Ethics in Categorizing Ethnicity and Disability in Research with Children

: The use of categories is a contested subject in social sciences. The use of social categories allows researchers to explore similarities, differences, and inequalities between groups of people. However, by using social categories, researchers run the risk of essentializing differences. The aim of this article is to problematize the procedural and relational ethics of using categories in research with children. Based on two research projects studying inclusion and exclusion in physical education, we examine the ongoing ethical dilemmas of categorizing children in terms of disability and ethnic background. The reflections are grounded in intersectional and relational ethical perspectives with a focus on how power is manifested in practices and structures throughout the research process. The data consist of field notes, transcripts of interviews with children and their parents, and the authors’ reflective accounts. The results are organized into three main themes: (1) How categories frame the research in its initial phases (informed consent and voluntary participation), (2) power relationships in context (navigating meanings of categories in the interviews and the relational ethics of generational ordering in combined interviews with children and their parents), and (3) (re)constructing stories and ensuring anonymity. In the discussion, we reflect on how singling out groups of children framed the research, how categories and power relations were negotiated and navigated in interviews and fieldwork, and how, in the reporting of the results, understandings of the children and their experiences were constructed. We argue that by not reflecting on the ethics of categorizing children in research, researchers are in danger of reproducing rather than challenging social inequality and discrimination.


