Abstract
The widespread privileging of children’s voices in recent times has triggered expansion of differing forms of qualitative enquiry that aim to “give children a voice.” Engaging children in research and eliciting their voices on matters that affect them is often showcased as being a more “authentic” way to capture children’s lived realities and afford their agency. Yet, the uptake of voice in qualitative enquiry, and how it may contribute to the privileging of particular ways of knowing (some) children’s lives, is rarely interrogated. Drawing on examples from our own research, in this paper we critically reflect on the frequent invoking of the term voice in qualitative health research with children. In doing so, we challenge claims of authenticity by exposing the tricky epistemological tensions and relations of power that are embedded within the production and legitimation of particular voices as being “correct” ways of knowing about health—including the ways our research intentions and methods contribute to these processes. We reflect on the methodological and epistemological value of silences, dissenting voices and other modes of expression to highlight forms of resistance to adult-led health agendas. We conclude by illustrating how dominant relations of power are (re)produced within and across research spaces, and through the mobilizing or pathologizing of particular young voices through research. Possibilities for advancing ways to harness children’s preferred modes of expression in qualitative research are also considered.
Introduction
The proliferation of research and efforts to give children and young people a voice on issues that affect them has expanded rapidly in recent times—often with reference to article 12 of the United Nations (UN) Convention on the Rights of the Child (UN, 1989). Article 12 states that children (defined as up to the age of 18 years) should be provided with the opportunity to express their views and such views given due regard in line with the age and maturity of the child (Ibid). Qualitative methods of enquiry are often deemed well suited to research conducted with children with their potential to sensitively and effectively elicit their voices on a range of topical concerns (Greene & Hill, 2005). Such approaches are often championed for offering a more authentic account of their perspectives and lived experiences. Indeed, there is now an extensive body of research from across different disciplines that aims to give children a voice on matters that affect them including from education and schooling (see Can & Göksenin, 2017; Forde et al., 2018), health (see From, 2017; Mengwasser & Walton, 2013), social work and social care (see Chapman et al., 2010), geography (see Kanngieser, 2012), psychology (see Fox, 2016), and the humanities (see Chawa et al., 2018).
While these contributions draw important attention to the value of listening to young people, the uptake of the term “voice” in research with children and young people is rarely interrogated (some exceptions are discussed shortly)—including the emergent epistemological incongruences and possibilities for reproducing dominant relations of power. In this paper, we aim to extend recent critiques of the use of voice in qualitative enquiry and specifically, problematize its use in health-related qualitative research with children and young people. We begin by charting some of the issues of representation and authenticity described in recent critiques (see for example, James, 2007; Komulainen, 2007; Spyrou, 2011) before illustrating with examples from our own research some of the tensions that emerge when seeking to “give children a voice” on topical health concerns. Our examples highlight how the frequent invoking of the term can inadvertently support existing social inequities and aid the reproduction of (rather than challenge to) dominant health and developmental discourses that position young people as being immature, risky or at risk. We draw particular attention to the ways in which research processes can legitimate and privilege some children’s voices as being “correct” reflections of health, while others are rendered silent or problematic.
Children’s Voices in Qualitative Research
The strength of qualitative enquiry lies in its ability to capture detailed accounts of participants’ perspectives and experiences, with a particular emphasis on understanding knowledge as being situated and co-constructed through the research process (Denzin & Lincoln, 2017). Qualitative research has been further championed for “giving a voice” to individuals or groups who are often under-acknowledged or otherwise marginalized (both in research and society more broadly). Examples of such research can be found across a range of fields include feminist studies (Belford & Lahiri-Roy, 2018; Haigwood, 2012), disability studies (Goodley & Runswick-Cole, 2015; Liddiard et al., 2018) and race studies (Chang, 2013; Louis et al., 2016).
The extension of voice-based agendas to childhood research has gained significant traction in recent times—and as part of a broader acknowledgment of the dominant tendency to disregard or marginalize children’s own perspectives and experiences (Lundy, 2007). Engaging children in research and eliciting their voices on matters that affect them is often showcased as being a more “authentic” way to capture children’s lived realities and afford their agency. Such approaches are further defended by recourse to their emancipatory and inclusive potential of an otherwise disenfranchised and “vulnerable” group. In relation to health, research with children has taken a focus on their thoughts on health and illness (Brady et al., 2015), health care policy and service provision (Persson et al., 2017), healthy lifestyles and practices (Martin et al., 2018) and health promotion (From, 2017; Mengwasser & Walton, 2013).
