Engagement, preference and identity: Proposing a developmental engagement framework

Abstract

Understanding how adolescents participate in science requires examining the developmental foundations of engagement and identity formation. This theoretical paper introduces the Developmental Engagement Framework, which positions science engagement as the overarching construct, with science identity representing a developmentally advanced subset that emerges when engagement experiences are integrated with self‐concept. This framework addresses persistent empirical puzzles in science identity research—including temporal instability and limited predictive validity in adolescent populations—by proposing that these patterns reflect normal neurodevelopmental processes rather than methodological limitations. During early‐to‐mid adolescence (ages 11–16), when self‐referential neural networks remain under development, science participation operates primarily through preference‐based engagement—experience‐driven participation supported by earlier‐maturing reward and habit formation circuits (basal ganglia, striatum). As prefrontal systems mature in late adolescence, some of this engagement becomes integrated with self‐concept, emerging as identity‐infused engagement. Our framework complements existing identity theories by specifying their neurodevelopmental prerequisites and optimal timing, drawing on identity process theory, neuroscientific evidence on value‐based choice and recent findings on daily experience consolidation. We demonstrate that science identity represents a maturational transformation of engagement rather than a separate construct. This perspective has implications for age‐appropriate measurement, intervention design and theoretical integration across developmental stages.

Statement of Contribution.

What is already known on this subject?

• Science identity frameworks have provided valuable insights into how individuals connect with science and revealed barriers faced by underrepresented students.

• Existing models were primarily developed from adult populations and assume mature capacities for self‐reflection and stable self‐representation.

• Adolescent identity measures consistently show high temporal instability and modest predictive validity, typically attributed to methodological limitations.

What does the present study add?

• The observed instability in adolescent science identity reflects normal neurodevelopmental processes—specifically, the protracted maturation of self‐referential neural networks continuing into early adulthood.

• The Developmental Engagement Framework positions science identity as a developmentally advanced subset of engagement that emerges when experiences are integrated with self‐concept.

• The framework distinguishes preference‐based engagement (ages 11–16), supported by earlier‐maturing reward circuits, from identity‐infused engagement (ages 17+), enabled by mature self‐referential networks.

INTRODUCTION

In recent decades, science education has increasingly recognized that meaningful engagement with science extends beyond factual knowledge to encompass personal values, attitudes and beliefs (National Research Council, 2009; Osborne et al., 2003). Within this context, ‘science identity’ has emerged as an influential framework for understanding how individuals form connections with science, offering critical insights into persistent equity issues by examining the barriers faced by underrepresented students (Carlone & Johnson, 2007; Hazari et al., 2010). The field's rapid growth, with hundreds of articles published in recent years (Danielsson et al., 2023; Sandrone, 2022), highlights its perceived importance in addressing challenges of student participation and engagement (Henriksen et al., 2015).

However, despite this proliferation of research, the application of science identity frameworks to adolescent populations reveals persistent empirical patterns that warrant closer examination. Studies consistently report that identity measures exhibit high temporal instability during these years (Perez et al., 2014; Potvin & Hasni, 2014) and demonstrate modest predictive validity for long‐term educational or career outcomes (Maltese & Tai, 2010; Trujillo & Tanner, 2014). While often treated as methodological limitations requiring correction, this review offers a different interpretation: these empirical patterns reflect normal neurodevelopmental processes characteristic of adolescence rather than measurement failures.

This paper examines science identity through a developmental cognitive neuroscience lens, arguing that the adolescent brain's unique maturational trajectory creates distinctive conditions for science engagement (Eccles, 2009). The neural systems supporting coherent self‐representation—particularly the self‐referential processing networks involving the medial prefrontal cortex (mPFC) and posterior cingulate cortex (PCC)—undergo protracted development that continues into early adulthood (Gogtay et al., 2004; Martinelli et al., 2013). Consequently, applying adult‐derived identity constructs to adolescents may involve measuring psychological capacities for which the neural architecture remains under development (Pfeifer & Berkman, 2018).

Building upon these neurodevelopmental insights, we propose the Developmental Engagement Framework, which positions science engagement as the overarching construct encompassing both preference‐based and identity‐infused participation. Following identity process theory, which demonstrates that identity commitments emerge through accumulated experiences (Bosma & Kunnen, 2001, 2008) and neuroscientific evidence positioning identity‐related processes as inputs to motivated behaviour (Pfeifer & Berkman, 2018), we conceptualize science identity as a developmentally advanced subset of engagement that emerges when neural prerequisites are met. During early‐to‐mid adolescence (11–16 years), when self‐referential processing networks remain under development, science participation operates through preference‐based engagement—a mode supported by earlier‐maturing neural systems for reward processing and habit formation (basal ganglia, striatum). This engagement‐first perspective enables robust science participation without requiring stable self‐representation. As prefrontal systems mature in late adolescence, accumulated engagement experiences can be integrated with self‐concept, emerging as identity‐infused engagement. Not all preference‐based engagement transitions to identity‐infused participation; rather, identity formation represents a maturational transformation that occurs for some individuals under supportive conditions (Branje et al., 2021). Rather than replacing existing frameworks, this neurodevelopmentally informed perspective complements identity theories by specifying their prerequisites and optimal developmental timing. This analysis proceeds by (1) critically evaluating current science identity paradigms through a developmental lens, (2) examining neurodevelopmental evidence regarding adolescent brain maturation, (3) presenting the Developmental Engagement Framework (Figure 1) and its integration with classical developmental theories and (4) discussing implications for supporting diverse pathways to science engagement.

FIGURE 1.

The developmental engagement framework: science engagement as an overarching construct.

CRITICAL ANALYSIS OF CURRENT SCIENCE IDENTITY PARADIGMS

Science identity research has made substantial contributions to understanding equity issues and the social dimensions of science learning, illuminating barriers faced by underrepresented groups and emphasizing the importance of recognition, belonging and cultural factors in science engagement (Carlone & Johnson, 2007; Hazari et al., 2010). These sociocultural frameworks have proven valuable for explaining how social structures and power dynamics shape opportunities for science participation. However, when viewed through a developmental lens, empirical patterns indicate that such models may benefit from integration with neurodevelopmental perspectives, especially in adolescent populations. Considering that identity formation depends on maturing cognitive and neural capacities for self‐reflection, social evaluation and perspective‐taking (Blakemore & Mills, 2014; Pfeifer & Berkman, 2018), integrating these insights clarifies why recognition processes are particularly salient—and sometimes unstable—during adolescence.