Introduction
The use of categories and the act of categorizing human beings in research is highly contested [1,2]. The underlying dilemma is what Gunaratnam [3] (p. 31) referred to as the "treacherous bind" of categories, in which researchers need social categories to address issues of inequality and discrimination, but at the same time, researchers need to critically de-and reconstruct these "discursively entangled" concepts. Several scholars have called attention to how categories such as children at risk, vulnerable children, disabled children, and ethnic minority children form ideological thought and political action [2]. Categories do not neutrally describe concepts, but rather contain political guidelines and can lead to stigmatization and hierarchization among people if left unexamined [3,4]. For example, scholars have pointed out the tendency of presenting people belonging to ethnic and cultural minorities in research literature only when they negatively deviate from what is considered normal/mainstream [5][6][7]. In disability research, diagnosis and categorization of children into disability groups is often objectively reported by portraying the disability as an inherent feature of the child rather than a constructed category. Within such research, the category and label of disability often connote marginality and stigma [8]. Hence, the use of categories in empirical research calls for ethical consideration.
The issue of categorization is particularly relevant in areas related to health, physical activity, and physical education [9,10]. While categorization of race and ethnicity in epidemiological research is considered important in order to generate knowledge to support public health initiatives [11], scholars have questioned how people are forced into broad categories that do not account for increasingly diverse populations and, furthermore, how studies often fail to recognize differences within groups of people [11,12]. Furthermore, scholars have pointed out how research regarding ethnicity in Western societies tends to center on the experiences of the minoritized other and is undermined by colorblind approaches [13]. As such, white researchers in the area of physical education have started to examine the ways in which the taken-for-granted nature of whiteness shape their professional identities and the research they engage in [13,14]. Scholars have also illuminated how children from minority backgrounds or with disabilities are often placed within homogenous and fixed categories and treated within deficit discourse [15][16][17][18][19]. Within this discourse, children of minority backgrounds or with disabilities are seen as lacking the skills, values, and norms to be recognized as good and competent students in physical education [20]. Scholars have suggested that research has contributed to (re)producing categorical thinking and othering by focusing on how children's characteristics, such as cultural background or ability, act as barriers to participation rather than examining how the subject in Western countries is racialized, white-centric, and embedded in thoughts of Eurocentrism and ableism [19,[21][22][23]. Furthermore, scholars have critiqued how many studies are based in a single issue approach (i.e., focused on disability, gender, or ethnicity alone), which runs the risk of missing how experiences are influenced by multiple aspects of individual lives or the marginalization and exclusion experienced by children who fall outside the scope of the category of focus [12,23,24]. To counter this, there has been a growing body of research applying intersectional frameworks to investigate the ways in which students' multiple identities are relevant to their physical education experiences [12,23].
While our reliance upon social categories in social research can reproduce dominant conceptions of the category in question, categories can also be used to mobilize political action and transformation [3]. As emphasized by feminist scholars in the field of physical education and health, this requires scholars to engage in critical reflections regarding the complex issues of power entailed in research relying on categories of difference [5,10]. The implementation of intersectional frameworks has been illuminated as important to address categorization and power relations in socially just ways [2,16,[25][26][27]. Intersectionality provides tools to understand the complex, dynamic, and contextual character of categories and how they are experienced in individual lives and in interactions [2,5]. Yet, the call (and responsibility) to challenge stereotypes and bring forward diversity within an intersectional framework raises ethical issues that are less discussed in the literature.
A large body of research drew attention to the issue of categorization in regard to how it is represented and implemented in official measurements and procedural ethics [11,28], however, there is still a need for more researchers to provide insight into how these challenges are navigated in specific research situations, contexts, and social relations [3,10,[29][30][31]. Based on two research projects studying inclusion and exclusion in physical education in schools among children with disabilities and of diverse ethnic backgrounds, we reflected upon the behind-the-scenes messiness of using social categories in research [32]. In the article, we problematized the procedural and relational ethics of using categories in research with children and reflected upon the relational encounters between children, parents, teachers, and researchers in the two projects. The aim was to contribute to the discussion of how categories framed the knowledge produced and the power relationships between the researcher and the participants. More precisely, the questions asked were: How does singling out minority groups of children frame our research? How are categories and power relations negotiated and navigated in our research? How are children and their experiences reconstructed in writing?
In this article, we build on the writings of Carolyn Ellis [30,33] to understand the relational ethics of categorizing children in research. Ellis stated that ethical research means more than getting a project approved by the ethics committee (procedural ethics) and replacing names with pseudonyms. Relational ethics "requires researchers to act from our hearts and minds, to acknowledge our interpersonal bonds to others, and initiate and maintain conversations" [30] (p. 4). The ethics of engaging children in research are widely discussed in childhood sociology literature [34][35][36][37]. While unequal power relationships are present in all research with human beings, research involving children is influenced to a greater extent by perceptions of their competence and vulnerabilityparticularly for children categorized as disabled or from an ethnic minority background [38]. Research centered on children's rights to be listened to and to take an active role in research that directly affects them requires a redistribution of power in the research relationship [5]. Yet, the redistribution of power in research is difficult to facilitate. Berry Mayall [27] argued that the asymmetrical power relationship of childhood versus adulthood is constructed as a principle of social categorization and generational organization that all researchers attempting to conduct research with children need to reflect upon. Ethical research practices with children require that we recognize children's inherent vulnerability while questioning their structurally constructed vulnerability [34][35][36]. Inherent vulnerability is a consequence of biological immaturity. Structural vulnerability, in contrast, arises as a consequence of, and is reinforced by, social and political structures that produce powerlessness in children. Intersectional and relational ethical perspectives allow researchers to consider ethics beyond their official rights and responsibilities, and instead base them on thoughtful/caring relationships, thus providing tools to balance power relationships in research situations [5,27].

Materials and Methods
The current article was based on two research projects exploring issues regarding inclusion and exclusion in the context of physical education in Norway. The Norwegian Social Science Data Services (NSD) approved the projects (Project A: 35845, Project B: 39074). Data collection for the projects took place from 2014 to 2015. Both projects conceptualized inclusion within the agenda of education for all, which emphasizes equity and respect for diversity as important principles guiding policies and practices [39]. However, the projects differed in terms of research design, the (non-)use of predefined categories, and how categories were used in terms of the selected analytical perspectives used to provide insight into inclusion/exclusion [24]. By exploring the two linked, yet distinct projects, we aimed to contribute the methodological literature regarding performing ethical research with children. As Cecchini [29] (this issue) argued, the risk of reproducing the stereotypic and marginalizing understandings that research seeks to challenge apply to all kinds of methodological approaches. In line with Ellis [30], we argue that investigating how we navigate these shared challenges will strengthen knowledge regarding how to construct socially just research. The following sections describe the two projects, the data, and the analytical approach used for this article.