Children’s involvement in health research is often formalized through research governance processes, whereby the Patient and Public Involvement (PPI) agenda has sought to ensure that all health research seeks to involve patients (including children) in the planning, funding, conducting and dissemination of research (Thompson et al., 2012). For example, INVOLVE, the National Advisory group for PPI in health research in the UK suggest that PPI in research should be “an active partnership between the public and researchers in the research process, rather than the use of people as the subjects of research” (INVOLVE, n.d.). Such a reframing of the role of patients from passive subjects to active agents is based on the premise that patients bring a different perspective; their experiential expertise, and that this is needed to ensure that research is valued, valid and feasible (Bissell et al., 2018). In response to this agenda, many PPI examples can be found across different health areas including paediatrics (Bate et al., 2016), palliative care (Mitchell et al., 2019) and mental health (Viksveen et al., 2017)—reflecting the growing tendency and efforts to place young people’s “voices” at the center of health research.
Examples of this kind are often commended for reflecting a more “authentic” account of children and young people’s lives and experiences, with some research even suggesting such approaches afford children’s agency, “empowerment,” and full participation rights. Yet, the notion of voice and how it may contribute to the privileging of particular ways of knowing (some) children’s lives, is rarely interrogated. Some recent critiques of the use of voice in qualitative enquiry in relation to children (and broader PPI initiatives [see for example, Bissell et al., 2018; Viksveen et al., 2017]) have drawn critical attention to issues of representation and the epistemological incongruences tied to claims of authenticity (see for example, James, 2007; Spyrou, 2011). Four related issues have been raised. First, the privileging of voice as a mode of expression downplays children’s alternative ways of indicating their preferences, perspectives and possible forms of resistance. Indeed, not using a voice may tell us more about children’s preferences than what is vocalized (see for example, Komulaninen, 2007; Lee, 2001; Lewis, 2010; Spyrou, 2016).
Komulainen (2007) illustrates this tension by unpacking the social and co-constructed nature of speech. Indeed, qualitative interviews are well recognized as a co-constructed process of meaning-making (Westcott & Littleton, 2005). Drawing on ethnographic field work with children with disabilities, Komulainen reveals how children’s voices are both constrained and enabled by the contexts in which speech may be enabled (or not). Understanding the (social and methodological) context in which experiences are voiced helps to uncover the ways (multiple) voices are mediated, constrained or enabled by the prevailing norms about children and childhood (Alldred, 1998). The importance of attending to how and which voices are made possible in any research encounter highlights the ways in which adults come to produce and understand the meaning of children’s voices—including any attention given to silences and the non-verbal (Lewis, 2010; Spyrou, 2016). How “voices” are heard and produced (methods), recognized (analysis) and taken-up (significance) underscores the interpretive and analytical processes and relations of power that ultimately represent the meaning(s) and value attached to children’s voices—a point we return to later in our discussion.
Second, and in some ways to guard against these tensions, a range of alternative methods for accessing children’s voices have been developed. Examples include the use of creative and arts-based methods such as drawings, photo-voice, visual diaries, map-making and so forth (see Thomson, 2008b, for examples of these methods). These participatory methods and related forms of enquiry have been differentially critiqued for their potential to fully enable children’s participation, with some evidence suggesting that participatory projects can remain adult-centered and led (Alderson, 2001; Davidson, 2017; Hart, 1992). The idea that such methods enable access to a greater “authenticity” of experience has been further challenged on epistemological grounds and the suggestion that such methods better capture “true” accounts of children’s realities (Eldén, 2012; James, 2007). Indeed, claims to authenticity present tricky epistemological tensions and the idea that voices “speak the truth” (Spyrou, 2011, p. 152; see also Mazzei & Jackson, 2009). Drawing on Geertz (1988), James (2007) cautions against the uncritical (and selective) representation of children’s accounts (however derived) as being a closer reflection of social reality. Children’s accounts are themselves shaped by prevailing discourses and reflect the social categories and positions from which children speak/are positioned (Eldén, 2012; Mayall, 2000, 2002; Thomson, 2008a).