Examining current science identity frameworks through a neurodevelopmental lens reveals implicit assumptions about cognitive and neural capacities that warrant closer scrutiny (Kim, 2018). Carlone and Johnson's (2007) foundational tripartite model—performance, competence and recognition—requires individuals to process social evaluation, integrate others' perceptions into their self‐concept and maintain stable self‐representations despite varying social contexts (Moore, 2008). Neurodevelopmentally, these capacities depend upon the coordinated functioning of self‐referential processing networks (medial prefrontal cortex) and social cognition networks (temporal–parietal junction). These neural networks undergo protracted development throughout adolescence (Blakemore, 2008; Pfeifer & Berkman, 2018), while simultaneously showing heightened sensitivity to peers' perceptions and evaluations—that is, how peers view and judge them (Somerville, 2013). This developmental pattern may explain why recognition proves powerfully motivating for adolescents yet produces unstable identity measurements: the sensitivity to peer‐based social feedback is present, but the neural architecture for integrating such feedback into stable self‐concepts remains under development.

Similarly, communities of practice frameworks (Lave & Wenger, 1991) emphasize identity negotiation through legitimate peripheral participation, assuming participants can navigate complex social dynamics while simultaneously developing domain expertise and constructing professional identities. This multifaceted identity work requires sophisticated social cognition, including perspective‐taking and understanding implicit community norms, alongside the capacity to reflect on one's developing identity within the community. These capabilities depend upon the integration of multiple neural networks—including prefrontal self‐referential networks and social cognition networks for navigating group dynamics—that continue maturing into early adulthood (Mills et al., 2014).

Quantitative approaches to science identity, while valuable for their scalability and analytic precision, implicitly assume cognitive and neural capacities that may not yet be fully developed in adolescent populations. For instance, Hazari et al.'s (2010) framework operationalizes science identity through self‐report items such as ‘I see myself as a science person’, enabling systematic investigation across large samples. However, this approach requires respondents to engage in explicit self‐categorization and abstract self‐reflection about their relationship with science. Such operations depend fundamentally upon mature self‐referential processing in the mPFC and posterior cingulate cortex—the neural substrates of self‐representation that exhibit continued development and limited functional integration during adolescence (Martinelli et al., 2013; Pfeifer & Peake, 2012). Consequently, while these survey instruments effectively capture momentary self‐perceptions, they may be assessing psychological capacities for which the underlying neural architecture remains under construction.

These observations suggest that the temporal instability and modest predictive validity observed in adolescent identity research may reflect not methodological inadequacy but rather the application of constructs that presume neural capacities still undergoing development. Rather than abandoning these frameworks, a neurodevelopmentally informed approach suggests they might be optimally applied during developmental periods when the requisite neural systems have matured, while complementary approaches better suited to earlier developmental stages could support the foundations upon which stable identities later emerge.

The empirical patterns observed in adolescent populations support this neurodevelopmental interpretation. Studies consistently document high temporal instability in identity measures, with substantial changes over periods as short as one academic year (Klimstra et al., 2010; Perez et al., 2014; Potvin & Hasni, 2014), alongside modest predictive validity for long‐term educational outcomes compared to other constructs such as self‐efficacy (Maltese & Tai, 2010; Trujillo & Tanner, 2014; see Stone, 2018, for neuroscientific perspectives on self‐efficacy). These patterns are consistent with the protracted development of self‐referential processing systems: measures assessing identity commitments show instability during the period when the neural networks supporting stable self‐representation undergo substantial reorganization.

Additionally, the predominant reliance on self‐report methodologies intersects with developmental considerations: sophisticated self‐reflection about identity commitments depends upon self‐referential processing systems that remain under development (Harter, 2012). The heightened social sensitivity characteristic of adolescence (Somerville, 2013) may also amplify social desirability effects, meaning measures may capture socially influenced momentary self‐perceptions rather than stable self‐concepts.

Cultural considerations add further complexity. Science identity frameworks predominantly emerged from Western, individualistic contexts (Aikenhead, 2006; Brickhouse et al., 2000), yet cultural neuroscience reveals systematic differences in self‐referential processing between individualistic and collectivistic cultures (Han & Ma, 2014; Salvador et al., 2022). Identity formation processes emphasizing individual self‐concept development may be neurobiologically, not merely socially, misaligned with students from cultural backgrounds where family and community recognition hold greater salience.

The heightened social sensitivity characteristic of adolescence (Mills et al., 2014; Somerville, 2013) creates additional interpretive challenges: declared identities may reflect immediate contextual influences and social positioning rather than stable, internalized self‐concepts, complicating the distinction between genuine identity development and context‐dependent self‐presentation.

These converging lines of evidence—conceptual analysis, empirical patterns, measurement considerations and cultural variations—suggest that current science identity frameworks would benefit from integration with developmental neuroscience perspectives when applied to adolescent populations. The observed temporal instability, modest predictive validity and contextual sensitivity are not methodological failures but rather patterns consistent with what would be expected when assessing psychological capacities dependent upon neural systems undergoing substantial development. Rather than replacing existing frameworks, which have proven valuable for understanding social and cultural dimensions of science engagement, a neurodevelopmentally informed approach can enrich current understanding by clarifying when and how different mechanisms of science engagement operate optimally across development. This perspective opens possibilities for developmentally responsive approaches that honor both the insights of identity research and the realities of the adolescent brain.

ADOLESCENT NEURODEVELOPMENT: UNDERSTANDING THE DEVELOPING MIND

The empirical patterns observed in current science identity research become comprehensible when examined through the lens of developmental cognitive neuroscience. The temporal instability and modest predictive validity in adolescents reflect a brain operating through mechanisms qualitatively different from adult cognition. Understanding these neurobiological foundations provides crucial context for why adult‐derived identity constructs require developmental consideration when applied to younger populations and points toward more developmentally suitable approaches to fostering science engagement.