Project A
The first project addressed students' experiences of inclusion and exclusion in multiethnic physical education classes. The aim of the project was to explore, from an intersectional perspective, the diversity of stories among students from diverse backgrounds [10,12]. In order to capture the complexity of lives in context [26], an ethnographic design using participant observation and semistructured interviews was chosen. Data consisted of field notes from participant observations of 56 physical education lessons and interviews with 17 students. At the end of the fieldwork, students were selected for interviews according to a generic purposive sampling technique [40]. In order to sample a diverse group of students, the participants were selected based on gender, ethnic background, social groupings, visible skills, and attitudes expressed toward the subject. The interview guide was designed to complement the field notes and generate rich accounts of experiences regarding welfare, learning outcomes, and perceived learning environments within physical education and in school in general. Based on an intersectional perspective, the interview guide also contained questions about family background, leisure-time activities and interests, and social relationships in order to understand the students' stories in a larger context. About one-third of the students in the classes were bilingual, having backgrounds from countries in South Asia, the Middle East, West and East Africa, and North America. All students except one were born and raised in Norway. The first author-a female, white, non-disabled, ethnic Norwegian PhD student-conducted the fieldwork and interviews. The observed lessons were spread over three semesters for the two classes. During the study period, two male and two female, white, ethnic Norwegian teachers were involved. The school, a public school located in the Oslo area, was contacted through the physical education teacher, and permission to conduct the research was obtained from the school management. In terms of procedural ethics, written informed consent was obtained from teachers and parents and oral consent from the students interviewed. Consent stated that all data would be handled with confidentiality, and interviewees were informed of the possibility of withdrawing at any time.
Because of its attention to context and recognition of individual agency, applying intersectional frameworks is viewed by many scholars in the field as one solution to address categorization and power relationships in the research process [5,10,16]. Viewing identities as multiple, fluid, and shifting, I decided to enter the field with an inclusive approach, not focusing on a special group of children and not knowing anything about how students might be categorized according to official measures of ethnicity. In order to identify classes where students had diverse ethnic and cultural backgrounds, the school chose classes for the fieldwork based on number of bilingual students present (approximately 40%). However, in the initial phases of the fieldwork, I was not familiar with which students were bilingual or not. This approach allowed me to explore how, when and which differences mattered, as well as how students negotiated different positions in the class. By applying an intersectional perspective crosscutting social division, such as gender and social class, individual differences and their opportunities and constraints could be recognized [12]. However, working with ethnicity in research requires specific ethical considerations, particularly in research where the illumination of individual experiences related to ethnicity is central [1,7]. Not knowing the ethnic backgrounds of the students also raised issues around power relationships between the researcher and the children. Blurring the lines of ethnicity challenged me to reflect critically upon the ways in which my own background informed the questions and observations [41]. However, it possibly made it more difficult to reveal how issues related to ethnicity often work in subtle ways [42].
While the project aimed at deconstructing essentialist and racialist images of the ethnic other, social categories were considered important in order to reveal power structures, exclusion, and social injustice [3,10]. One of the strategies for this was the application of an intersectional lens in combination with a thematic narrative analysis of the ethnographic data [43]. Children's stories can provide insight into the structural and contextual processes that produce inequality and exclusion. This insight can be gained by considering the processes of positioning and identification in relation to categories of difference (i.e., by paying attention to how the students identify themselves and others in terms of ethnicity, being fit and sporty, or physical appearance) [22]. The thematic narrative was important, as it stressed that interviews were analyzed separately and that extracts were not separated from the interview when interpreting the sequences of text. Extracts were first interpreted in light of the data, i.e., the interview as a whole, interviews with peers, and field notes. Secondly, the data were interpreted in light of the national and political contexts and previous research in the field of physical and general education.