Third, as suggested, research with children may serve to privilege particular voices (and silence others)—thereby contributing to existing social inequities. The tendency to report the singular, educated, and articulate child voice downplays the diversity and individuality of children and their experiences. Such diversity draws attention to childhood as a structural category experienced differently by children of differing class, gender and ethnicities—and other markers of social identity (James, 2007; Mayall, 2002, Thomson, 2008a). Mayall (2002) suggests that researchers should move beyond simplistic reporting of children’s voices and instead, expose and examine the “different positions from which children speak” (p. 67). Critical attention to the ways in which our research processes and methods differentially position, produce and privilege (some) children’s voices and ways of knowing is thus needed to expose the ways in which research may inadvertently contribute to the reproduction of children’s “marginalized” status.
Fourth, the elicitation of children’s voices through research (and other participatory mechanisms) continues to be derived from and located within adult frameworks of power and necessitates the legitimation of children’s perspectives as being valid. Evidence of tokenism is documented in the literature (Alderson, 2001; Hart, 1992; Lundy, 2007), but less criticality has been directed toward the ways in which adult-informed frames of reference shape the analysis of children’s voices—often in very subtle ways. The privileging of particular voices (and the possibilities for marginalizing and pathologizing others) through the processes of interpretation and reporting can reflect preferred (adult) ways of knowing. Alldred and Burman (2005) thus call on researchers to reflect on their own interpretive, authorial and editorial decisions when seeking to represent children’s voices, including how such processes ultimately determine which voices are deemed most significant to the topic under investigation.
These issues of authenticity and representation have been relatively overlooked in health-related research with children and young people as efforts to garner children’s voices on health areas remains increasingly popular and as part of broader PPI agendas in health care. Drawing on examples from our own work, we extend the aforementioned critiques to illustrate how the privileging of some voices can reflect the imposition and reproduction of dominant (adult/child) power relations—ultimately producing “correct” voices about health. Through our methodological reflections, we aim to enhance the conceptual elaboration of the notion of voice in qualitative research with children and offer new directions for qualitative forms of enquiry with young people. To that end, we start with a brief context to the public health agenda in relation to children’s health and how this has triggered efforts to “give children a voice” on topical health issues, before proceeding with examples from two different studies with children (aged 9–10 years) (Fairbrother et al., 2016) and young people (aged 15–16 years) (Spencer, 2013a).
Privileging and Pathologizing (Some) Children’s Voices on Health
For children and young people in particular, concepts of health are often defined by risk discourses. Public health has been particularly active in arousing concerns about the threats to children’s wellbeing in relation to obesity, cigarette smoking, alcohol and mental health (Fullagar, 2012), with some children positioned as being “riskier” or “at risk” than others (Kelly, 2003, 2006). Those experiencing multiple forms of disadvantage and inequity, for example, have been found to be more likely to engage in “risky” health practices (Green et al., 2013; Hanson & Chen, 2007; Tu et al., 2016). Children and young people from lower socioeconomic groups are both more likely to smoke (Taylor-Robinson et al., 2017) and more likely to start smoking at an earlier age (Green et al., 2013). Similarly, socioeconomic inequalities have been linked to both poorer diets (Ranjit et al., 2015) and lower levels of physical activity (Sterdt et al., 2014). Because of this, there have been concerted efforts to elicit children’s perspectives on a variety of health areas and to harness these within the development of “relevant” forms of health promotion for children and young people.
Against this backdrop, Fairbrother et al. (2016) sought to explore how children from socioeconomically contrasting communities in the North of England understood the relationship between food and health. 53 children aged 9–10 years participated in interviews and debates in friendship groups in schools (phase one) and a sub-sample of eight children and their parents participated in in-depth individual interviews in the home (phase two) A flexible topic guide was employed with the aim of being responsive to, and probing ideas emerging spontaneously from, the participants themselves (Ridge, 2003). In keeping with child-centered methodologies (e.g., Christensen & James, 2008), all interviews and debates included the use of visual prompts and gave children the option of creating drawings themselves with the aim of maintaining children’s interest and enjoyment and prompting discussion. In this way, the study design and selected methods were guided by sensitivity to the well-documented power differentials between an adult researcher and child participant (Greene & Hill, 2005; Punch, 2002) but also the research topic. The elicitation of children’s voices was thus aided by, and produced through, the four principles outlined by Holloway and Jefferson (2000) as follows: 1) the use of flexible open-ended questioning rather than closed questions, 2) eliciting stories, 3) avoiding why questions, and 4) following up using respondents’ own ordering and phrasing. In part, this approach sought to ensure children felt comfortable and at ease to share their ideas throughout the interviews and discussions.
