combination of registry data and learning data) embraced by governments and technology companies (e.g., EDUCA, n.d; Mertala, 2024), but there is also growing interest in the scholarly sphere in idiographic learning analytics, which develops insights about individual students through longitudinal data analysis (Saqr, 2023). Thus, drawing on Gee (2015), discourses that emphasize the importance—or even the necessity—of datafication for education and children can be conceptualized as the “Big ‘D’ Discourse,” a larger discursive context within which various “small ‘d’ discourses,” like the ones introduced in this study, are examined.

When it comes to future research, so far only limited work has explored how discourses (written with either an uppercase or lowercase letter) manifest in the realm of everyday schooling. Furthermore, as datafication is a global phenomenon, comparative studies are needed to identify which discourses are universal and which are more context-specific. In particular, discourses that “travel” across contexts should be critically examined, as they are often decontextualized and far removed from the realities of childhood (see, e.g., the ongoing discussion about “digital natives”; Mertala et al., 2024). While the four discourses of children and the datafication of education outlined in this paper are not exhaustive, they nevertheless offer conceptual and analytical lenses for such studies.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This work was supported by the Kulttuurin ja Yhteiskunnan Tutkimuksen Toimikunta of the Reserach Council of Finland,(355523).

ORCID iD

Pekka Mertala

Data Availability Statement

The data are available upon reasonable request. *

Notes

References

Alasuutari

Karila

(2010) Framing the picture of the child. Children & Society 24(2): 100–111. Available at: https://eric.ed.gov/?id=EJ871853

Bergen

Labonté

(2020) “Everything is perfect, and we have no problems”: detecting and limiting social desirability bias in qualitative research. Qualitative Health Research 30(5): 783–792. DOI: 10.1177/1049732319889354.

Bradbury

(2020) Datafied at four: the role of data in the ‘schoolification’of early childhood education in England. In: Jarke

Breiter

(eds) The Datafication of Education. London: Routledge, 8–22.

Couldry

(2018) Deconstructing datafication’s brave new world. New Media & Society 20(12): 4473–4491. DOI: 10.1177/1461444818775968.

Crawford

(2021) The Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. New Haven: Yale University Press.

Dencik

Hintz

Redden

, et al. (2019) Exploring data justice: conceptions, applications and directions. Information, Communication & Society 22(7): 873–881. DOI: 10.1080/1369118X.2019.1606268.

DiCicco-Bloom

Crabtree

(2006) The qualitative research interview. Medical Education 40(4): 314–321. DOI: 10.1111/j.1365-2929.2006.02418.x.

DigiOne . (no date). DigiOne. Available at: https://www.digione.fi/digione-eng/ (Accessed: 20 December 2024).

Dong

Mertala

(2021) Preservice teachers’ beliefs about young children’s technology use at home. Teaching and Teacher Education 102: 103325. DOI: 10.1016/j.tate.2021.103325.

10.

Educa . (no date). Educa Flagship. Available at: https://educaflagship.fi/en (Accessed: 20 December 2024).

11.

European Data Protection Supervisor. no date. Data Minimization. Available at: https://www.edps.europa.eu/data-protection/data-protection/glossary/d_en

12.

Fagerlund

Leino

Niilo-Rämä

, et al. (2024) Kohti Digiosaamisen Strategista Kehittämistä: Kansainvälinen Monilukutaidon Ja Ohjelmoinnillisen Ajattelun Tutkimus (ICILS 2023). Jyväskylä: Koulutuksen tutkimuslaitos, Tutkimuksia, Vol. 40. Available at: https://jyx.jyu.fi/bitstream/handle/123456789/99007/ICILS2023-raportti.pdf?sequence=1&isAllowed=y

13.

Fagerlund

Palsa

Mertala

(2025) Exploration of domains of educational purpose in K-12 data literacy education research. Educational Research Review 46: 100663. DOI: 10.1016/j.edurev.2024.100663.

14.

Fairclough

(2003) Analysing Discourse. London: Routledge.

15.

Fairclough

(2011) Semiotic aspects of social transformation and learning. In: Rogers

(ed) An Introduction to Critical Discourse Analysis in Education. 2nd edition. New York: Routledge, 225–236.

16.

Garner

. (no date). Data in education: Part One of two. In: Learning A–Z. Available at: https://www.learninga-z.com/site/resources/breakroom-blog/data-in-education (Accessed: 20 December 2024).

17.

Gee

(2015) Discourse, small d, big D. In: Tracy

Ilie

Sandel

(eds) The International Encyclopedia of Language and Social Interaction. Hoboken: Wiley-Blackwell, 1–5. DOI: 10.1002/9781118611463.wbielsi016.

18.

Grönfors

(2011) Laadullisen tutkimuksen kenttätyömenetelmät. Hämeenlinna: SoFia-Sosiologi-Filosofiapu Vilkka. Available at: https://vilkka.fi/books/Laadullisen_tutkimuksen.pdf.