Project B
The second project (Project B) was a multimethod research project that addressed inclusion in physical education as experienced by children with disabilities and their parents [44]. The understanding of inclusion in this project was directed toward children's rights to "participate in regular physical education with their peers while receiving the supplementary aid and support services needed to take full advantage of the curriculum and the social, physical and academic benefits it aims to provide" [45] (p. 3). Although the selection of participants proceeded from a medical categorization of individual characteristics of physical, mental, or emotional functioning, disability in this project was understood within an interactional approach [18,46]. The interactional approach to disability recognizes that disability is not experienced or lived in the same way by all individuals; rather, it is understood as contextual, situational, and individually dependent [46]. The aim was to better understand the interactions between these personal and environmental elements and what supported or hindered inclusion in physical education. Overall, Project B consisted of one survey study and one interview study with children and parents. The data for this article were limited to the interview study.
Purposeful criteria-based sampling was used to recruit participants in the interview study. The main criteria were that the child was identified to have a disability and attended general school. Participants were recruited at a rehabilitation center specialized in physical activity for children with disabilities. In the study, 15 children with disabilities (nine boys and six girls) and 26 parents (10 fathers and 16 mothers) were included. Seven children were diagnosed with cerebral palsy (CP), five with Down syndrome, four with physical disabilities, three with learning disabilities, two with Asperger spectrum disorder (ASD), and four with other disabilities, such as visual impairment or an unspecified diagnosis.
The interviews were conducted while the families attended a three-week stay at a rehabilitation center. The overall themes in the semi-structured interview guide were: (a) Children's placement in physical education, (b) children's experiences with the activities and organization of physical education, c) social relationships with peers and teachers, and d) experiences with the learning climate. The interviews were recorded and transcribed verbatim. The exceptions were two interviews in which I wrote notes to log the conversation because the child preferred not to use a recorder. The interviews were conducted as combined interviews with the child and the parent together or as individual interviews, depending on the participants' preferences (12 interviews with children and parent(s) together, three interviews with the child alone, and 13 interviews with the parent alone). The second author-a female, white, non-disabled, ethnic Norwegian PhD student-conducted the interviews. Written informed consent was obtained from children and parents.

Data and Analytical Approach
Throughout the two projects, the authors wrote reflexive accounts to raise awareness and reflect on how their backgrounds and experiences might have influenced the questions asked and the understandings constructed [47]. Field notes, transcripts of interviews, and the reflective accounts written throughout the two projects formed the basis of the discussion in this article. We first discussed the ethical dilemmas we experienced while working with social categories in the projects. From the discussion we developed the aforementioned research questions. Based on the research questions-which incorporate ethical challenges from the initiation of projects to the write up of the results-we reread our field notes, the interview transcripts, and our reflective accounts to better understand the relational ethics of navigation and negotiating social categories in our project. The selected extracts and narrations problematize our research in terms of negotiating social categorization, power and generational ordering, and our relationships with the participants. The narrations take the reader backstage of the research and offer confessional tales regarding the relational ethics of performing research with children categorized in terms of disability or ethnic belonging [32]. All names of participants in the extracts are pseudonyms to maintain anonymity.

Results
This section presents the procedural and relational ethical dilemmas of using categories and the categorization the we experienced in our research, from initiating the project to the final phase of reporting the knowledge constructed. The results are organized according to the research topics: (1) How categories framed the research in its initial phases, (2) power relationships in context, and (3) (re)constructing stories and ensuring anonymity, which are related to the ethical dilemmas involved in reconstructing the children and their experiences in this article. We each bring in our own voice to reflect upon our projects, A and B.

How Categories Framed the Research in its Initial Phases
The following section explores how categories built into the definition of the projects influenced the approaches used to recruit participants. In both studies, the analyses of the data illuminated several ethical issues and communicative challenges regarding the navigation of generational ordering in research with children. In the initial phases, Project B encountered ethical issues regarding informed consent, while in Project A the challenges were related to ensuring voluntary participation.

