Gender was self-reported during the assessment sessions. Ethnicity information was collected either from parents using an online survey (Qualtrics software, Prove, UT) or from schools using a physical or digital questionnaire. Ethnicity questions were based on Office for National Statistics (ONS) recommendations (ONS, n.d.). In addition, we report on participants’ Year Group, including Year 3 (7–8 years), Year 4 (8–9 years) and Year 5 (9–10 years).

This was included in the DEER Study assessment battery from October 2022 to March 2023 and data was collected from 196 participants across seven schools in five boroughs in East London. Body mapping complemented the other measures in the DEER study by providing a non-verbal form of expression and helping to ensure that participants continued to engage well with the assessment battery and did not experience survey fatigue (Britton et al., 2020). Body mapping took approximately 10–15 min for participants to complete (a blank version including instructions is shown in Supplementary material 1). The A4 sheet included (1) a body template for participants to color in red where they experience embodied stress, (2) a ‘recent worries’ bubble for participants to write any recent worries they have had and, (3) a space for participants to draw and describe a symbol to represent their interpretation of resilience. To ensure that children understood the task, a researcher from the data collection team (DP, MN, AM, DSS) explained the activity either to the whole group or to small groups of participants, depending on whether they had finished the other tasks at a similar time. The researcher went through the instructions as written on the A4 sheet to ensure consistency across groups. Of note, 34 participants completed an earlier version of body mapping using multiple colors to highlight embodied stress. However, this made interpretation of the locations of embodied stress more ambiguous, therefore we changed this to red only. These 34 participants’ data were not included in analysis relating to the location of embodied stress. Participants were provided with HB pencils and red colouring pencils for drawing and writing.

Children could provide as much information as they wished. As part of our safeguarding and distress protocol, if a participant became distressed during body mapping, they were told they could take a break or stop taking part if they wished, and the data collection team informed the school. If a researcher was worried about a participant’s safety after reviewing the content of their body map following data collection, they discussed this with the DEER Study principal investigators (JYFL, DO) and, if deemed necessary, informed the participant’s school in case further safeguarding action was required.

We followed manifest content analysis, which involves describing the surface structure of participants’ responses and what is obvious from the data (Bengtsson, 2016). To develop the codes for each data category (worries, body parts and resilience symbols), AM and DSS looked over the data multiple times to gain familiarity. These researchers then independently inputted the data into NVivo 12, which involved transcribing the symbols and coloring data into textual descriptions. The researchers then met to compare the transcripts and develop accurate final versions of them. This ensured a standardized approach to preparing the body maps for analysis. Next, the researchers independently inductively coded a subset of the data, and met to ensure consensus in the coding strategies. The researchers then independently coded all the data and met to develop a final coding framework, with any differences discussed and resolved by agreement. The researchers then recoded the data based on this final coding framework. They met regularly to refine the coding framework and discuss emerging themes and categories (Bengtsson, 2016). During these discussions, the researchers referred back to the raw body map data if needed for clarification and deeper interpretation, particularly for the drawings. Final codes were grouped into themes and subthemes for the worries and resilience data, whilst the coloring data remained as singular codes such as ‘top of the head’. Final themes and descriptions for each were discussed and agreed upon between AM and DSS, with input from JYFL.

As an example of this process, for the resilience data, we initially created separate codes for support from parents and support from family members to distinguish specific depictions of parents from others, such as a ‘cousin’ and ‘my family’. However, we then combined this into a singular code of support from parents and other family members as all of this data related to receiving emotional support from family when feeling sad or worried. Along with other codes related to family and friendship, we then developed the theme social networks, which related to the category “Interpersonal level”. Table 1 displays examples of how codes, themes and categories were developed from the worries and resilience data.

Analysis was carried out on qualitative differences between boys and girls for resilience findings (RQ2). Given that few participants reported ‘other’ and ‘prefer not to say’ for their gender, we did not include participants of these categories of response in this comparison. Owing to relying on children’s drawing and writing, at times it was challenging to read or interpret participants’ responses. Therefore, any responses that were unclear, for example due to ambiguous intended meaning or difficulty reading handwriting, were coded as ‘unclear’. Worry or resilience responses that were missing were coded as ‘invalid’. Worry responses such as ‘no worries’ or ‘no’ were coded as ‘no worries’. Resilience responses which were negative, such as emotions like ‘angry’, ‘worried’, ‘sad’ or a worry being described, were coded as ‘negative descriptions’.

