The Challenge of Reflecting ‘Social-Contextual’ Childhood Disability in Quantitative Research: A Narrative Review of the Literature

Introduction

Disability has been described as ‘the quintessential post-modern concept, because it is so complex, so variable, so contingent, so situated’ (Shakespeare and Watson 2001, 19). Such is its dynamic and disputed nature that drafters of the landmark United Nations Convention on the Rights of Persons with Disabilities (CRPD) (2006) were ultimately unable to agree on a definition of disability, opting instead for an ‘open-ended description’ in the final document (Series 2019, 81). This reticence reflects concern around the power of definitions, models and conceptualisations: that ‘how we understand disability transforms how we respond to it, that questions of epistemology and ontology ultimately influence the material conditions and relationships of power that shape lives’ (Series 2019, 77). For almost 50 years, social science has debated the meaning of disability. Yet, despite the attention it has received, disability ‘remains an enigma’ (Watson and Vehmas 2020, 4).

These complexities pose a particular problem for quantitative research, which must not only navigate between varied and often contradictory conceptualisations of disability but must also operationalise them for measurement. It is perhaps unsurprising then that social science research has tended to prefer qualitative approaches, which are inductive rather than deductive, accommodate complexity, and are sensitive to the social context in which data is produced. For childhood disability research in particular, a key benefit is that such methods are responsive to children’s preferred means of communication, enabling children to participate in research as ‘authors of their own stories’ (McLaughlin, Coleman-Fountain and Clavering 2016, 35). Qualitative approaches allow for many creative methods that enable children’s voices, particularly marginalised voices, to be heard (Merrick et al. 2019; Stafford 2017).

Despite these and many other advantages for qualitative methods, our article focuses on the quantitative rather than qualitative study of childhood disability. This is not to promote any single methodology over another or to overlook the many shortcomings of traditional quantitative childhood disability research. However, we take the position that thoughtfully designed quantitative research has a contribution to make to addressing the inequalities of disability (Besio et al. 2022; Blackburn, Read and Spencer 2007), since ‘when disabled children are missing from the statistics, there is every likelihood they are being forgotten in policymaking processes’ (Beckett and Callus 2023, 672). Indeed, signatories to the CRPD are obliged, under Article 31, to collect appropriate statistical data. We must therefore find workable solutions to the challenges of disability definition and measurement.

There are numerous difficulties with current quantitative approaches, and a goal of this article is to raise awareness among disability researchers and users of disability data about the limitations of these traditional strategies. Limitations are both practical (for example, they yield inconsistent data) and theoretical (they reflect an outdated model of disability). In the following article, we explore these limitations and their implications for childhood disability research.

Limitations of traditional approaches to disability measurement

A frequently cited challenge with traditional approaches to quantitative disability measurement is the inconsistent estimations of disability prevalence that they generate across a given population. For example, exploring data from the same representative sample of the Norwegian population, Molden and Tøssebro found disability prevalence rates of between 9.7 and 27.8%, depending on which of eleven measures they used to operationalise disability (Molden and Tøssebro 2010). Altman used five indicators from the same iteration of a US population survey and found disability prevalence to range from 1.4% to 29.7% (Altman 2001). That these inconsistencies are a result of the many different approaches taken to disability measurement has been ‘shown, confirmed and replicated’ (Grönvik 2009, 2). However, the differential estimates of prevalence that alternative measures yield are, as Molden and Tøssebro (2012) point out, only the most conspicuous of several challenges that arise from the poorly understood connections between different disability measures and the alternative groups they define.

A number of typologies of disability measures are offered in the literature. For example, Fujiura and Rutkowski-Kmitta (2001) distinguish between illness, or impairment-based schemes; limited-function, or restricted life activity schemes; and ecological perspectives, which attempt to capture the person-environment interaction. Altman (2001) proposes a similar typology that includes both biophysiological and social role/behaviour measures, but she also includes administrative/legal measures that, in her own research, derive from information about the individual’s receipt of social security or pension on account of disability. Ravaud, Letourmy and Ville add to this list a simple self-attribution measure of disability (‘regardless of its nature’) by the respondent themselves (Ravaud, Letourmy and Ville 2002, 534). With some notable exceptions, the exact nature of the relationship between these different operationalisations of disability in survey research and the groups that they define—that is, whether there exist any consistent patterns in the demographic profile of groups identified by alternative measures—has rarely been explored (Altman 2001; Molden and Tøssebro 2010; Pettinicchio and Maroto 2021; Platt 2018; Putz and Glickman 2019; Ravaud, Letourmy and Ville 2002; Weeks et al. 2021). This is regrettable, since Molden and Tøssebro (2010) note the diverging demographics of groups classed as ‘disabled’ under alternative measures or studied in different contexts, even when these are drawn from the same population. For example, Altman (2001) found considerable differences in the gender and ‘race’ of respondents, depending on the measure used. In particular, she found that men were comparatively over-represented in the group of people defined as disabled under her administrative definition, a finding echoed by Grönvik (2009). Even similar ‘types’ of disability measures are sensitive to specific question wordings and response categories, as Weeks et al. (2021) found when comparing two ‘limited-function’ approaches in the same US survey. Here, the group defined as disabled by a question set with four ‘severity’ categories was more likely to include older and more ethnically diverse people than a similar set with binary ‘yes/no’ responses. Ravaud, Letourmy and Ville (2002) also found age-related differences in the likelihood of self-reporting a disability, with older people more likely to be in a group of people who reported a restriction in their activities but who did not also self-report a disability.