The protracted development of self‐referential brain networks

Mature self‐representation depends upon the coordinated functioning of several interconnected brain networks that undergo protracted development extending well into the third decade of life (Oyserman et al., 2017). Research on self‐identity has established that self‐referential processing relies primarily on the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC) and angular gyrus—regions collectively forming the ‘self‐memory system’ (Martinelli et al., 2013; Northoff et al., 2006). These networks, which form the core of the brain's Default Mode Network (DMN), support the integration of autobiographical memories, self‐referential evaluation and the maintenance of coherent self‐narratives that characterize adult identity formation (Buckner et al., 2008). The mPFC, in particular, activates consistently during self‐referential mental activity (Gusnard et al., 2001) and represents the self‐concept in terms of subjective importance (Levorsen et al., 2023), a sophisticated cognitive operation requiring substantial neural maturation. However, the functional integration of these networks develops gradually throughout adolescence, suggesting the capacity for stable, coherent self‐representation is neurobiologically limited during the teenage years.

The sophisticated nature of self‐referential processing becomes evident when considering its cognitive demands. Mature identity formation requires not merely the recognition of personal characteristics but their weighted integration according to subjective importance hierarchies—what researchers term ‘self‐importance’ mapping (Levorsen et al., 2023). This process involves complex interactions between memory systems, evaluative processes and future planning capabilities, all of which depend on coordinated prefrontal function. The DMN, which underpins this integration, shows reduced connectivity and altered activation patterns during adolescence compared to adulthood (Blakemore, 2008), providing a neurobiological explanation for the instability observed in adolescent self‐concepts.

Structural neuroimaging studies consistently demonstrate that the prefrontal cortex continues to mature until approximately age 25, with the mPFC showing a particularly protracted development (Gogtay et al., 2004). During adolescence, these regions undergo extensive remodelling through synaptic pruning and myelination, which meaningfully shapes their functional characteristics. Grey matter volume decreases while white matter increases, reflecting the elimination of unused connections and the strengthening of frequently used pathways (Casey et al., 2008). This ongoing structural change means adolescent self‐referential processing may operate qualitatively differently from adult patterns, offering a direct neural explanation for the temporal instability observed in adolescent identity measures.

Functional neuroimaging supports this developmental account, revealing that adolescents show heightened activation in regions associated with social evaluation and reduced activation in areas supporting long‐term planning (Pfeifer & Peake, 2012). During self‐referential tasks, adolescents demonstrate increased activity in the medial prefrontal regions associated with mentalising about others' thoughts, suggesting their self‐concepts are heavily influenced by perceived social evaluation (Pfeifer & Berkman, 2018). This pattern contrasts sharply with adult self‐referential processing, which shows greater independence from immediate social context and stronger integration with autobiographical memory systems.

Recent research has further illuminated the dynamic nature of self‐identity during development. Huang et al. (2025) demonstrated using Bayesian network modelling that self‐identity functions within a complex network of affective and regulatory processes, with identity coherence playing a crucial role in anxiety management (see also Liu et al., 2024, for neural perspectives on self‐efficacy). This finding indicates that identity instability during adolescence may not merely reflect measurement error but rather genuine developmental volatility that has important implications for emotional well‐being and educational engagement.

The hypersensitive social brain and alternative neural pathways

In stark contrast to the slow development of self‐referential networks, the neural circuits responsible for processing social information and rewards are not only mature but become hypersensitive during adolescence. This creates a ‘developmental mismatch’ that is central to understanding adolescent behaviour (Steinberg, 2013). The adolescent brain demonstrates distinctive patterns of social information processing that differ markedly from adult social cognition, with implications extending beyond individual identity formation to encompass social learning environments.

Adolescents demonstrate heightened sensitivity to social evaluation and peer influence, with neuroimaging studies revealing exaggerated responses to social feedback compared to adult patterns (Somerville, 2013; Towner et al., 2023). Research on adolescent decision‐making in social contexts illustrates consistent activation in regions supporting the integration of affective, cognitive control and social information processing (van Hoorn et al., 2019). Adolescents show heightened activation in social brain networks when being observed by others, and the experience of social exclusion activates neural networks associated with physical pain processing (Masten et al., 2009). This neural overlap suggests that exclusion from science‐oriented peer groups may be experienced as genuinely distressing, while inclusion provides powerful motivation for continued engagement.

Conversely, social rewards such as peer approval show particular salience during this period (Izuma et al., 2008), often outweighing other forms of reinforcement. Research demonstrates that social and monetary rewards activate overlapping neural circuits within the striatum, but social rewards show enhanced activation during adolescence compared to adulthood. This heightened social awareness means adolescents are particularly attuned to peer reactions and vulnerable to social influences that can either support or discourage science participation (Mills et al., 2014).

The neural architecture supporting social cognition relies upon distributed networks including the mPFC, temporal–parietal junction and superior temporal sulcus, enabling theory of mind abilities and understanding of others' mental states (Schurz et al., 2014). These networks follow a protracted development timeline, suggesting that adolescents' capacity for navigating complex social dynamics in educational settings may differ qualitatively from adult expectations.

This social sensitivity is further shaped by cultural context. Cultural neuroscience reveals systematic differences in the neural substrates of self‐referential and social cognition between individuals from collectivistic versus individualistic backgrounds. East Asian cultures show increased neural activity in regions related to inferring others' minds and emotion regulation, whereas Western cultures show enhanced activity in areas related to self‐relevance encoding (Han & Ma, 2014; Salvador et al., 2022). Such findings suggest that Western‐centric emphases on individual identity achievement may be neurobiologically, not just socially, misaligned with students from different cultural backgrounds.

Crucially, the neural circuits underlying these reward and social processes, particularly the basal ganglia (striatum), mature earlier than the prefrontal networks required for stable identity. The striatum is a hub for neuroplasticity (Graybiel & Smith, 2011) and plays a central role in reward‐based learning and in transforming goal‐directed actions into automatic habits through repeated practice and reinforcement (Petrie et al., 2025; Yin & Knowlton, 2006). The striatum's role as a neuroplasticity hub is particularly relevant, as it receives massive input from aminergic systems including dopaminergic innervation and projections from nearly every region of neocortex, positioning it ideally to support gradual preference consolidation (Graybiel, 2008).