Reflexivity is important to build an awareness of the relational and emotional aspects of research (Warin, 2011). The first author (AM) is a qualitative researcher with an existing interest in ABR methods, which may have impacted the design of the body mapping method. However, our multidisciplinary team, including researchers from psychology, psychiatry, medicine and the social sciences, ensured that we had a diverse range of perspectives and input into the study design. The first and second (DSS) authors remained conscious of how our positionalities would influence the lens through which we interpreted and analyzed the data. We are both adult British women without children and from outside of East London, and so have different lived experiences from the participants. However, we both have experience of working with children in our academic and professional careers. As adult researchers, we could not dissolve the age-based power differences between ourselves and the participants, but rather attempted to use this position to advocate for children’s voices (Caldairou-Bessette et al., 2020). During analysis, we discussed how our ages and cultural and educational backgrounds may have impacted how we made sense of the data, particularly for drawings which can be more challenging to interpret. Our use of manifest content analysis was an attempt to give participants ‘interpretive authority’, staying close to their words and depictions to avoid mis- or over-interpretation (Carter and Ford, 2013). That said, we came to view our findings as “sensitive and reflexive co-construction” (Carter and Ford, 2013, p. 10) between ourselves and the participants.

Table 2 displays the demographic characteristics of the participants. 48.47% of participants were boys, 46.43% girls and 2.55% identified as ‘other’. The sample was ethnically diverse, with roughly equal numbers of participants across year groups (Table 2). Below, we first outline the findings of RQ3 and RQ4 as they contextualize the findings of RQ1 and RQ2. To aid interpretation, we have included tables to display the percentage of participants who reported the themes across research questions.

Table 3 shows the 11 worry themes derived from participants’ written responses (Figure 2). Detailed codes within each worry theme are included in Supplementary Table S1. The most common worry related to Education, largely expressed as anxiety about academic performance. The next most common worries were Social relationships and interactions, followed by Physical health (of the self and others) and Family and the home.

This section displays the results from 162 participants who completed the second version of body mapping (Figure 3). Participants displayed a range of somatic experiences associated with recent stress, resulting in 51 body part categories represented across the body maps. Although largely connected to specific body parts, 5.55% of participants depicted recent stress as a whole-body experience. Table 4 displays an overview of the body parts, which account for those colored by at least 5% of participants, categorized into areas of the body.

The head was the most frequently colored body part (74.07%). Whilst the majority of this was the top of the head (49.38%), the entire head was colored by 22.22% of participants. Of those who colored the top of the head, 4.94% colored the ‘brain region’ by representing cloud-like shapes in the top of the head or labeling their coloring as ‘brain’. One participant clarified next to their body map that “I get stressed in the brain and head.” The hands were the second most frequently colored region (43.83%), with 30.23% of participants coloring both hands and fewer coloring only the right (6.17%) or left (3.09%) hands, the fingers (2.47%) or the thumbs (1.85%). One participant drew a test paper in their left hand. The abdominal/stomach area (40.74%) was the third most frequently colored body part, with 6.17% of participants representing the middle of the stomach, often with a small circle. One participant drew a butterfly in the abdominal area of their body map, reminiscent of ‘butterflies in your stomach’.

The fourth most frequently colored area was the feet (37.65%), with 27.78% of participants coloring both feet and fewer coloring only the left (4.32%) or right (3.09%) foot or just the toes (2.47%). The legs were colored by 28.40% of participants; 13.58% of participants colored both legs, whilst only 4.94 and 4.32% colored only the left or right, respectively. Either both knees or just the right knee (categorized under legs) were colored by 5.56% of participants. The chest area was colored by 25.93% of participants. 7.41% of participants drew hearts; the majority of these were positioned in the chest, but one participant drew this in the throat. One participant, who drew a heart in the left side of the chest, wrote “here.”

Table 5 displays the resilience themes derived from participants’ drawings and descriptions. Detailed codes within each resilience theme are included in Supplementary Table S2. These themes fall into two categories related to how resilience is conceptualized (concepts of resilience) and the factors which contribute to it (socioecological resilience factors). Descriptions of each of these categories and included themes are expanded upon below.