While there is often an assumption that these different groups are ‘best viewed as overlapping populations, with considerable but not total communality’ (Fujiura and Rutkowski-Kmitta 2001, 69), several studies have found that the anticipated overlap between groups of people defined as disabled under differing measurement strategies is often much smaller than expected (Altman 2001; Amilon et al. 2021; Molden and Tøssebro 2012; Platt 2018). These and other challenges have consequences for ease of comparison across studies and, ultimately, for ‘building a cumulative research basis’ (Molden and Tøssebro 2012, 341).

Much of this is older research, coinciding with a fertile period for the examination of disability measurement in the literature, stretching from the publication of the International Classification of Functioning, Disability and Health (ICF) in 2001 to the adoption and implementation worldwide of the Washington Group’s Short and Extended Set of disability questions (World Health Organization 2001). Our literature search identified few more recent studies examining these important questions in such detail.

While, as we have discussed, approaches to operationalising disability for measurement are varied, the dominant conceptualisation of disability underpinning many of them is the same, deriving broadly from an individual’s personal characteristics, implicitly referencing the so-called ‘medical model’, although it is rarely framed as such. In this model, disability is understood simplistically as an individual physical or mental deficit, indistinguishable from impairment. For example, in the UK, Barnes and Mercer (2010, 32) highlighted the ‘near total reliance on a functional limitations approach to disability in the collection of official statistics and governmental research’ (a functional limitations approach considers disability as a restriction or limitation in a named activity, such as walking a given distance or hearing others talking in a given setting). More recently, the UK’s 2021 census asked if participants had a long-term physical or mental health condition or illness lasting or expected to last twelve months or more and whether it limited their day-to-day activities ‘a little’, ‘a lot’ or ‘not at all’. The assumption that it is the condition alone—not any environmental factor—that imposes the limitation is what relates this survey question to the medical model (Office for National Statistics 2024).

Equally, in their reviews of data sources on childhood disability both in the UK and internationally, Read, Blackburn and Spencer (2010) and Cappa, Petrowski and Njelesani (2015) independently conclude that most draw on questions formulated to reflect a medical model in order to capture disability. In both these studies, the researchers found that impairment-based approaches—those that focus on disability as an illness or health-related condition that requires a remedy—were the most common. Even in nationally representative birth cohort studies, considered a particularly valuable source of information about children’s development (Canova and Cantarutti 2020) and often with an interdisciplinary interest in the multiple domains of children’s lives (Counihan et al. 2023), medical model-informed questions about disability persist. For example, in 2017/2018, the Growing Up in Ireland study asked primary caregivers whether their child had ‘any on-going chronic physical or mental health problem, illness or disability’ and if they were ‘hampered in their daily activities’ by this ‘problem’. This narrow emphasis on impairment as it relates to a supposed ‘norm’ in the measurement of disability is particularly problematic for research into disabled childhoods, given the wide differentials in the nature and pace of children’s physical and mental change across ages, cohorts and time.

The endurance of medical model-informed disability measures in much disability research disregards many decades of scholarship in the social sciences, as well as recent developments in international human rights frameworks, which are critical of traditional conceptualisations of disability and which instead understand it as a ‘social-contextual’ phenomenon (Shakespeare 2006, 9). Although social science continues to debate the role of impairment in the production of disability, the social-contextual understanding of disability has been helpfully characterised as a ‘family of social contextual approaches’ (Shakespeare 2006, 2), which, despite their differences, all accept the ‘revolutionary kernel of truth’ at the heart of the social model (Bickenbach 2020, 67): that disability is ‘a social, relational, political and cultural entity that is lived through as a very personal experience though shaped by some very public encounters’ (Goodley 2017, xi). Callus and Beckett refer to a ‘broad church of ideas found in disability studies regarding how best to interpret disability’, the differences between which ‘are much less important than their similarities’ (Callus and Beckett 2023, 7).

The identification of these broadly opposing conceptualisations of disability is now approximately fifty years old, and they have been explored and variously challenged and defended in the succeeding decades. ‘Social-contextual’ disability has been taken forward in the literature, and many alternative models within Shakespeare’s ‘family’ of approaches proposed. These, like the original ‘British social model’ itself, are historically, geographically and otherwise situated, reflecting the contexts within which they were developed. Amongst these ‘second-wave social model critics’ (Goodley 2017) are Swain and French, whose ‘affirmation model’ emphasised ‘positive social identities, both individual and collective, for disabled people grounded in the benefits of lifestyle and life experience of being impaired and disabled’ (Swain and French 2000, 569). Explicitly linked by these authors to the Disability Arts Movement, but with wider applications, including in sport, this model brings impairment firmly back into the frame as the basis for a positive disability identity (Kavanagh 2012; McCormack and Collins 2012). Also seeking to re-establish a place for impairment in disability was Carol Thomas, whose ‘social relational model’ acknowledged that activity restrictions may arise directly from impairment, but who ‘understands disability to be those restrictions of activity that result from the exercise of the power to exclude’ (Thomas 2004, 29).