This distinction is critical: while the neural networks supporting stable, reflective identity remain under construction, the neural circuits supporting learning, habit formation and preference consolidation are fully operational and highly active. The adolescent brain's heightened plasticity (Steinberg, 2013) and reward sensitivity make this period ideal for forming lasting skills and affinities through repeated positive experiences, with dopaminergic reorganization contributing to enhanced learning capacity alongside variability in preferences and commitments.

Implications for a developmentally‐informed approach

This convergent evidence from developmental cognitive neuroscience suggests several important implications for understanding adolescent science engagement. The concept of neuroplasticity provides a crucial framework for understanding how brain development relates to science engagement. Unlike the relatively stable neural architecture of adulthood, the adolescent brain exhibits heightened plasticity that enables rapid learning and adaptation but also contributes to instability in cognitive and affective processes (Steinberg, 2013). This enhanced plasticity suggests that adolescent science engagement might be better understood as a dynamic learning process rather than a stable identity formation process (Gopnik et al., 2004).

First, the protracted development of self‐referential processing networks indicates that stable identity formation may be neurobiologically premature during early‐to‐mid adolescence. Educational approaches that rely heavily on fostering a stable ‘science identity’ may be attempting to engage psychological processes for which the neural architecture is not yet fully available. The sophisticated cognitive operations required for mature identity formation—including the weighted integration of personal characteristics according to subjective importance hierarchies—depend on neural systems that remain under development throughout the teenage years.

Second, the earlier maturation and heightened sensitivity of reward and habit‐formation systems suggest that ‘preference consolidation’ represents a more developmentally appropriate mechanism for fostering science engagement during this period. This approach would shift the educational focus from cultivating a declarative ‘identity’ to providing the repeated, positive and skill‐building experiences that allow scientific thinking and practices to become habitual and preferred. The basal ganglia learning mechanisms operate through incremental strengthening of stimulus–response associations rather than requiring sophisticated self‐referential processing, making them more suitable for adolescent populations.

Third, the heightened social sensitivity of the adolescent brain presents both challenges and opportunities. While peer influence can create vulnerability to social pressures that discourage science engagement, this same social sensitivity can be strategically leveraged to support learning through carefully designed collaborative experiences and peer recognition systems that work with, rather than against, adolescent developmental characteristics.

Individual differences in neural development reveal substantial variation in the timing and trajectory of brain maturation, with some individuals showing earlier or later development of key neural networks (Mills et al., 2014). These individual differences may contribute to the heterogeneity observed in adolescent responses to science education and suggest the need for personalized approaches that accommodate diverse developmental trajectories. Furthermore, emerging research on neurodiversity reveals that atypical patterns of brain development may confer advantages for certain types of scientific thinking (Baron‐Cohen, 2009), challenging deficit‐based models of neurodevelopmental variation and highlighting the importance of accommodating diverse pathways to science engagement.

These neuroscientific insights suggest that a new conceptualisation of science engagement is needed—one that honours the dynamic and adaptive nature of the developing mind while providing foundations upon which more sophisticated identity commitments can later be built. Rather than viewing adolescent inconsistency and apparent lack of commitment as problems to solve, developmentally informed approaches recognise these characteristics as normal features of a period characterised by remarkable learning capacity and adaptive potential.

THE DEVELOPMENTAL ENGAGEMENT FRAMEWORK: TWO MODES OF SCIENCE PARTICIPATION

Building on identity process theory (Bosma & Kunnen, 2001, 2008), neuroscientific evidence on value‐based choice (Pfeifer & Berkman, 2018) and recent findings on daily experience consolidation (Branje et al., 2021), we propose the Developmental Engagement Framework (Figure 1). This framework positions science engagement as the overarching construct that encompasses two developmentally sequenced modes of participation: preference‐based engagement (early‐mid adolescence, ages 11–16) and identity‐infused engagement (late adolescence, ages 17+). Science identity thus represents a subset of science engagement—specifically, engagement that has been integrated with self‐concept through mature self‐referential neural networks.

These modes are complementary rather than competing. Preference‐based engagement establishes the foundation for later identity formation, though not all engagement necessarily transitions to identity‐infused participation. This framework resolves apparent empirical inconsistencies in the literature by recognizing that early adolescent science participation can be robust and consequential without requiring stable identity structures.

Science engagement encompasses two developmentally sequenced modes distinguished by their neural foundations and cognitive mechanisms. Not all preference‐based engagement transitions to identity‐infused participation; individual and contextual factors influence this developmental trajectory.

Science engagement as the overarching construct

We conceptualize science engagement as the umbrella construct encompassing all forms of sustained, motivated participation in science‐related activities across the developmental span. This definition is deliberately broad, recognizing that the psychological and neural mechanisms supporting such participation change fundamentally across adolescence. Critically, not all science engagement requires identity formation. Individuals can participate meaningfully and consistently in science without necessarily integrating science into their self‐concept—a distinction that becomes particularly important during early‐to‐mid adolescence when neural systems supporting stable self‐representation remain under development.

Following Pfeifer and Berkman's (2018) value‐based choice perspective, we propose that identity‐related processes represent one form of ‘value input’ to motivated behaviour, but not the only form. During periods when self‐referential processing networks are immature, other value inputs—particularly those derived from immediate experience, positive reinforcement and habit formation—can effectively drive sustained engagement. This perspective aligns with identity process theory (Bosma & Kunnen, 2001, 2008), which demonstrates that identity commitments emerge through accumulated experiences rather than preceding them. As Branje et al. (2021) note, ‘commitment might be consolidated in daily experiences’—suggesting that engagement provides the experiential foundation from which identity can later emerge.

Within this overarching engagement construct, we distinguish two developmentally sequenced modes that predominate at different life stages, supported by distinct neural systems and psychological mechanisms.