Informed Consent
In Project B, the processes of recruitment and securing children's informed consent presented several ethical challenges. The Norwegian ethical guidelines recommend consulting parents before engaging children under the age of 15 in research, thereby allowing parents to act as gatekeepers. However, children also have the right to receive enough information about the project to make an informed decision on whether or not they want to participate in the research [34][35][36]. In the initial stages, I often felt a sense of powerlessness by this dependency on the goodwill of parents and how parents described the research to their child [36].
As participants have the right to opt in or out of research without having to explain why, it is often difficult to know why people refuse to participate [34,36]. In the communication with parents for Project B (i.e., before the interviews), parents voluntary offered some reasons why they refused to partake. Some parents refused to take part fearing that the research would stigmatize the child and construct differentness. Several children in the project had previously experienced a school system in which inclusion was just a rhetorical ornament, while in practice they experienced exclusion and marginalization. One parent explained that she did not want her daughter to take part in the research because of the value-loaded term inclusion. According to the mother, the daughter was tired of always feeling different and in need of adaptation. By refusing to inform the daughter of the project, she was protecting her from yet another place where she was singled out because of her impairment.
The ethical challenge of informed consent was also apparent in the interviews. In some of the interviews, it appeared to be the parent wanting to take part in the project to learn more about the physical education setting, while the child took part because their parent had strongly encouraged them to do so. In such cases, ethical regulations and guidelines designed to protect children can also construct children's vulnerability in research. In the case above, the parent and the researcher both constructed this situation. These situations required that the researcher pay attention to what was going on, recognize and see both the parent and the child, read the relational cues, and (re)act in the best interest of the child [33]. In some situations, the interviews were cut short because I could see that the child was tired, bored, or uncomfortable.

Voluntary Participation
At the beginning of Project A, seven of the students did not give their consent for participation. The NSD guidelines stated that the project could be initiated, however, that these students should not be included in the data. Yet, entering a field where not everyone had given their consent provided several ethical dilemmas. For example, how could children be removed from the social interplay among peers in a class? Was there any way that non-participating children could be part of the data without violating their decision not to be part of the project? At the beginning of the fieldwork, I decided to note the non-participating children in terms of only their gender in the field notes. As the fieldwork unfolded, I gained a rapport with some of the students who did not participate, and four of the seven later decided to join the project. At this point, I was more familiar with the students and could go back to the earlier field notes to write the students into some of the accounts.
There are ethical concerns regarding entering a setting where some people have not given their consent for participation [1]. Why, then, was entering this particular field and finding ways to include the non-participating children so important that the researcher decided to challenge one of the cornerstones of research ethics? At the beginning of the fieldwork, a number of the non-participating children were identified by one of the physical education teachers as bilingual and having an ethnic minority background. As the purpose of the project was to generate stories of children with ethnic minority backgrounds and to investigate the inter-ethnic relations between students of minority and majority backgrounds [12,41], the non-participating students constituted an important target group. Issues related to recruitment and consent are particularly urgent in research relying on certain categories as the starting point [10].
In Project A, the children had the opportunity to gain insight into the project before consenting. Because consent to research is conceptualized as a process, the fieldwork design allowed me to build trusting relationships with the not-yet participants to familiarize them with the project before they made a decision [30]. Building trust in research takes times and requires fieldwork.