More recently, the international United Nations (UN) system, especially the work of the World Health Organization (WHO), provides another context through which recent developments in defining disability can be read. The UN’s focus on disability has taken place ‘within a paradigm of rights that has been emerging since the United Nations Universal Declaration of Human Rights of 1948’ (Rioux and Carbert 2003), culminating in the publication of the CRPD in 2006 and the emergence of the so-called ‘human rights model of disability’. In contrast, the WHO’s work in this area has historically placed disability within the context of health and medicine. Its publication in 2001 of the ICF conceives of disability as a ‘biopsychosocial’ phenomenon, presented in the tool as ‘integrating’ the ‘opposing’ medical and social models (World Health Organization 2001).

These and other conceptualisations of disability offer scope to reconsider how disability might be defined and measured for quantitative research. Yet there is a persistent use of outdated measures across numerous disciplines (demography, education, public administration, international development, epidemiology, rehabilitation, health, etc.), which suggests a lack of engagement among some researchers with these contemporary ideas. However, proponents of ‘social-contextual’ disability are in turn criticised for a similar lack of engagement with their colleagues in other disciplines, due partly to their fundamentally different epistemological perspective (Bickenbach 2020; Grönvik 2007). This disconnect is not sustainable if quantitative disability research, and its associated data, are to meet the ambitious goals now being asked of them. Since how we measure disability in quantitative research both reflects and informs how it is understood (Bennani 2023), and because the increased availability of large-scale population data and the capacity to process it present an opportunity to do so, then there is an urgent need for disability measurement of all kinds to engage with contemporary social science understandings of disability, and vice versa.

In this article, we are interested in how a discussion of theory can inform measurement. Our interest is in how childhood disability is studied quantitatively, although this examination needs to be situated within a broader discussion of disability measurement and childhood disability research. Acknowledging that medicalised disability measures represent a well-established approach across many disciplines and that finding workable alternatives reflecting ‘social-contextual’ disability is challenging, in this paper we seek nevertheless to identify some ways forward. To do so, we consider the quantitative research context within which disability measures are put to use, exploring how our assumptions drive the study designs we use, which in turn influence the demands made of disability measures in research.

Scope and Methodology

To address our subject, we have chosen a narrative approach, which allows researchers to ‘describe what is known on a topic while conducting a subjective examination and critique of an entire body of literature’ (Sukhera 2022). This type of approach is most suited to problems that require ‘clarification and insight’ rather than to those that require data, which are more suited to meta-analysis or systematic review (Greenhalgh et al. 2018). While ours is not a systematic review, nevertheless, responding to Turnbull et al. (2023) in this section, we have sought to be transparent about our general process, within the bounds of our selected narrative mode.

Drawing on existing literature to inform the challenge of operationalising ‘social-contextual’ childhood disability for quantitative research is challenging, due in particular to the scarcity of work that specifically examines disability measurement as it relates to childhood disability theory. Our review has therefore required the integration of several strands of literature across multiple disciplines, including disability studies, children’s studies and quantitative measurement. We have particularly focused on methodological, theoretical or commentary articles or book chapters (rather than primary studies) in our selected domains (much already collected to inform a previous publication) (Swift et al. 2021). Although we hoped to avoid referencing literature situated within the disciplines of rehabilitation, healthcare or special education, due to the scarcity of relevant papers, we have been obliged to include these in our discussion, where such literature engages directly with the question of disability measurement as it relates to models or theory.

Also, due to the scarcity and older age of the most relevant literature, two limitations in our selection are the focus largely on the global North, in particular the UK and Scandinavia, and that some references are over twenty years old.

The review begins with an attempt to understand why the mutual disconnect between the social science perspective on disability and much quantitative research has been particularly evident within childhood disability research. It points to the new focus of the last decade on childhood disability from social science but suggests that disability measurement in quantitative research has remained distant from or even unaware of these developments. The importance of clarity around the conceptualisation of disability underpinning any disability measure is noted, and the particularly negative impact on children of the persistence of the medical model is highlighted. Alternative theoretical frameworks and preliminary approaches, which have been built on children’s own reports of disability, are examined for their potential to inform measurement. Finally, several wider issues relating to the quantitative study of childhood disability are considered important context for future recommendations around disability measurement for this group.

We have generally used the term ‘disabled children’ throughout the discussion to reflect our understanding of disability as a social phenomenon.

Childhood Research and Disability Research – Exploring Synergies

We previously identified a disconnect between the design of disability measures and contemporary social science perspectives on disability. There are additional obstacles to mutual engagement in children’s disability research. These include the reported absence of theory specific to the lives of disabled children on which measurement might draw (Stalker 2012; Watson 2012) and a particular focus in contemporary social science scholarship on qualitative approaches to childhood research, deemed more appropriate to foregrounding the child’s own voice and therefore more in line with a key principle of contemporary childhood studies (Tisdall 2012; Richards and Clark 2018). This marked preference for qualitative methods in childhood disability research may be a key reason that quantitative approaches have become largely the preserve of ‘applied’ disciplines such as healthcare or education and remain untroubled by the ideas that had such a far-reaching impact elsewhere. In short, a disciplinary divide has been matched by a methodological divide, thus enabling and reinforcing the disconnect.