19.

Hiltunen

Ahonen

Hienonen

, et al. (2023) PISA 2022 Ensituloksia. Helsinki: Opetus- Ja Kulttuuriministeriön Julkaisuja 2023:49. Available at: https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/165295/OKM_2023_49.pdf?sequence=4&isAllowed=y

20.

Hiltunen

Kotila

Lehtola

, et al. (2024) Perustaidoilla Kohti Tasa-Arvoisempia Mahdollisuuksia: Neljännen Luokan Oppilaiden Matematiikan Ja Luonnontieteiden Osaaminen. Jyväskylä: Koulutuksen Tutkimuslaitos, Tutkimuksia, Vol. 41. Available at: https://www.jyu.fi/fi/file/perustaidoilla-kohti-tasa-arvoisempia-mahdollisuuksia

21.

Högström

Paakkari

Nilsberth

, et al. (2024) Sticky, slippery, and smooth access–exploring the affective potentials of access to writing technology in lower secondary classrooms. In: Learning, Media and Technology, 1–13. DOI: 10.1080/17439884.2024.2436566.

22.

James

Jenks

Prout

(1998) Theorizing Childhood. Cambridge: Polity Press.

23.

Jarke

Breiter

(2019) Editorial: the datafication of education. Learning, Media and Technology 44(1): 1–6. DOI: 10.1080/17439884.2019.1573833.

24.

Jenks

(2005) Childhood. 2nd edition. London: Routledge.

25.

Johnson

(1995) Constructing the child in psychology: the child-as-primitive in Hall and Piaget. Journal of Phenomenological Psychology 26(2): 35–57. DOI: 10.1163/156916295X00044.

26.

Kalalahti

Varjo

(2012) Tasa-arvo ja oikeudenmukaisuus perusopetukseen sijoittumisessa ja valikoitumisessa. Kasvatus & Aika 6(1): 5–20.

27.

Klette

(2018) Individualism and collectivism in Nordic schools: a comparative approach. In: Paulsen

(ed) Sustainable Modernity: The Nordic Model and beyond. London: Routledge, 59–78.

28.

Lakka

(2024) Takarivin Tekstikäytänteet. Doctoral dissertation. Helsinki: University of Helsinki.

29.

Mascheroni

(2020) Datafied childhoods: contextualising datafication in everyday life. Current Sociology 68(6): 798–813. DOI: 10.1177/0011392119865762.

30.

Matthews

(2007) A window on the ‘new’ sociology of childhood. Sociology Compass 1(1): 322–334. DOI: 10.1111/j.1751-9020.2007.00030.x.

31.

McCoy

(2020) Gen Z and Digital Distractions in the Classroom: Student Classroom Use of Digital Devices for Non-class Related Purposes . Nebraska College: Faculty Publications of Journalism & Mass Communications, Vol. 116. Available at: https://digitalcommons.unl.edu/journalismfacpub/116

32.

Mertala

(2019) Wonder children and victimizing parents–preservice early childhood teachers’ beliefs about children and technology at home. Early Child Development and Care 189(3): 392–404. DOI: 10.1080/03004430.2017.1314270.

33.

Mertala

(2020) Paradoxes of participation in the digitalization of education: a narrative account. Learning, Media and Technology 45(2): 179–192. DOI: 10.1080/17439884.2020.1696362.

34.

Mertala

(2021) Koulutuksen digitaalinen datafik(s)aatio. Kasvatus & Aika 15(1): 43–61. DOI: 10.33350/ka.100161.

35.

Mertala

(2024) Datafixation of education. In: Schiefner-Rohs

Hofhues

Breiter

(eds). Datafizierung (in) der Bildung: Kritische Perspektiven auf digitale Vermessung in pädagogischen Kontexten. Bielefeld: Transcript Verlag. Digitale Gesellschaft, Vol. 59, 31–42.

36.

Mertala

López-Pernas

Vartiainen

Saqr

Tedre

(2024) Digital natives in the scientific literature: A topic modeling approach. Computers in Human Behavior 152(108076). DOI: 10.1016/j.chb.2023.108076.

37.

Mertala

Palsa

Slotte Dufva

(2020) Monilukutaito koodin purkajana: ehdotus laaja-alaiseksi ohjelmoinnin pedagogiikaksi. Media & Viestintä 43(2): 74–93. DOI: 10.23983/mv.91079.

38.

Meuser

Nagel

(2008) The expert interview and changes in knowledge production. In: Bogner

Littig

Menz

(eds) Interviewing Experts. London: Palgrave Macmillan, 17–42.

39.

Moss

Dillon

Statham

(2000) The “child in need” and “the rich child”: discourses, constructions and practice. Critical Social Policy 20(2): 233–254. DOI: 10.1177/0261018300020002.