Power Relations in Context
In both projects, we experienced asymmetrical power relationships in the interviews with the children. Some of these could be sensed during the interviews and were recorded in the reflective accounts, while others became more evident when reading through the transcripts.
In Norwegian! Obviously! Not...no. If you think that I am Persian, then for sure you think "Ooh she is probably used to such Persian stuff and things like that," but no, I am Norwegian, Norwegian, Norwegian! In different ways, the extracts illustrate the asymmetrical power relations in the interview and how categories are often connected to stereotypic images of minority ethnic groups as other [15]. Though I sought to ask children to openly reflect upon ethnicity, bringing up ethnicity as a topic appeared to have different meanings for the children. For some children, their ethnicity was related to experiences of being treated differently (to an unarticulated Norwegian standard). In the interview situations, these different meanings became evident through the children's use of binaries, such as we/you or us/them. In retrospect, considering multiple interpretations of the interviewer-interviewee relationship and how the students navigated power structures in the interview situation provided insight into their contextualized meaning-making in a larger societal context. For example, in the case of Maya, one interpretation is that her statement was a response to how she perceived the researcher's ability, as a representative of the majority culture, to define her. Her response can also be interpreted as a resistance to power if she expected the researcher to devalue Persian culture in comparison to Norwegian. Maya's response reveals how the question evoked associations toward power relationships in a larger societal context, in which the meaning of the binary categories of Norwegian and Persian are locked and ranked. Additionally, in the interview with both Maya and Navid, the students addressed the researcher with "What do you mean?" and "If you think..." This direct confrontation alerted me of the ways in which participants could resist and challenge power structures by questioning the content of social categories.

Navigating the Generational Ordering in Interviews
In Project B, navigating the generational ordering in the interviews posed different challenges depending on whether the child was alone or with their parent. In the combined interviews, I had to navigate the triangle of communication patterns and negotiations between the child, parent, and myself. In these interviews, my attempts to structure the communication were complicated by the need to develop rapport with the child and the parent, while simultaneously, communicative negotiations were in play between the parent and the child. A less successful example is from an interview with Annabelle and her father: Father: Every year that went by, the distance between them [referring to Annabelle and her peers] grew and grew. She is barely in contact with them now. It's healthy for the class that there is someone a bit different-that everybody doesn't look alike. But sometimes it gets a bit rough. In the field notes written directly after this interview I wrote down feeling uncomfortable in the interview and inadequate as an interviewer. Annabelle was playful in the interview and was not particularly interested in follow the line of questions introduced. My concerns were that Annabelle, who was in a situation in which she had little control, experienced that we talked about her more than with her, and that the way we talked about her constructed her experiences of bullying and exclusion within a personal tragedy model of understanding disability. In this interview, I was left with a feeling that I had co-constructed yet another disempowering arena for Annabelle.
In the interviews, parents also added to the child's story if they believed that the child struggled to answer. Silence in interviews could be both powerful and painful. While the researcher often interpreted the children's silence as a thinking pause, parents seemed to interpret it as the child being uncertain of what to answer. This occasionally led parents to answer on behalf of their child. In some situation, parents even took control of the communication by acting as an interviewer as well as answering the questions directed at the child. The following example with Timothy and his father particularly illuminates several of the issues discussed: Timothy: I like to be physically active and things like that.
Interviewer: Yes? What do you like best in physical education?
Timothy: Ehhh ball games and dodgeball and games like that.
This interview demonstrates a situation in which the father took control over the communication and steered his son's story toward a narrative they had shared several times before. Some of the children and their parents had attended numerous medical/treatment interviews previously. Because of the familiarity with medical interviews, I had to work to get beyond the medical narratives that the families had told several times, while also honoring and supporting the stories shared in the interview [33].
Another challenge was the participants' avoidance of the topic of peer relationships in school. One case was an interview with a young girl and her mother. During the interview, the conversation ran smoothly, and both the child and her mother shared their stories of marginalization and exclusion in physical education. However, during the interview, I could sense that there was more to the story than what was told. Both the mother and the child were reluctant to share stories of peer relationships and friendships. The interview was cut short by the girl leaving to attend a physical activity session, and the mother and I continued the interview alone. After the daughter left, I introduced questions around peer relationships and social exclusion, and the mother narrated a maternal story of seeing her child being more and more socially excluded within her peer group. Analyzing the data, the researcher noted several contextual factors that could have influenced the situation. Identity projections are always socially situated, and what a person says is contextually bound. Discussing the diverse experiences of being a parent when the child is present might encourage enactment of two different and even conflicting socially situated identities. Enactment of conflicting identities could explain why, in some interviews, the children and their parents avoided speaking directly about peer relationships when both the child and parent were present. As with the parents, the children also protected their parents from sensitive information during the interviews. Sensing this protectiveness, I was reluctant to contribute to a conflicting situation for the parent and child by asking them to reflect further on the child's difficulties while they were present when it was clear that the child struggled socially at school.
I knew prior to the interviews that the setting would pose various challenges. However, by analyzing the data with relational ethics of categorization in mind, it become clear that not enough reflection was given to the implications of relationships and communication in the interviews.
Foreseeing some of these challenges might have helped me to prepare and navigate the generational ordering better or to construct a caring environment that allowed for more detailed descriptions from both parents and children.