At the Individual level, gender differences were found within the theme of Interests and leisure. Boys depicted Sports (14.89% vs. 3.30%, 14 vs. 3 responses) and Videogames and media (8.51% vs. 1.10%, 8 vs. 1 responses) more often than girls. However, within this theme, only girls drew Toys (3.30%, 3 responses) and Engagement in the arts (8.79%, 8 responses). The theme of Interpersonal level: Social networks was more commonly expressed by girls than boys (19.78% vs. 6.38%, 18 vs. 6 responses). This included aspects of friendship, such as socialising and playing with friends, and family support.

This study employed a novel body mapping tool to explore resilience, recent worries and somatic stress among children aged 7–10 in East London. Aligned with existing literature, worries around education and social relationships were prominent. Somatic experiences of stress reflected common complaints in children, yet body mapping also revealed unique experiences. Resilience was shown to be an interplay of factors across socioecological levels, with personal resources and strong social relationships particularly valued by participants. We also found gender differences regarding resilience factors at the individual and interpersonal levels. Our qualitative research design enabled us to gather data on perceptions and experiences which can be challenging to capture with quantitative methods, particularly with children.

Findings of RQ1 revealed that children commonly defined resilience as an individual capacity to persevere, which aligns with earlier dominant notions of resilience which remain widely held. However, children also used abstract imagery to define resilience, such as through depictions of nature and animals. Whilst some of these depictions were only created by a small number of participants, such as flowers, rainbows and Planet Earth, examining these exceptions can provide insight into diverse and nuanced perspectives (McPherson and Thorne, 2006). For instance, participants associating resilience with nature is not commonly reported in the literature, but alludes to it as an entwinement of mind, body and environment. We can then draw parallels with the growing evidence of the benefits of nature-based interventions, including arts-in-nature, for promoting mental health and resilience in CYP (Moula et al., 2022; Obeng et al., 2023; Overbey et al., 2023). Children often make sense of the world in abstract and imaginative ways, and so drawing enabled children’s shared and unique “ideas, beliefs, and metaphors” (Pridmore and Bendelow, 1995, p. 473) to become part of the research process. Through acknowledging these less common findings, we can gain a deeper understanding of diverse concepts of resilience that are important to children.

In the analysis of resilience factors (RQ1), we situated participants’ responses within a socioecological framework (RQ1), whereby resilience develops from “a hierarchically organized set of protective systems that cumulatively buffer the effects of adversity” (Roisman et al., 2002, p. 1216). Under15standing the factors operating at different levels of influence can inform strategies that promote resilience (Twum-Antwi et al., 2020). The majority of participants’ responses emerged at the individual and interpersonal levels (Figure 6). Sports and other hobbies were frequently cited, which play a role in the development of instrumental and social skills and social networks (Gilligan, 2000). Although hobbies reflect individual interests, their benefits can relate to their social and relational qualities, for example participation in sports has been shown to increase social competence and empathy in adolescents (Caldarella et al., 2019) and arts activities has been found to improve relationship building and a sense of belonging in CYP (Zarobe and Bungay, 2017).

At the interpersonal level, family, friends and pets were commonly depicted. These factors underscore the significance of social support in fostering resilience (Gartland et al., 2019). Research suggests that positive family and social support systems are associated with resilient outcomes in children experiencing social adversity (Gartland et al., 2019). Additionally, pets have been shown to increase children’s resilience and self-worth (Purewal et al., 2017), including the reduction and prevention of depressive symptoms (Gadomski et al., 2015). These findings are corroborated by a recent meta-analysis which found that social support-based interventions yielded the largest positive effects for action-based resilience outcomes (i.e., functional and/or performance-related activities in relation to the intervention) (Liu et al., 2020). This highlights the value of interventions which facilitate supportive relationships between children and important people in their lives, including family, peers and educators (McDonald et al., 2019).