Another factor contributing to the limited influence of contemporary understandings of disability on quantitative childhood disability research may be the historical lack of interest in childhood disability within social science itself. Disabled children are said to have been ‘twice-ignored’, being ‘routinely excluded from the category “child” by mainstream theory’ and yet also forgotten by disability theory, which has not taken their specific experience into account (Munn 1997, 484). In the last two decades, however, the ‘new’ sociology of childhood—whereby there has been a reconsideration of children, no longer viewed as submissive or simply as the property of the adults around them, but rather as social actors with rights (Mallett and Runswick-Cole 2014)—seems to have opened up a useful exchange between childhood and disability research. There are parallels between the two disciplines:

Both are driven by a desire to turn those they work with, children and disabled people, from the objects of study to the subject and to present them as active agents, and both approaches are also firmly wedded to the concepts of rights and participation (Watson 2012, 192).

This has been characterised as a mutually beneficial engagement, which extends the understanding of both childhood and of disability in those fields (Kelly 2005).

Early evidence for the increased attention being paid to disabled childhoods was seen, for example, in the significant 1998 Economic and Social Research Council-funded Life as a Disabled Child qualitative research project (funded under the Children 5–16: Growing into the 21st Century programme, grant L129251027). The last decade has also seen the publication of several edited collections specifically focused on childhood disability (for example, Beckett and Callus 2023; Boggis 2018; Callus and Beckett 2024; Curran and Runswick-Cole 2013; McLaughlin, Coleman-Fountain and Clavering 2016; Runswick-Cole, Curran and Liddiard 2017; Traustadóttir et al. 2015). Equally, at a similar point, the broader study of childhood was finally alerted to the question of disabled childhoods, with, for example, the publication of a special issue of Children and Society (Abbott et al. 2012) dedicated solely to disabled children and young people. Latterly, the burgeoning field of disabled childhood studies, first proposed as a distinctive discipline at the start of this productive last decade or so (Curran and Runswick-Cole 2014) and explicitly committed to bringing together disability studies and childhood studies (Goodley, Runswick-Cole and Liddiard 2020), has had considerable impact in research.

Disability Models in Childhood Research

While social science scholarship has finally turned its attention to reframing childhood disability, the field of disability measurement continues to be largely oblivious to the wide range of themes and contexts by and through which childhood disability is being considered elsewhere, and in measurement the medical model continues to have considerable staying power. Some see the model as particularly pernicious for children, for whom an additional dimension of ‘normalcy’ comes into play, that of time. Writers in the two disciplines of disability studies and childhood studies associate these ideas of what is ‘normal’ at a particular stage of a child’s life— developmental ‘milestones’—with the discipline of developmental psychology. They share a view of this approach as a particularly influential and negative manifestation of the medical model in the lives of disabled children. Developmental psychology’s focus on what it considers as normal development is seen, by these disciplines, as doubling down, for disabled children, on the concept of ‘normalcy’, which is strongly rejected in disability studies (Davis 2016). When developmental psychology adds in the dimension of time, the ‘normal’ child against which disabled children are assessed becomes a linear progression through a pattern of various stages or developmental childhood ‘milestones’ that is supposedly ‘predictable’ and ‘universal’ (Mallett and Runswick-Cole 2014, 41) but that is instead ‘narrow and individualistic’ (Traustadóttir et al. 2015, 17). In other words, children are compared to not just one standard but several, each attached to a particular age group. Children who don’t fit one or more of these models are ‘othered’ and judged as being below the normal category (Mallett and Runswick-Cole 2014). Consequently, ‘the idea that developmental psychology stands in implacable opposition to Childhood Studies has become something of an orthodoxy’ (Tatlow-Golden and Montgomery 2021, 3). By extension, this opposition has also been adopted in the disability studies literature where it relates to children. Amongst the features of developmental psychology said to reinforce a narrow ideal of ‘normal’ childhood and thereby to have negative consequences for all children, in particular disabled children, are ‘clinical diagnostic criteria such as standardised measures of intelligence or self-concept, which are equated with the medical or clinical approach’ (Llewellyn and Hogan 2000, 159) and whose ‘norms’ are themselves based largely on samples which exclude disabled children (Cheong and Johnston 2013; Llewellyn and Chung 1997); an ‘ethnocentric’ approach which is based on Western assumptions and which overlooks ‘the different capacities, competences, interests and developmental trajectories of poorer and less privileged children’ (Tatlow-Golden and Montgomery 2021, 3); and the perception of children as ‘becomings’ on their way to adulthood rather than ‘beings’ with agency and their own rights (Uprichard 2008; Traustadóttir et al. 2015). Bloch (2000, 258) points out that, from as early as the 19th century, ‘child development knowledge’ and its associated research have ‘been used as part of the normalising of subjectivities, helping to identify who and what was normal, who and what was abnormal and required intervention’. The impact of these assumptions has been far-reaching for all children, but particularly for those with disabilities. However, as Tatlow-Golden and Montgomery (2021, 7) suggest, there are developmental psychologists, such as Bronfenbrenner, whose work is open to a contextualised view of children’s lives and that ‘can engage in fruitful dialogue with Childhood Studies’. Yet ‘mainstream developmental studies as currently practiced are indeed still often myopic to meaningful understandings of context’ (Tatlow-Golden and Montgomery (2021, 7).