Despite such intentions, during the course of the study it became apparent that such processes seemed to aid the production of some children’s voices more readily than others—most notably those located within the socioeconomically advantaged area. Children from this context were particularly apt and articulate in their debates on the relationship between health and food, demonstrating their knowledge of healthy diets. These children seemed keen to share their understandings and ideas, often eloquently and fluently extending the discussions with phrases such as, “And can I tell you like?” At other times, these children took on the role of the interviewer, asking each other questions and facilitating and extending subsequent discussions to share their ideas with phrases like, “Shall we listen to Lizzy?” In some ways, these positive dynamics could be taken as evidence of a highly “successful” interview, with minimal “interference” from the moderator (Krueger & Casey, 2015). Yet such dynamics resulted in a significant imbalance in data generated across different groups of children. Because of this, some children’s voices were clearly more evident than others, with transcripts varying considerably in length between different groups. Despite best intentions, the study and its methods ultimately produced much more data from the children from the school situated in the socioeconomically privileged area (c.f. Spyrou, 2011).
This imbalance was particularly pronounced during the analysis and writing-up stages of the study. The piecing together of more “disjointed” conversations that were shorter in length or featured “matter of fact” answers and silences were less amenable to conventional coding processes in qualitative analysis, where the focus is often on the text itself (Miles et al., 2013). While some analytical approaches (e.g., conversational analysis) attend to the non-verbal, silences, pauses and other gestures, greater analytical work and interpretation on the part of the researcher is arguably needed in order to determine the meaning of the non-vocal. Such difficulties not only raise a note of caution about how such utterances and silences are interpreted by the researcher’s interpretation of utterances and silences (Spyrou, 2016) but, in this particular study, further guided the selection of data and quotations to substantiate the analysis.
For example, as part of the conventions of qualitative data reporting, quotations are frequently, yet selectively, used to illustrate key ideas (Miles et al., 2013). Methodological decisions of this kind are often guided by identifying “typical” or commonly expressed views, but also through explaining qualitatively the relationship of the quotations included to those that were left out (Mason, 2002, p. 184). Because of their fluency in relation to the goals of the enquiry, quotes generated from children located in the socioeconomically advantaged area were used more readily in the initial stage of writing up. In this way, the study reporting could have inadvertently contributed to the idea that articulate voices reflect “correct” understandings of health and through privileging some children’s voices, while silencing others. To avoid such concerns, a concerted effort was thus made to redraft the findings and redress any such imbalances—but this again illustrates the centering of adult-led decisions on the analysis and reporting of children’s voices.
In a separate study, Spencer (2013a) also sought to elicit a range of young voices on health. The main goal of the enquiry was to examine the challenges of empowerment in relation to health promotion with young people (Spencer, 2013a). Problems of production and interpretation were similarly evident but also extended to how such voices were “heard” and received as part of the study’s dissemination activities. To avoid priming around dominant health discourses, a combination of ethnographic methods was used to deepen understanding of young people’s own concepts of health, health risks and their priorities for health promotion. Methods included 6 months intensive participant observation, focus groups and in-depth individual interviews with 55 young people (aged 15–16 years) from diverse social, educational and economic backgrounds. Data were generated in a school and local social settings such as youth clubs, music and sports clubs and the town center.
Findings are reported elsewhere and revealed the relative successes of using a combination of qualitative methods for accessing young people’s notions of health. These perspectives included the importance young people attached to having fun and the detrimental impacts of being judged by others (adults and peers) on their social and emotional wellbeing (Spencer, 2013a, 2013b). Yet such findings proved problematic during dissemination activities (e.g., presentations delivered during school assemblies, written summaries of findings for participants/school personnel) and largely because young people’s “different” ways of talking about health were often (mis)understood by teachers and other adults as a product of young people’s lack of knowledge of health or evidence of their (im)maturity. Indeed, far from suggesting possibilities for agency, evidence of enjoyment or fun was taken as young people’s inability to take things seriously or their tendency to engage in reckless and risky behaviors.