At the organizational level, whilst only two participants described education in relation to resilience, highlighting this as an exception reveals the stark contrast with how commonly school was instead described as a worry. This suggests educational settings as ideal sites for interventions to improve psychosocial wellbeing. For example, school-based programmes focused on mindfulness have been shown to help students overcome social, emotional and behavioral problems (Schonert-Reichl and Lawlor, 2010). The limited findings on community and policy levels are unsurprising given the young age of participants. However, the inclusion of faith as a community factor is concurrent with research suggesting religion to be an important coping strategy and protective factor for many communities (Foy et al., 2011; Pielock et al., 2016; Mhaka-Mutepfa and Maundeni, 2019). Our findings mirror one systematic review which found less evidence for school and community factors associated with resilient outcomes in children experiencing social adversity compared with individual factors (Gartland et al., 2019). Whilst our intentional focus on children in East London contributes to the growing body of literature on resilience in children from diverse backgrounds, there is scope for further research which investigates these wider socioecological factors.

Gender differences in the prevalence and trajectories of mental health difficulties and types of resilience resources that CYP use have been explored extensively (Rescorla et al., 2007; Martin and Hadwin, 2022; Yoon et al., 2023; Mar et al., 2024). Our results found gender differences in resilience factors that align with research on the gendered socialization of boys and girls (RQ2). For example, relational resilience, whereby resilience arises through fostering mutually empathic connections (Jordan, 2023), is believed to be more common in girls who are socialized to see themselves as relational entities (Miller, 1986). Girls in our study more often depicted interpersonal coping strategies (e.g., family and peer relationships). Boys chose individualistic, competitive and task-oriented strategies (e.g., sports and videogames), which aligns with research on their socialization (Frydenberg and Lewis, 1993). However, access to recreational resources is an important community protective factor for resilience (Zolkoski and Bullock, 2012) and the peer acceptance gained through participation in sports may contribute to feelings of self-esteem in adolescents (Daniels and Leaper, 2006). These peer and community aspects of sports may explain the prevalence of team sports described by boys in our sample. These results emphasize the importance of children having access to a range of resources that go beyond gender stereotypes, as this may constrain children’s capacity to develop resilience. Gender-informed and gender-specific mental health interventions may be more effective at supporting CYP’s mental health and resilience (Friedrich et al., 2010; Herrmann et al., 2024). However, few trials of resilience-supporting interventions assess their gender-specific effects, leaving uncertainty about whether targeting gender-related protective factors through interventions could lead to better mental health outcomes (Dray et al., 2017). Future research could elucidate this by incorporating analysis of the effect of interventions by gender. There is also a need for more research which explores risk and protective factors in the lives of transgender, non-binary, and gender diverse CYP to inform tailored interventions (Johns et al., 2018; Chew et al., 2020), particularly given existing research which highlights that these CYP face increased risk of emotional distress, suicidal ideation and severe mental health conditions (Eisenberg et al., 2017; Holt et al., 2023).

The results of RQ3 revealed 11 main worry themes identified from children’s responses. Many of these align with existing literature on the worries of children aged 12 and below, including school, health, social relationships and death (Silverman et al., 1995; Muris et al., 1998, 2000; Christie and MacMullin, 2016). School and academic performance emerged as primary worries, consistent with studies highlighting these as among the most common areas of worry in primary school-aged children (Silverman et al., 1995; Muris et al., 1998, 2000) and young people during the COVID-19 pandemic (Cauberghe et al., 2021; Shukla et al., 2023). Social relationships and interactions was the second most common worry category, reflecting the substantial role that managing social relationships plays in early life stress (Cheetham-Blake et al., 2019). This relational aspect of worry emerged across categories, such as concern about the health of others and public and private self-consciousness. This co-exists with the finding of social networks and supportive interactions as resilience factors, with research suggesting that learning to navigate social relationships and maintaining strong peer and adult-child relationships can help build resilience (Waaktaar et al., 2004; Graber et al., 2016; Twum-Antwi et al., 2020).

The body maps captured 51 body parts reflecting somatic stress symptoms (RQ4). The top of the head was most frequently colored, followed by the hands and abdomen. Headaches and abdominal pain are two of the most common somatic complaints among children and adolescents (Goodman and McGrath, 1991; King et al., 2011), and both have been associated with pressure to perform academically (Passchier and Orlebeke, 1985; Randall et al., 2019). This aligns with the high frequency of worry responses in our study about academic performance. The prevalence of the abdomen and chest also reflect somatic symptoms and measures used to detect anxiety in common diagnostic and scale tools. Whilst the association of stress with the head and hands is less common in these tools, these body parts are associated with common somatic traits of stress and anxiety (e.g., headaches, sweating, shaking and clenching) and were also commonly represented as such in a body mapping study on embodied experiences of anxiety (Vaughan et al., 2023). The prevalence of the hands could also reflect academic anxiety and physical discomfort from writing, for example one participant drew a test paper in their hand. 12 body parts were unique to individual body maps and 45 were found on ten maps or less, supporting the idea that embodied experiences of stress can be highly individualized (Vaughan et al., 2023).