The especially negative impact of the medical model on the lives of disabled children (Underwood et al. 2020) renders the outdated disability measures that are so often used to explore their lives particularly troubling. Even very recent, thoroughly considered, and officially sanctioned measures, such as the Washington Group on Disability Statistics Child Functioning Module (WGCFM),¹ continue to focus on limited functioning, viewed as an ‘intervening mechanism’ between impairment and participation (Madans et al. 2004, 5). This language is reminiscent of the ICF, on which the WGCFM is explicitly based (World Health Organization 2001). The tool is highly influential in what Bennani has described as ‘the broader project of quantification’ (Bennani 2023, 7). While the ICF is claimed to be ‘the only fully worked out, culturally and linguistically tested, and evidence-based classification of functioning and disability’ (Bickenbach 2020, 57), there remain those who question its principles and approach. Chief among the criticisms is that, despite the document’s assertions to the contrary, its underlying conceptual framework remains that of the medical model (Barnes and Mercer 2010; Hammell 2004; Michailakis 2003; Mitra and Shakespeare 2019; Söder 2009). Although the tool has an expressed ambition to apply ‘in many scientific, clinical, administrative and social policy contexts’ (World Health Organization 2001, 242), it is in healthcare and rehabilitation that it remains most widely used (Cerniauskaite et al. 2011; Maribo et al. 2016). The lack of classification for the domain of ‘Personal factors’ and the absence of any distinction between the domains of ‘Activity’ and ‘Participation’ (the authors apparently found there to be ‘no conceptually robust way of distinguishing’ between the two (Bickenbach 2020, 63)) are seen as illustrating unhelpful ambiguity in the tool (Shakespeare 2014). In contrast, the domains of body functions and structures are described in considerable detail. The ICF has been an important reference point in the realm of disability measurement for children through its role as the basis of the widely disseminated WGCFM. While it offers a more holistic view of disability on which measurement could be based than its earlier iterations, in practice the tendency remains, even in supposedly ICF-informed measurement approaches, to focus on impairment.

In light of the recent engagement between childhood studies and disability studies and the long-acknowledged shortcomings of the medical model, why then the continuing adherence to the medical model for quantitative research into childhood disability? On the surface, the model’s focus on impairment may appear simple to operationalise for quantitative research. However, as the inconsistencies in the resultant estimates of prevalence indicate, this supposed simplicity is an illusion. While at first, the associated concepts may seem ‘common to everyday experience’, in the process of operationalisation, ‘layers of nuance are revealed’ (Fujiura and Rutkowski-Kmitta 2001, 70), and, duly, ‘disability breaks down when we start to scrutinise it’ (Goodley 2017, 10).

This complexity poses significant challenges to disability measurement, which, Altman emphasises, should be clear about the ‘definition of disability, the related conceptual component(s) that is/are measured and the nature of the operationalisation of that concept in the course of data collection’ (Altman 2001, 96). Her call for transparency regarding the framework underpinning the disability measure being deployed reflects a growing understanding that there is no ‘right’ or ‘wrong’ measure of disability, only different concepts being measured each time (Madans, Loeb and Altman 2011). Alternative approaches are ‘imperfect proxies that capture only a fraction of the complex reality that is disablement’ (Fujiura and Rutkowski-Kmitta 2001, 92). Researchers seeking to honour the call for measures that reflect contemporary theoretical frameworks of childhood disability are therefore faced with a challenge: finding frameworks appropriate to the study of childhood disability that move on from the medical model and are amenable for use in measurement.

The ’family’ of ‘social-contextual’ models may provide an alternative (Shakespeare 2006, 2). A key challenge with operationalising the social-contextual model for measurement lies with the focus on environmental issues external to the individual. This is a particular difficulty in large-scale population research, where

‘the person is the principle unit of analysis . . and it follows that individual characteristics lie at the core of identification methods. The incorporation of measures capturing interactions with social context or environmental setting is relatively unexplored on a large-scale basis’ (Fujiura and Rutkowski-Kmitta 2001, 85).

Social-contextual disability also rejects the binary ‘able-bodied v. disabled’ conceptualisation of disability. Ideologically, the desire to separate people into one category or another is seen as a manifestation of a ‘societal obsession with finding disability’ (Goodley 2017, xi) and reinforces distinctions ‘between diverse forms of embodiment’ that are ‘at best an expediency, and at worst a violent imposition of epistemic and/or material power’ (Shildrick 2020, 34). This challenges the requirement in many quantitative research designs to separate a group of people that experience disability from a wider population. This ideological challenge is matched by a practical one, as the inconsistent prevalence data yielded by apparently comparable disability measures outlined in the introduction suggests. Bickenbach et al. point out that ‘scientifically speaking, there are no inherent or intrinsic boundaries to the range of variation in human abilities; ability±disability is a continuum, and the complete absence of disability, like the complete absence of ability, is a limiting case of theoretic interest only’ (Bickenbach et al. 1999, 1182).

There have been some early attempts to address these challenges. For example, there is preliminary literature examining ways in which environmental barriers might be operationalised for research (Magasi et al. 2015; Baird et al. 2022), although, as yet, there are few agreed valid and reliable measures in use (Reinhardt et al. 2011). The challenge of generating a binary disability ‘category’ through measurement has also been partially addressed, if not yet overcome, by the WHO’s Model Disability Survey (MDS),² which avoids imposing an a priori and absolute binary classification on individuals by using the captured data in each context to develop a localised scale of 0–100 for capacity and performance (defined as per the ICF), which is then partitioned at the analysis stage ‘using “fit for purpose” cut-offs’ to define the necessary categories (Sabariego et al. 2021, 2). This approach acknowledges the continuum of disability while delivering on the ultimate need, in this context, for prevalence data.