Participants from this study were, however, acutely aware of the dangers of expressing opinions that departed from adult-defined “legitimate” perspectives and how these might be (mis)understood and interpreted by adults. Because of this, participants often stated that “there’s no point” in voicing their opinions since any counter perspective simply reinforced the idea that they were unruly or disruptive. These young people often expressed their views by remaining silent and in an effort not to reproduce the very same (risk-based and developmental) discourses they sought to contest. While such silences can be interpreted as evidence of their resistance during the analysis stages (see Spyrou, 2016), reporting on such silences was also problematic and because a lack of vocalization supported the idea that these young people were largely apathetic and unwilling or unable to engage in “important” discussions. In this context, both young people’s voices and silences were pathologized as evidence of their (im)maturity and stage of development. Indeed, participants in this study felt that their voices were only permitted when they aligned with dominant (read adult) perspectives on health (see also Alldred & Burman, 2005; Greene & Hill, 2005).
Reporting examples of young people’s resistance to adult-led health agendas was thus especially difficult because of the potential to inadvertently support the idea that young people were disruptive and unruly. When offering a counter perspective (e.g., adults are hypocritical and engage in health risk behaviors more often than young people), these young voices seemed to strengthen the idea that they were unable to take things seriously or were disrespectful to adults. Involving young people in dissemination activities therefore required careful mediation to ensure their perspectives were not “misheard” or misunderstood. Examples of this kind highlight some of the potential “risks” and unintended consequences of presenting “uncensored” dissenting voices (see Lather, 2009; Mazzei, 2009; Spyrou, 2011)—particularly when such voices reflect a challenge to dominant social norms or adult/child hierarchies. Yet through the processes of mediating and sanitizing accounts in order to avoid misappropriation of young people’s meanings (e.g., through the removal of offensive language when describing adult health behaviors), the adult researcher once again controls the analytical decision-making and how and which voices are ultimately represented to convey (socially sanctioned) meanings of health (see also Alldred, 1998). Through privileging voice, other forms of contestation or resistance to adult-led agendas may be overlooked.
Reflections on this study highlight how adult-led decisions about which voices were deemed correct (or otherwise pathologized) are closely tied to the underlying structural conditions and discourses that define and mediate the legitimacy of young people’s voices (Alldred, 1998; Mayall, 2002). The reframing of young people’s voices as evidence of their immaturity, ignorance or lack of competence (especially when such voices reflect a challenge to adult ways of knowing) highlight how dominant developmental trajectories influence how young people’s voices on health are heard and acted upon (or not). Despite suggesting young people’s agency, the processes through which young voices are harnessed, analyzed and reported in research thus remain deeply embedded within adult/child power relations, which continue to define and legitimate the value of children’s voices (James, 2007; Spencer, 2014).
Discussion
The privileging of children and young people’s voices through research has offered new opportunities to understand their perspectives and experiences on a range of issues. Yet, the uncritical uptake of voice-based research as a “good thing” downplays the complex relations of power that shape our research intentions and processes. Indeed, power mediates how all research is produced (Holland et al., 2010; Spyrou, 2016) and reflects “a process of representation” (James, 2007, p. 268). Through our methodological reflections, we have sought to expose some of these processes and their effects and how (despite best intentions), differing research methods and contexts can contribute to the production and legitimation of some children’s voices, while rendering others silent or problematic—highlighting too, how prevailing developmental and risk-based discourses can shape how children’s voices are represented and (mis)understood (Alldred, 1998).
Greater criticality is thus needed in terms of how dominant discourses influence not only how young voices are accessed and understood, but also how they shape what children are enabled to say—as well as how we understand the meanings of what is said (or not said) (Spyrou, 2011, 2016). In relation to health, there has been little criticality in terms of who is enabled to take an active role in health research and how particular approaches to research—including the contexts in which research takes place, can privilege some children’s voices, while silencing others. Indeed, a preoccupation with ensuring that (any) children are involved in health research projects to ensure that research governance requirements (e.g., PPI agendas) are fulfilled may reinforce the potential that the usual (assenting) voices are included.