Mode 1: Preference‐based engagement (ages 11–16)

Preference consolidation is defined as a neurodevelopmental process through which repeated positive experiences with scientific activities strengthen sustained affinity and engagement through associative learning mechanisms operating in earlier‐maturing neural circuits (basal ganglia, striatum). This mode of engagement does not require stable self‐representation; instead, it operates through experiential learning, positive reinforcement and habit formation—processes that are neurobiologically available during early‐to‐mid adolescence when self‐referential networks remain under development.

The neural basis for preference‐based engagement rests on reward processing systems (ventral striatum, nucleus accumbens) and habit formation circuits (dorsal striatum, basal ganglia) that mature earlier than prefrontal self‐referential networks (Galvan et al., 2006; van Duijvenvoorde et al., 2016). During early adolescence, these systems exhibit heightened sensitivity to rewarding experiences and increased capacity for procedural learning (Davidow et al., 2016). Importantly, these neural systems operate effectively without requiring integration with self‐concept—enabling robust learning and sustained engagement through mechanisms independent of identity.

Consider a 13‐year‐old student who repeatedly engages in hands‐on physics experiments at school. Through accumulated positive experiences—successful demonstrations, peer collaboration, teacher encouragement, intrinsic satisfaction from problem‐solving—the student develops a strengthened preference for physics activities. This preference is neurobiologically encoded through:

1. Reward learning: Positive experiences activate ventral striatal circuits, creating positive associations with physics activities (Galvan, 2010).

2. Habit formation: Repeated engagement strengthens procedural representations in dorsal striatal circuits, making physics participation feel increasingly natural and automatic (Tricomi et al., 2009).

3. Skill development: Accumulated practice builds genuine competence, further reinforcing positive experiences through mastery experiences.

Critically, this student might enthusiastically state ‘I love physics!’ and demonstrate sustained, self‐directed engagement in physics activities, yet may not yet conceptualize physics as part of who they are. When asked ‘Are you a physics person?’ The student might struggle to answer or provide an uncertain response—not because their engagement is superficial or insincere, but because the neural systems supporting stable self‐representation (mPFC, PCC) have not yet matured sufficiently to support identity integration (Pfeifer & Berkman, 2018). This is developmentally normal and does not diminish the authenticity or importance of their engagement.

Preference‐based engagement is characterised by

1. Activity‐centredness: Students are drawn to engaging in science activities rather than explicitly identifying as ‘science people’.

2. Experience dependence: Engagement strength directly reflects accumulated positive experiences.

3. Context sensitivity: Preferences may vary across domains (strong preference for biology but not chemistry) or contexts (enjoys science at summer camp but not in formal school).

4. Dynamic exploration: Preferences remain fluid, allowing active exploration across domains before consolidating more stable patterns.

5. Temporal variability: Day‐to‐day engagement may fluctuate more than identity‐based participation, as it relies on immediate experiential factors rather than a stable self‐concept.

These characteristics, often interpreted as indicating ‘weak’ or ‘unstable’ identity in empirical studies (Perez et al., 2014; Potvin & Hasni, 2014), are instead markers of developmentally appropriate engagement that does not yet require identity structures.

Mode 2: Identity‐infused engagement (ages 17+)

Identity‐infused engagement represents the integration of science participation with self‐concept, enabled by the maturation of self‐referential processing networks. This mode becomes neurobiologically viable during late adolescence (approximately ages 17 and beyond) as the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC) and associated self‐referential networks approach mature functioning patterns (Mills et al., 2014; Pfeifer & Berkman, 2018).

In this mode, science engagement is not merely something the individual enjoys or is skilled at; it becomes part of how they understand and define themselves. The individual can engage in self‐referential processing about their relationship with science (‘I am a science person’), evaluate science‐related activities in relation to self‐concept and integrate science into their broader narrative identity. These processes require neural capacities that are not fully available during early‐to‐mid adolescence.

The transition from preference‐based to identity‐infused engagement is neither automatic nor universal. Not all individuals who develop strong science preferences during early adolescence will subsequently integrate science into their identity. This transition appears to depend on:

1. Neural maturation: The prefrontal self‐referential processing networks must reach sufficient maturity to support stable self‐representation (Pfeifer & Berkman, 2018).

2. Accumulated engagement: Bosma and Kunnen's (2001, 2008) identity process theory suggests that identity commitments emerge through accumulated experiences. Sustained preference‐based engagement during early‐to‐mid adolescence provides the experiential foundation from which identity can potentially emerge.

3. Social recognition: Recognition from others as a ‘science person’ can facilitate integration of science into self‐concept (Carlone & Johnson, 2007) though such recognition becomes psychologically meaningful only when neural systems supporting self‐referential processing are sufficiently mature.

4. Opportunity structures: Continued access to science learning environments, quality instruction and supportive relationships influences whether accumulated preferences crystallize into identity commitments (Branje et al., 2021).

5. Individual agency: As self‐referential processing matures, individuals actively evaluate whether science ‘fits’ with their emerging sense of self, making deliberate decisions about identity commitments.

A critical implication of this framework is that robust, sustained science engagement during early‐to‐mid adolescence does not guarantee later science identity formation. Two primary developmental pathways illustrate this variability. In the first pathway, strong preference‐based engagement during ages 11–16, when sustained through continued positive experiences, becomes integrated with self‐concept as neural maturation enables identity integration, emerging as stable science identity by late adolescence. For example, a student with consistent positive biology experiences throughout middle school develops deep biological expertise and, by age 18, clearly identifies as ‘a biology person’ and pursues university studies in biology. The preference‐based engagement of early adolescence has been transformed into identity‐infused engagement supported by mature self‐referential processing networks.

However, the second pathway demonstrates that individuals may maintain strong preference‐based engagement without ever integrating science into self‐concept. An individual might continue to enjoy amateur astronomy throughout adulthood—subscribing to astronomy magazines, attending public lectures at the local planetarium, maintaining a backyard telescope and experiencing genuine satisfaction from stargazing activities—yet when asked ‘Are you an astronomy person?’ Respond negatively or with uncertainty, instead characterizing astronomy as ‘a hobby I really enjoy’ rather than as integral to self‐concept. This person pursues a non‐science career and does not integrate astronomy into their identity, yet the engagement remains robust and consequential. This pattern does not represent developmental failure or incomplete maturation; rather, it reflects a valid form of sustained engagement that operates through preference mechanisms rather than identity structures, demonstrating that identity integration is not necessary for meaningful, lifelong science participation.