However, there remain questions as to whether the social model, whose basic principles broadly inform most contemporary social science examinations of disabled childhoods, is even appropriate for the study of disabled children. It has been criticised (at least in its ‘strongest’ form) due to its ‘one-dimensional’ analysis, which offers only ‘a relatively small window through which to examine the lives of disabled children’ (Watson 2012, 194). Connors and Stalker criticise the assumption that the social model will apply equally to children as to adults, asserting that the ‘social model of disability has paid little attention to disabled children, with few attempts to explore how far it provides an adequate explanatory framework for their experiences’ (Connors and Stalker 2007, 19). Overall, while the recent engagement between childhood studies and disability studies may have opened up research into childhood disability to a wide array of ideas from both disciplines, there is a claim that the relevance of any of these theories to the lives of disabled children specifically has yet to be fully examined (Carpenter and McConkey 2012). Connors and Stalker argue that few studies ‘have focused specifically on children’s perceptions and experiences of impairment and disability or explored the implications of these for theorizing childhood disability’ (Connors and Stalker 2007, 19).

Bringing in the Child’s Own Perspective

Three studies from our literature search attempt to fill the gap between children and young people’s own reported experience of disability and disability theory (Brady, Franklin and RIP:STARS Collective 2023; Connors and Stalker 2007; Egilson and Traustadóttir 2009). In the first, Connors and Stalker are attracted to the work of Carol Thomas because of its ‘definitions of disability which relate directly to people’s lived experience’ (Connors and Stalker 2007, 20). They found that the conclusions she drew from her work with disabled women, which built on the tradition of the social model, were also relevant to understanding the views expressed by the disabled children that they themselves interviewed. They cite examples of ‘impairment effects’, ‘barriers to doing’, and in particular the importance of perceived ‘barriers to being’.

In the second study, Egilson and Traustadóttir (2009) investigated the adequacy of alternative conceptualisations of disability to reflect the school participation experiences of children with physical impairments in Iceland. While acknowledging the work of Carol Thomas in seeking to make room for subjective experiences of disability, the authors nevertheless found that it was the so-called ‘Nordic relational approach’ that was most consistent with their findings. Their research confirmed ‘the situational and contextual aspect of disability’ and ‘the complexity of the interaction between the individual child and his or her surroundings’, which they feel are key to the Nordic relational model (Egilson and Traustadóttir 2009, 60). They strongly reject the ICF, which they say defines participation as ‘what people do although it may not be what they would prefer to do’ (Egilson and Traustadóttir 2009, 55).

In a third and more recent article, a group of neurodivergent young people engaged in a rare example of co-produced research reflecting on how participants’ own experiences of disability relate to disability theory (Brady and Franklin 2023). They emphasise the ongoing dominance of medical and individualised approaches to disability in their lives, particularly in their interactions with the mental health system. They embrace the social model, which ‘moved us away from problematising and blaming ourselves, and using learnt negative terminology to describe ourselves, (and) empowered us to think differently about the reactions and responses we receive’ (Brady and Franklin 2023, 1386). They also report disabling attitudes such as narratives of pity and vulnerability and reject a world where their impairment becomes their dominant identity. All of these themes are reflected in the literature of disabled childhood studies.

Of these three studies, it is perhaps the conclusions of Connors and Stalker (2007) that offer the most potential for measurement. They identified four ways in which children experience disability: through impairment, difference, other people’s behaviour towards them and material barriers. These four dimensions may offer a means by which children’s own understanding of disability could be operationalised for research. On the basis of Connors and Stalker’s study, impairment is different from that referenced in much quantitative research, in that it relates mainly to what Thomas called ‘impairment effects’—for example, tiring easily or experiencing frequent pain—and not to the physical or intellectual difference from the norm itself (Connors and Stalker 2007). In addition, the researchers point to Thomas’ ‘barriers to being’ in the children’s lives, relating both to the poor management of difference, particularly in the school setting, and also to the belittling words and actions of others.

Much contemporary research into childhood disability also highlights the existence of ableism in children’s lives, for example, in the education system (Goodley 2017), in leisure spaces (Hodge and Runswick-Cole 2013) and in the representation of disability in children’s literature, film, etc. (Cooper 2013; Sigurjónsdóttir 2015). As Hodge and Runswick-Cole point out, ‘disabled children will not always be able to articulate their experience of ableism. They may assume that the fault lies in them because of their impairment rather than in disabling environments’ (Hodge and Runswick-Cole 2013, 322). Connors and Stalker themselves point to the absence of a language with which children can talk about their experiences of disability (Connors and Stalker 2007). However, Brady and Franklin illustrate how young people can identify ableism in their lives, although they counsel that ‘introducing disabled young people to their oppression with an aim of raising critical consciousness needs to be approached carefully and ethically’ (Brady and Franklin 2023, 1389). Others have pointed to children’s general ambivalence or indeed disavowal of a disability identity (Watson et al. 1999; Connors and Stalker 2007), although conversely Brady and Franklin show how young people may embrace this identity. Attempts to probe the presence of ableism in the lives of children should nevertheless be attentive to the potential sensitivities around labels and instead consider the many ways in which ableism and difference might be experienced in a child’s life.