40.

Nolin

(2020) Data as oil, infrastructure or asset? Three metaphors of data as economic value. Journal of Information, Communication and Ethics in Society 18(1): 28–43. DOI: 10.1108/JICES-04-2019-0044.

41.

Olson

Rampaul

(2013) Representations of childhood in the media. In: Lemish

(ed) The Routledge International Handbook of Children, Adolescents and Media. London: Routledge, 49–56.

42.

O’Neill

Selwyn

Smith

, et al. (2022) The two faces of the child in facial recognition industry discourse: biometric capture between innocence and recalcitrance. Information, Communication & Society 25(6): 752–767. DOI: 10.1080/1369118X.2022.2035030.

43.

Paakkari

Antti

(2020) Entangled Devices: An ethnographic study of students, mobile phones and capitalism. Helsinki, Finland: University of Helsinki.

44.

Palsa

Fagerlund

Mertala

(2025) Exploring the futures of datafied welfare state education: thematic analysis of sociotechnical imaginaries. Social Inclusion 13: 10087. DOI: 10.17645/si.10087.

45.

Pangrazio

Mavoa

(2025) Studying the datafication of Australian childhoods: learning from a survey of digital technologies in homes with young children. Media International Australia 194(1): 53–68. DOI: 10.1177/1329878X231162.

46.

Pangrazio

Sefton-Green

(2022) Learning to live well with data: concepts and challenges. In: Pangrazio

Sefton-Green

(eds) Learning to Live with Datafication. London: Routledge, 1–16.

47.

Perttola

(2018) Data on tullut luokkahuoneeseen. In: Sanoma Pro. Available at: https://www.sanomapro.fi/data-on-tullut-luokkahuoneeseen/ (Accessed 20 December 2024).

48.

Polkinghorne

(1995) Narrative configuration in qualitative analysis. International Journal of Qualitative Studies in Education 8(1): 5–23. DOI: 10.1080/0951839950080103.

49.

Puschmann

Burgess

(2014) Metaphors of big data. International Journal of Communication 8: 1690–1709. https://ijoc.org/index.php/ijoc/article/view/2169/1162

50.

Ryan

(2016) Childhood as Discourse. Oxford: Oxford University Press.

51.

Salomaa

Palsa

(2019) Medialukutaito Suomessa. Kansalliset Mediakasvatuslinjaukset . Helsinki: Opetus- Ja Kulttuuriministeriön Julkaisuja 2019, Vol. 37. Available at: https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/161951/OKM_2019_37.pdf?sequence=1&isAllowed=y (Accessed 20 December 2024).

52.

Saqr

(2023) Modelling within-person idiographic variance could help explain and individualize learning. British Journal of Educational Technology 54(5): 1077–1094. DOI: 10.1111/bjet.13381.

53.

Savirimuthu

(2020) Datafication as parenthesis: reconceptualising the best interests of the child principle in data protection law. International Review of Law, Computers & Technology 34(3): 310–341. DOI: 10.1080/13600869.2019.1590926.

54.

Selwyn

(2003) Doing IT for the kids': Re-examining children, computers and the information society. Media, Culture & Society 25(3): 351–378. DOI: 10.1177/01634437030250030.

55.

Silvennoinen

Tedre

Valtonen

(2024) Datafikoituva peruskoulu–tasapainoilua lapsen henkilötietojen suojan ja opetuksen tavoitteiden välillä. Lakimies 122(5): 655–678.

56.

Smith

(2012) Producing governable subjects: images of childhood old and new. Childhood 19(1): 24–37. DOI: 10.1177/09075682114014.

57.

Smith

(2015) Deconstructing discourses to rupture fairytales of the “ideal” childhood. In: Wyn

Cahill

(eds) Handbook of Children and Youth Studies. Singapore: Springer, 505–520. DOI: 10.1007/978-981-4451-15-4_50.

58.

Uitto

Lassila

Jokikokko

, et al. (2023) Using artefacts in narrative pedagogies: a case from beginning teachers’ peer group meetings. European Journal of Teacher Education 46(1): 35–49. DOI: 10.1080/02619768.2021.1889506.

59.

Uprichard

(2008) Children as “being and becomings”: children, childhood and temporality. Children & Society 22(4): 303–313. DOI: 10.1111/j.1099-0860.2007.00110.x.

60.

Uusitalo

(2015) Hallinnan Tekemiset: Mediakasvatuspolitiikan Diskursiivinen Tuottaminen. Doctoral dissertation. Tampere: Tampere University.

61.

Vashko

(2023) Kuva näyttää karun tiedon Pisa-tuloksista: Näin valtava on muutos. In: Ilta-Sanomat. Available at: https://www.is.fi/kotimaa/art-2000010038169.html (Accessed 20 December 2024).