This two‐pathway perspective explains why longitudinal studies find variable relationships between early science engagement and later STEM participation (Maltese & Tai, 2010; Sadler et al., 2012). The presence of preference‐based engagement during early adolescence provides a supportive foundation for potential later identity formation, but this transition depends on multiple factors including continued positive experiences, social recognition, opportunity structures and individual agency. Critically, robust science engagement during early‐to‐mid adolescence should be valued in its own right, not merely as a precursor to identity formation. Both pathways—preference‐to‐identity consolidation and preference maintenance without identity integration—represent authentic and consequential forms of science participation across the lifespan.

Integration with classical developmental theories

The Developmental Engagement Framework builds upon rather than replaces foundational developmental theories. We integrate three classical theoretical perspectives—Piagetian cognitive development, Vygotskian sociocultural theory and Bronfenbrenner's ecological systems theory—which offer complementary lenses for understanding science engagement across adolescence. These theories were selected because they (1) represent enduring frameworks widely applied in science education research, (2) address different but interconnected aspects of development (cognitive, social and contextual) and (3) provide established conceptual vocabulary for connecting neurodevelopmental evidence with educational practice.

Piaget's theory of cognitive development (Piaget, 1972; Inhelder & Piaget, 1958) provides a foundational framework for understanding the changing nature of adolescent reasoning, particularly the transition to formal operational thought characterized by hypothetical‐deductive reasoning and abstract thinking. This framework is particularly relevant for understanding science engagement because scientific reasoning—hypothesis generation, systematic experimentation, abstract model building—relies heavily on formal operational capacities.

The Developmental Engagement Framework extends Piagetian insights by specifying the neural mechanisms underlying the emergence of formal operations. The protracted development of the prefrontal cortex and associated executive function networks (Luna et al., 2015) provides the neurobiological basis for Piaget's observations about the gradual emergence of abstract reasoning during adolescence. Importantly, just as Piaget noted substantial variability in the age at which formal operations emerge (and recognized that not all adults consistently employ formal operational reasoning), our framework acknowledges individual differences in the timing of neural maturation supporting identity formation.

The preference‐based engagement mode aligns well with Piaget's concrete operational stage (approximately ages 7–11) extending into early adolescence. Students at this stage can engage meaningfully with hands‐on science activities, observable phenomena and concrete problem‐solving—all of which can support robust preference consolidation without requiring the formal operational capacity for abstract self‐reflection. As formal operational thought emerges and consolidates in middle‐to‐late adolescence, students become increasingly capable of the abstract reflection about science and self that supports identity formation.

Vygotsky's (1978) sociocultural theory emphasizes that higher‐order psychological functions develop through social interaction and cultural mediation. This perspective is critical for understanding both preference consolidation and identity formation as fundamentally social processes shaped by cultural context, social relationships and mediated learning experiences.

For preference‐based engagement, Vygotsky's concept of the zone of proximal development (ZPD) illuminates how guided participation with more knowledgeable others builds science competence and positive affect. Science educators, parents, peers and mentors create scaffolded experiences within students' ZPD, facilitating skill development and positive science associations that consolidate preferences. Importantly, this social scaffolding operates effectively during early‐to‐mid adolescence without requiring students to develop explicit science identities. The supportive social context nurtures engagement through the social processes Vygotsky described.

For identity‐infused engagement, Vygotskian theory helps explain how identity formation is inherently socio‐culturally constructed. As Carlone and Johnson (2007) demonstrated, recognition from others as a ‘science person’ plays a critical role in identity development—a profoundly Vygotskian insight. However, our neurodevelopmental framework adds that such recognition becomes psychologically meaningful only when neural systems supporting self‐referential processing are sufficiently mature to incorporate others' perceptions into self‐concept. Prior to this neural maturation, social recognition may be experienced positively but cannot yet be fully integrated into stable identity structures.

Bronfenbrenner's (1979, 2005) ecological systems theory conceptualizes development as occurring within nested environmental contexts—microsystem (immediate settings), mesosystem (interconnections between settings), exosystem (external settings that indirectly affect the individual) and macrosystem (cultural values and ideologies). This multi‐level framework is essential for understanding how science engagement develops across diverse contexts and cultures.

Our framework integrates Bronfenbrenner's insights by recognizing that both preference consolidation and identity formation are profoundly shaped by ecological contexts, though these contexts may operate differently at different developmental stages. During early‐to‐mid adolescence, microsystem factors (quality of science classroom, teacher–student relationships, peer group attitudes toward science) and mesosystem connections (alignment between school and family science values) directly influence preference development. The macrosystem—cultural beliefs about who can be a ‘science person’, gender stereotypes about science, societal valuing of STEM—shapes the experiences available to youth and the social messages they receive.

During late adolescence, as identity‐infused engagement becomes viable, these same ecological factors influence whether accumulated preferences crystallize into identity commitments. However, the psychological meaning of these ecological influences changes as self‐referential processing matures. A supportive science teacher may facilitate preference consolidation in a 13‐year‐old through positive reinforcement; the same teacher may facilitate identity formation in an 18‐year‐old by providing recognition and affirming the student's emerging self‐concept as a scientist.

Crucially, Bronfenbrenner's theory helps explain cultural variation in science engagement patterns. Different cultural macrosystems emphasize individual versus collective identity to varying degrees (Markus & Kitayama, 1991), potentially influencing when and whether identity‐infused engagement becomes salient. In more collectivist cultural contexts, science participation might remain primarily preference‐based or role‐based throughout adolescence, with identity integration occurring later or taking different forms than in individualistic Western contexts. This cultural sensitivity is critical for avoiding inappropriately universalizing Western developmental timelines.

These three classical theories—Piagetian, Vygotskian and Bronfenbrenerian—address complementary facets of science engagement development: cognitive readiness (Piaget), social mediation (Vygotsky) and contextual embeddedness (Bronfenbrenner). The Developmental Engagement Framework integrates these established perspectives with contemporary neuroscientific evidence, specifying the neural mechanisms underlying observed developmental patterns while honoring the social and contextual insights of these foundational theories. Table 1 summarizes the developmental characteristics distinguishing preference‐based and identity‐infused engagement modes.