Standard Quantitative Research Procedures and Disability Measurement in Children

This review has so far highlighted the persistence of medical model-informed disability measures in childhood disability research. It has highlighted a lack of specific theory on which childhood disability measures might draw but has also outlined how ‘social contextual’ disability and research exploring children’s own reports of disability might inform an updated approach.

The challenges for quantitative researchers approaching the study of disability in children do not stop, however, with the scarcity of appropriate theoretical frameworks based on those children’s own voices or of measures that reflect them. There are additional complications, common to both qualitative and quantitative research but requiring different mitigations in both. Four of these are discussed in detail in the analysis that follows.

1. The first challenge is the high level of ‘surveillance’ of children by adults in the research process (Watson et al. 1999, 11), with their lives ‘more often than not narrated by adults’ (Callus and Beckett 2024, 2). Change first requires that the child’s own voice be referenced in the development of any conceptual framework and related disability measure, as we have discussed. With the next step—implementing the selected measure in a study—there are both ethical and practical challenges with consulting children directly. Although reference to the child’s own reported narrative is an important principle in childhood research, there is little guidance in the literature on the ethics of asking children directly in surveys about their experience of disability. Generally, there is an acknowledged tension between enabling children to fully participate in research that concerns them while also wishing to protect them. Qualitative, rather than quantitative, methodologies seem to be regarded, at least in social science, as the most appropriate for child participation. Indeed, a recent UNICEF working paper emphasises that certain (unspecified) methodologies can, in themselves, be exclusive and therefore unethical for such children (Thompson, Cannon, and Wickenden 2020). While these researchers do not explicitly advocate for qualitative over quantitative methodologies, their warning against ‘rigid or heavily standardized, prescriptive approaches’ (Thompson, Cannon, and Wickenden 2020, 34) clearly presents particular challenges for quantitative research, since standardisation of the survey question and its administration form key features of this methodology. While large-scale population surveys often compromise on standardising the type of respondent, allowing proxy adults (teachers, parents) to act as respondents on behalf of children (Read, Blackburn and Spencer 2010), they do not traditionally offer the same flexibility in response format, for example, to maximise children’s involvement. However, frameworks and principles for involving children with disabilities call for a variety of data collection tools and for methods of administration that reflect the heterogeneity of disabled children (Jenkin et al. 2020; Thompson, Cannon and Wickenden 2020), suggesting that hearing the voice of all children is of higher priority than the absolute standardisation of the instruments concerned. This perspective could offer the necessary impetus for the development of appropriate self-report measures for disabled children. However, given the many ways in which children do or don’t identify with disability (Beckett and Callus 2024; Priestley, Corker and Watson 1999), ideally surveys would include questions appropriate for children that approach the issue from several perspectives, rather than relying solely on simplistic questions about whether a child considers themselves to be ‘disabled’ or not.

2. A second challenge for research into disabled childhoods is negotiating the ‘fluidity’ of a child’s rapid change over time (Watson 2012, 200), a fact that layers additional longitudinal complexity onto the existing heterogeneity of these children’s lives. There is a general recognition that the mutable and diverse trajectories of a child’s development over time call for longitudinal designs that will allow for the variety of children’s pathways through the lifecourse (Emerson 2012; Stalker 2012; Watson 2012; Ytterhus 2012). Potential assistance for quantitative researchers with meeting this challenge comes from the increasing number of large-scale longitudinal studies of children internationally (Hansen and Joshi 2007; Emerson 2012), although there are limitations in the approaches to disability measurement used in these studies (Cappa, Petrowski and Njelesani 2015; Read, Blackburn and Spencer 2010).

3. Questions of study design also inform the third challenge, which is the need to honour the principle proposed for all research into childhood and disability of conceiving disability as only one of several factors that affect the lives of all children (e.g., gender, ethnicity, etc.) and the importance of considering children with disability as children ‘first and foremost’ (Stalker 2012, 173). Research informed by this principle should research children alongside, rather than separately from, their peers (Priestley 1998; Traustadóttir et al. 2015) and should consider disability alongside other social categories (Toft and Franklin 2020). Although we have already discussed the lack of appropriate theoretical frameworks through which to understand childhood disability, implementing the recommendation that childhood disability research be mainstreamed may in fact call on us to find general theories of childhood that accommodate disability rather than specific childhood disability theory. ‘To view disabled children as “disabled first” is to deny them their rights as children’ (Priestley 1998, 210). Duane has something similar to say about gender, which she argues ‘should remain under the larger rubric of childhood studies’ (Duane 2017) since ‘to separate children out by gender, to draw attention to the workings of gender, runs the risk of imposing the very gendered dualisms on children that scholars seek to undo’. There is a similar danger with our current approaches to studying disability.