In the absence of verbal data, AM led an activity and discussion with an advisory group of six students in Years 3, 4 and 5, from a primary school in East London, to discuss the findings on resilience. After discussing how they would define resilience, the students were shown the resilience topics and asked to select six which resonated with them and visually represent these on a ‘coat of arms’. Regarding concepts of resilience, the idea of perseverance emerged, with one child describing resilience as “when life is hard, just keep going.” Whilst only one participant in the study depicted Planet Earth as a symbolic reminder of resilience, it resonated with two children from the advisory group, one of whom related it to her love of nature and the outside world.

For resilience factors, videogames were a favourite hobby for several children, one of whom particularly enjoyed playing with their friends. Whilst the impact of videogames on children’s health is often perceived as negative, there is growing evidence of their positive psychological effects (Johnson et al., 2013), including lower prevalence of mental health difficulties and peer relationship problems (Kovess-Masfety et al., 2016). Team sports like football were also commonly described by the group, and for one child their enjoyment was related to being good and football and the sense of achievement that came with this. At the interpersonal level, several children described family, friends and pets as sources of resilience. The home was also depicted, with one child expressing that “my family is my home.” Compared with study participants, all children in the advisory group expressed that they enjoyed school at times, when prompted during the discussion.

Our method draws upon the draw and write approach (Gauntlett and Horsley, 2004), which helps children convey complex ideas and recall events (Horstman et al., 2008). Compared with traditional research methods, drawing may be less anxiety-inducing for children (de Jäger et al., 2016), even for sensitive topics (Hill et al., 1996; Horstman et al., 2008). However, criticisms of draw and write include the challenges of interpretation and analysis of children’s drawings and its limited sensitivity to social context (Backett-Milburn and McKie, 1999; Angell et al., 2014). Whilst AM and DSS referred back to participants’ responses during analysis, some clarity of meaning may have been lost due to reliance on researcher interpretations. Whilst we carried out consultation with an advisory group, a refinement of our method could incorporate the ‘draw-write-tell’ approach (Angell and Angell, 2013), which combines drawing with verbal interpretations. It has been argued to more effectively equalize adult-child power imbalances and centralize children’s voices (Angell et al., 2014). Further research incorporating this approach could be done to evaluate the efficacy of the body mapping method utilized in this study for understanding children’s subjective experiences.

Our study fits within a growing body of literature using body mapping to explore embodied symptoms and experiences, including embodied anxiety and non-anxiety in a general population sample (Vaughan et al., 2023), embodied feelings following a surfing intervention for children with autism (Britton et al., 2020) and as an assessment of embodied trauma in refugees and asylum-seekers (O’Brien and Charura, 2023). With its origins in art therapy (Solomon, 2002), body mapping has been used as treatment for trauma and dissociation in children (Santen, 2007, 2015). Combined, this points to the potential for future research to explore the clinical application of this body mapping approach. For example, authors have suggested that body mapping could be applied to the diagnosis and treatment of anxiety (Vaughan et al., 2023) and assessment of embodied trauma (O’Brien and Charura, 2023). Digital and web-based approaches could be combined with body mapping to offer rapid and dynamic assessment of children’s embodied experiences (Edwards, 2021; Ludlow, 2021; Ticho, 2021; Shanneik and Sobieczky, 2023). This reduced reliance on drawing and movement may make the method more inclusive for individuals who are not comfortable with drawing and those with reduced mobility. Body mapping is also a promising educational tool (Chenhall et al., 2013; Senior et al., 2014; Helmer et al., 2015; Lys et al., 2018; Cook et al., 2022; Henderson et al., 2023). As our version is simple and suitable for large groups, it holds potential as a tool for teachers and others working with children to explore embodied experiences across a range of issues, including mental health (Barnes et al., 2024). Further research is needed to evaluate these potential applications of body mapping.