TABLE 1.

Developmental characteristics of two modes of science engagement.

Characteristic	Preference‐based engagement (early‐mid adolescence)	Identity‐infused engagement (late adolescence)
Developmental period	11–16 years	17+ years
Primary neural systems	Basal ganglia Striatal circuits Reward processing systems	Medial prefrontal cortex (mPFC) Posterior cingulate cortex (PCC) Self‐referential networks
Engagement mode	Experience‐based Activity‐driven Reward‐sensitive	Reflection‐based Self‐evaluative Meaning‐making
Key mechanisms	Associative learning Habit formation Preference consolidation	Self‐referential processing Narrative integration Value‐based choice
Participation characterization	‘I enjoy doing science activities’ ‘Science is fun’	‘I am a science person’ ‘Science is part of who I am’
Measurement approach	Behavioural engagement indices Interest scales Activity participation records	Self‐concept measures Identity scales Recognition‐based items
Temporal stability	Moderate (dynamic, exploration phase)	Higher (stable commitments)
Predictive validity for future outcomes	Predicts continued exploration and skill development	Predicts career choices and long‐term persistence

The framework's integration with established developmental theories demonstrates that neuroscientific evidence does not replace but rather complements and enriches existing educational theory. By specifying the neural prerequisites for different forms of engagement, we can better understand when and under what conditions different theoretical insights apply most powerfully across adolescent development.

Implications for understanding individual and cultural variation

The Developmental Engagement Framework has important implications for understanding substantial individual and cultural variation in adolescent science engagement patterns. Rather than treating this variation as problematic heterogeneity, the developmental perspective suggests that variation reflects normal differences in developmental timing, cultural values and neural architecture that should be accommodated rather than eliminated.

Neuroimaging research demonstrates considerable individual variation in the timing of brain maturation, particularly for association cortices including prefrontal self‐referential networks (Gogtay et al., 2004; Tamnes et al., 2017). This neural variability translates directly into variation in when individuals become neurobiologically capable of stable identity formation. Some adolescents may develop mature self‐referential processing by age 16, enabling earlier identity integration; others may not reach this developmental milestone until their early 20s. Both patterns represent normal variation rather than developmental delay or acceleration.

This individual variation has several implications:

1. Age is an imperfect proxy for developmental readiness. Chronological age provides useful general guidance about typical patterns, but individual assessment of developmental status may be necessary for precision.

2. Early identity formation may not indicate advanced development. An apparent ‘science identity’ expressed by a 13‐year‐old might reflect social learning of identity language rather than genuine integration of science with self‐concept supported by mature neural architecture.

3. Later identity formation does not indicate deficiency. Students who maintain preference‐based engagement through age 18 or beyond are not ‘behind’; they may simply be following a normal developmental trajectory characterized by later neural maturation.

Cross‐cultural research documents substantial variation in how identity is conceptualized, experienced and expressed across cultural contexts (Arnett, 2008; Markus & Kitayama, 1991). Western individualistic cultures tend to emphasize personal identity and individual uniqueness, whereas many non‐Western collectivist cultures emphasize social roles, relationships and group membership. These cultural differences may profoundly influence when, whether and how science becomes integrated with self‐concept.

In cultural contexts where individual identity is highly salient, the transition from preference‐based to identity‐infused engagement may occur readily as neural maturation permits. In contexts where collective identity or social roles predominate, science participation may be organized around family expectations, community service or social role fulfilment rather than personal identity—even among adults with fully mature prefrontal cortexes. This does not indicate developmental immaturity but rather reflects culturally shaped pathways to sustained science engagement.

Educational implications include

1. Avoiding cultural bias in assessment: Identity‐based measures may underestimate science engagement in cultural contexts where personal identity is less salient.

2. Respecting diverse pathways: Multiple developmental routes to sustained science participation should be recognized as equally valid.

3. Cultural adaptation of framework: The two‐mode framework should be applied flexibly, recognizing that the relative importance of each mode and the typical timing of transitions may vary across cultural contexts.

Gender differences in science identity have been extensively documented, with girls and women often reporting weaker science identities despite equivalent or superior academic performance (Gonsalves & Danielsson, 2020; Hazari et al., 2010), with these patterns as emerging as kindergarten (Patrick et al., 2009). Our developmental framework suggests an additional consideration: stereotype threat and social recognition biases may differentially affect the transition from preference‐based to identity‐infused engagement for girls.

Even when girls develop strong science preferences during early‐to‐mid adolescence through positive experiences, the transition to identity integration during late adolescence may be impeded by

1. Lack of recognition: Teachers, parents and peers may be less likely to recognize and affirm girls as ‘science people’ (Carlone & Johnson, 2007), depriving them of social feedback necessary for identity consolidation.

2. Stereotype activation: As self‐referential processing matures, girls become more cognitively capable of internalizing cultural stereotypes that ‘girls aren't good at science’, potentially inhibiting identity integration even when strong preferences and competence exist.

3. Role model scarcity: Limited representation of women scientists in media, textbooks and local environments may make it psychologically difficult for girls to envision science as part of their future selves.

Critically, these barriers operate at the level of identity formation while potentially leaving preference‐based engagement intact. A girl may maintain strong interest in and aptitude for science throughout adolescence yet fail to develop a science identity due to social‐contextual factors that impede the preference‐to‐identity transition. Interventions targeting recognition, stereotype reduction and role model exposure may be particularly critical during the late adolescent period when identity integration becomes neurobiologically viable.

Table 2 presents educational and assessment implications of these two modes of engagement. The framework suggests that educational approaches should be developmentally calibrated to match the predominant mode at each life stage. During early‐to‐mid adolescence (ages 11–16), when preference‐based engagement predominates, educators can effectively foster science participation through hands‐on experiences, skill development and positive reinforcement that capitalize on reward‐sensitive neural systems (basal ganglia, striatal circuits). Experiential learning, peer collaboration, incremental challenges and diverse activity exposure align with the neural foundations of preference consolidation. During late adolescence (ages 17+), when identity‐infused engagement becomes viable, educational contexts can appropriately emphasize identity exploration, career planning and self‐reflective practices that leverage mature self‐referential processing networks. Mentorship, role models and identity‐affirming recognition become psychologically meaningful as students develop the capacity to integrate science into self‐concept.