   Of the few general theories of childhood that are occasionally applied to disabled childhoods, one that seems to be most commonly used across a range of fields, including in writings on social work (Bricout et al. 2004), psychology (Llewellyn and Hogan 2000), education (Sontag 1996), family research (Ferguson 2012) and disability studies (Fougeyrollas et al. 2019; Traustadóttir et al. 2015), is Bronfenbrenner’s ecological model of human development (Bronfenbrenner 1979). Emanating originally from the discipline of developmental psychology, it has been used in the study of childhood disability, with Traustadóttir et al. (2015) pointing out that socio-pedagogical approaches from this discipline correspond with the Nordic social-interactional approach to disability, which they argue is highly relevant to the study of disabled children. Bronfenbrenner’s bioecological theory of human development offers an opportunity to view disabled children in terms of their status as ‘children first and foremost’. It also incorporates the dimension of time, an aspect of children’s lives that more generalised theories of disability do not accommodate.

4. A final challenge also concerns the way in which traditional approaches to study design seem themselves to have solidified outdated conceptualisations of childhood disability and associated disability measures. For example, research to compare outcomes for disabled children, defined somehow with an often decontextualised ‘disability’ label, with those of their peers requires a binary disability measure on which to define its groups. This is a requirement that classic approaches to measurement (such as those that allow for disaggregation by age or gender) have appeared, on the surface, to be readily able to accommodate. However, as we have discussed, the complex and contingent nature of disability does not easily allow for such a ‘blunt’ approach.

   Perhaps, rather than looking for measures that will deliver on this simplistic requirement, we need to reconsider our research questions and design alternative studies to answer them. Exploratory research approaches that allow groups to emerge from the data, rather than being imposed a priori (an approach that is particularly vulnerable to bias), would seem to be an appropriate first step.

Conclusion and Recommendations

This review has identified a compelling need for the measurement of disability in children to move beyond simplistic measures that reflect the medical model in favour of engagement with the rich contemporary social science literature of childhood disability to examine how it might inform future approaches. The challenges of operationalising ‘social-contextual’ understandings of disability for measurement as an alternative to impairment-based approaches have also been explored, together with some early suggestions as to how this might be accomplished. However, the review has also pointed to questions about the validity of existing models and theories of disability, whose relationship to the specific experiences of disabled children has yet to be fully explored. Three notable exceptions have been examined, together with their potential for informing a disability measure or measures (Brady and Franklin 2023; Connors and Stalker 2007; Egilson and Traustadóttir 2009). Finally, we have explored four additional contextual issues for consideration by those who wish to explore childhood disability through a quantitative lens.

A key recommendation arising from the discussion is for researcher transparency regarding the conceptual frameworks that underpin the disability measures they select. Researchers should be informed by the rich literature around childhood disability emanating from childhood and disability studies when establishing their position. Impairment, or functioning-based approaches to measurement, should be used with extreme caution and only in parallel with survey items that consider the more important features of ‘social-contextual’ disability, such as environmental factors and equal participation. However, research has indicated that children do report ‘impairment effects’, and it is appropriate that these should also be captured in quantitative research (Connors and Stalker 2007). The presence of ableism in children’s lives is a legitimate target for research and could be captured through judicious survey questions. Whatever disability measures are deployed should be contextualised, for example, in school or at home. It is unlikely that one size will fit all.

Several other more general principles of research with disabled children apply to quantitative as much as qualitative research. Researchers should make the direct participation of children as respondents a priority, and the selected disability measures should facilitate this. However, simplistic questions that explore the level at which children do or do not identify as disabled are unlikely, on their own, to yield much meaningful data. The growing literature exploring children’s own perspectives on their disability should inform how such questions are framed and the language used. General research principles applicable to quantitative approaches also include the importance of longitudinal methodologies, which are most appropriate to capture the rapid change over time in children’s lives. Children should also be studied holistically and together, considering disability alongside other social categories such as gender, ethnicity and social class. Conceptual frameworks that accommodate disability as one of several factors influencing their lives, such as Bronfenbrenner’s bioecological theory of human development, may be more appropriate than specific models of disability in children.

Overall, this review has concluded that research questions and the studies designed to answer them should be more open to exploring the many ways in which children’s lives differ from each other and discovering the relationships between these differences and factors such as gender or disability. Quantitative researchers should be cautious about designing studies that use a disability measure a priori to define a group of children in order to examine differences between this group and others or to estimate prevalence. Until we understand more about the workings of disability in children and how this might be comprehended through quantitative data, it is almost impossible to be sure which disability measures are appropriate to separate any single group of children from another. Without appropriate measures, we are likely to continually encounter the challenges and inconsistencies of quantitative childhood disability research, which have been outlined in this review. Armed with a fresh understanding of disablement in children and measures of disability that reflect it, we will be better placed to harness the power of large-scale data to improve their lives.

These recommendations, which question a number of the bases on which childhood disability research has been conducted to date, show how much work there is to be done. However, there are currently positive developments in many of these areas, with, for example, a growing literature exploring the views of children themselves; ongoing examinations of how the environment might be brought into the quantitative study of disability; and the increasing availability of longitudinal data exploring the lives of children from before birth. However, the evolution in our understanding of disability from the static ‘disability as impairment’ to the more dynamic ‘social-contextual’ conceptualisation continues to pose challenges for disability measurement, particularly when relying on traditional measurement approaches. As in many other areas, the complex concept of disability in quantitative research is troubling our assumptions and undermining traditional approaches but is thereby providing the opportunity for productive reflection.

Competing Interests

The authors have no competing interests to declare.