Spyrou (2011) highlights that, “when children speak they do so by drawing from the repertoire of their inherited social languages and speech genres which constrain to some extent what they can say” (p. 159). In the context of widening inequalities in child health, widening socioeconomic inequalities and increasing child poverty, it is particularly important to explore the contexts within which children are involved in research (Taylor-Robinson et al., 2019). Mason’s (2002) emphasis that interviews are “highly dependent upon people’s capacity to verbalise, interact, conceptualise and remember” (p. 64) and therefore can never be taken as a direct representation of people’s understandings outside the interview context resonates here. It is thus by no means enough to work out our sampling frame based upon including children from different backgrounds, researchers must also take a reflexive stance in their analysis and writing up to guard against perpetuating inequalities and marginalizing those from disadvantaged backgrounds.
Such critiques are not only relevant to our data collection methods and sampling strategies but extend to the contexts chosen to access children—and, as we have argued, how children are positioned within such contexts. Adult-controlled contexts (such as the school setting) have been found to enable or inhibit some children’s participation (Allen, 2007, 2008; David et al., 2001; Robinson, 2011). Here, we have illustrated how dominant ideas about what it means to be healthy can determine the extent to which children’s perspectives are given credence—particularly when these reflect an alternative perspective. Paying attention to how our methods produce socially sanctioned or dissenting voices underscores the importance of reflecting on what children say in relation to what is possible for them to say in any given context—as well as what is possible for us to hear them saying (Alldred & Burman, 2005, p. 176). Analysis of the non-normative, dissenting voices and silences is thus central to how knowledge about childhood and children’s lives is produced, and in order to guard against the tendency for such voices to be overlooked or problematized as evidence of some young people’s disruptive or “risky” practices—thereby serving to strengthen negative discourses about young people.
Concluding Comments
Our reflections highlight how particular research methods and contexts can inadvertently privilege some voices, while pathologizing others. Such processes can contribute to the reproduction of socio-economic and adult/child hierarchies—thereby undermining the emancipatory ideals embedded with the idea of eliciting voices. Methodological reflexivity needs to encompass a critical reflection on how we come to interpret and represent children’s perspectives, but crucially how such representation may inadvertently contribute to children’s marginalized and disenfranchised status. The dominant tendency to evaluate the value of young voices in line with developmental trajectories and age-based competencies highlights the need to develop approaches that expose the tricky issues of power in research that shape the representation of children’s voices—rather than suggesting that giving children a voice simply diffuses unequal adult/child power relations (Davidson, 2017).
Developing and enacting methodological sensitivity and reflexivity is not without its own difficulties, particularly when much research is set within defined timeframes or institutional or governance frameworks that typically operate from an adult lens. Nevertheless, we encourage researchers to engage in some “uncomfortable” questioning about their research goals and approaches. Such questions may include: How do the study’s aims and theoretical underpinnings contribute to children’s status and positioning—both during the research process but in society more broadly? How do our sampling processes encourage some young people’s opportunities to participate, while others are excluded? How do our methods enable children to reveal, question, challenge and offer perspectives that run counter to what we think we know? And how do we respond to such perspectives? How do we encourage children’s dissent or refusal to participate? How can we involve children in the interpretation of meanings? How do we decide what gets counted in the analysis, and what are the effects for children and understandings of childhood?
By developing critical reflexivity of our research processes, we can offer greater transparency about the nature of social relations guiding the production and legitimation of some children’s voices—including their situated, multi-layered nature (Davidson, 2017; Kraftl, 2013). Crucially, we underscore the importance of recognizing and understanding the effects of children’s positionality within the research process—including how this positionality enables (or not) different perspectives to emerge and the extent to which they are taken seriously. Affording children’s agency through research demands greater criticality including attention toward how divergent “voices” are interpreted and understood within broader discourses of childhood. In particular, as adult researchers we need to be in tune to the ways in which our methodological and interpretive decision-making produces and validates (some) children’s voices and in particular ways (Alldred & Burman, 2005)—including possibilities to reproduce existing social inequities. Without such criticality, opportunities for children’s voices to meaningfully inform health agendas will remain dependent upon the ways in adults’ privilege some children’s perspectives as “meaningful,” “appropriate” or “correct” voices on health.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Hannah Fairbrother's research was supported by a University of Sheffield PhD Studentship. Grace Spencer's research was funded by an ESRC studentship.