TABLE 2.

Educational and assessment implications of two engagement modes.

Aspect	Preference‐based mode	Identity‐infused mode
Educational focus	Hands‐on experiences, skill development, positive reinforcement, activity variety, success experiences	Identity exploration, career planning, self‐reflection, connecting science to personal values and goals
Intervention strategies	Experiential learning, peer collaboration, incremental challenges, immediate feedback, diverse activity exposure	Mentorship, role models, identity affirming recognition, science‐self integration activities, career counselling
Assessment methods	Behavioural indicators (participation frequency, task persistence), skill demonstrations, interest inventories	Self‐report identity scales, narrative interviews, future self‐projections, career aspiration measures
Temporal expectations	Expect fluctuation and exploration across domains; do not expect stable commitments	Expect greater stability in commitments; support consolidation of emerging identity
Success indicators	Active participation, positive affect during activities, skill progression, continued voluntary engagement	Self‐identification as ‘science person’, integration of science into future plans, stable commitment despite obstacles
Risk factors	Accumulated negative experiences, lack of competence‐building opportunities, absence of social support	Lack of recognition, stereotype threat activation, absence of role models, misalignment between science and emerging self‐concept

Note: Educational approaches and assessment methods should be developmentally calibrated to match the predominant engagement mode at each life stage.

Assessment methods should similarly match the developmental mode. For preference‐based engagement, behavioural indicators (participation frequency, task persistence), skill demonstrations and interest inventories provide valid measures of science engagement without requiring stable self‐representation. For identity‐infused engagement, self‐report identity scales, narrative interviews and future self‐projections become appropriate as neural systems supporting stable self‐concept mature. Attempting to measure identity during early‐to‐mid adolescence may yield unreliable results not because students lack genuine science engagement, but because the neural architecture supporting identity integration remains under development. Recognition of these developmental differences enables age‐appropriate support for science participation across adolescence.

CONCLUSIONS

This analysis began with the observation that science identity frameworks, despite valuable contributions to understanding equity and social dimensions of science learning, face persistent empirical challenges when applied to adolescent populations (Perez et al., 2014; Potvin & Hasni, 2014). The temporal instability and modest predictive validity consistently observed in adolescent identity measures have often been treated as methodological limitations. However, examination through developmental cognitive neuroscience reveals these patterns as reflections of normal neurodevelopmental processes rather than measurement failures.

The convergent evidence demonstrates that the neural architecture required for stable self‐representation undergoes protracted development extending into the third decade of life (Gogtay et al., 2004). Self‐referential processing networks that support mature identity formation remain immature throughout adolescence (Martinelli et al., 2013), while reward processing and habit formation systems mature much earlier (Yin & Knowlton, 2006). This developmental mismatch provides a compelling neurobiological explanation for why adult‐derived identity constructs may be inappropriate for younger populations and why preference consolidation may represent a more developmentally suitable alternative.

The proposed Developmental Engagement Framework makes several theoretical contributions. First, it offers a neurobiologically grounded explanation for empirical puzzles that have challenged science identity research, reframing apparent methodological problems as meaningful developmental phenomena. Second, it bridges established developmental theories and contemporary neuroscientific findings, demonstrating how classic insights from Erikson (1968), Piaget and Vygotsky can be enhanced through integration with brain research. Third, it challenges deficit‐based models that pathologize adolescent inconsistency (Harter, 2012), instead recognizing developmental characteristics as normal features of a period characterized by heightened learning capacity.

The developmental continuum model recognizes preference consolidation and identity formation as complementary rather than competing processes, operating optimally at different developmental stages. Early adolescence may be better served by approaches emphasizing repeated positive experiences and skill development that leverage heightened neuroplasticity and reward sensitivity (Steinberg, 2013). Identity‐focused approaches become more appropriate during late adolescence when neurobiological prerequisites for stable self‐concept formation are established.

The framework's emphasis on individual and cultural variation represents another important contribution. Rather than seeking universal solutions, this perspective recognizes that developmental trajectories and cultural values vary substantially across populations (Mills et al., 2014; van Buuren et al., 2022). Neurobiological evidence for cultural differences in self‐referential processing (Han & Ma, 2014; van der Cruijsen et al., 2023) provides compelling support for approaches that accommodate rather than standardize these differences, while acknowledging that individual differences in neurodevelopmental timing create natural variation in when different students may be ready for different types of science engagement experiences (Baron‐Cohen, 2009).

Several limitations must be acknowledged. Translation from neuroscientific findings to developmental theory involves considerable uncertainty (Bruer, 1997; Howard‐Jones, 2014), and most brain research has been conducted with relatively homogeneous populations in laboratory settings. Longitudinal studies tracking both neural development and science engagement remain rare (Ansari & Coch, 2006), limiting understanding of how these processes unfold in real‐world contexts.

These limitations point towards important directions for future research. Longitudinal studies examining relationships between brain development and science engagement could provide crucial evidence about optimal timing for different educational experiences. Cross‐cultural investigations could illuminate how neurodevelopmental patterns vary across contexts. Research on individual differences in developmental timing and trajectory could reveal diverse pathways to science engagement that current frameworks may overlook.

The ultimate goal is advancing theoretical understanding that better serves young people developing relationships with science. By recognizing adolescence as a distinct developmental period with unique characteristics and capabilities, we can develop approaches that work with rather than against natural developmental processes (Gopnik et al., 2004). This perspective honors both the remarkable potential and distinctive realities of the developing mind, providing foundations for more inclusive support of diverse pathways to science engagement.

AUTHOR CONTRIBUTIONS

Mihyun Son: Conceptualization; writing – review and editing; validation; investigation; formal analysis; supervision; writing – original draft; visualization. Hunkoog Jho: Conceptualization; investigation; funding acquisition; writing – original draft; methodology; visualization; formal analysis; project administration; data curation; supervision; resources; software.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.