(Re)constructing the Stories-Ensuring Anonymity
Contextualizing individual experiences is essential in intersectional research, as is seeking to understand the complex ways in which multiple identities shape opportunities and constraints in interactions with others [26]. However, retelling the rich stories of lives in context challenges researchers in terms of ensuring anonymity. This is especially the case in ethnographic work [48]. In Project A, emphasizing diversity in order to break with stereotypes and homogenous narratives regarding physical education and ethnicity was central to the purpose of the project. However, in research conducted in a defined environment or institution, chances are high that participants recognize both themselves and others, making it "difficult to ensure that data are totally unattributable" [49] (p. 341). Changing a name is not enough. In Project A, there was also the risk that other students and teachers would recognize participants in the project, as there were few teachers involved and only two school classes in the sample. Furthermore, there was the risk that recognizability might extend beyond the institution (e.g., that individual students would be recognized by parents or within a community). For example, the Pakistani community in Norway is portrayed as an interconnected social network [50].
From an intersectional perspective, children's disabilities and backgrounds, such as ethnicity, culture, and religion, are important for research. However, these characteristics also make the participants more vulnerable. How can researchers meet the requirements of anonymity while simultaneously highlighting the different aspects of the participants? Reflecting upon their own research on why so few students pursue degrees in physical education teacher education, Flintoff and Webb [10] discussed how a small sample made it necessary to stick to broad generalizing categories for their participants, leaving out any individual viewpoints their participants held regarding identity. They reflected: "These decisions are compromises and are very much at odds with our theorizing of identities as multiple, fluid and shifting" [10] (p. 580).
In Project A, one of the ways the researcher dealt with the challenge of ensuring anonymity within a framework of intersectionality was through applying a thematic narrative analysis of the data, as this allowed the researcher to maintain the truthfulness and nuance of the children's stories throughout the analysis [43]. While analyzing the data, the researcher (re)constructed the stories as close to the lives of the children as possible to maintain their rich detail [30]. However, in preparing narratives for publication, the researcher went through the stories and chose to leave out some details or rewrite certain aspects (e.g., writing the occupation of parents in more general terms or changing the gender and/or number of siblings). The researcher also made thorough reflections regarding parts of the research where it was more important to safeguard anonymity than others (e.g., if it was conceivable that publication could have direct negative consequences for the participant in question) [30]. One example in the data was a child from a religious minority in Norway. Talking about the importance of religion in the family, the child stated that their parents held a liberal view on religion and that a sibling was an atheist. At the end of the interview, the child brought up this topic again, expressing the importance of this information not being connected to the family. In such cases, it is essential to be aware that some pieces of data cannot be published [30].

Discussion and Concluding Remarks
Although social categories are essential for the generation of knowledge regarding social inequality and marginalization [3], the use of categories in research is a highly contested practice and calls for researchers to engage in ethical reflection [1,2]. The current study provided examples of how researching diversity and inclusion creates situations requiring ethical considerations that cannot be solved through official guidelines. The article explored how singling out minority groups of children framed the research, how categories and power relationships were negotiated and navigated in interviews and fieldwork, and how, in the writing of this paper, categories were negotiated at the cross-road of intersectionality, relational ethics, and the procedural ethics of ensuring anonymity. These issues are particularly urgent within research related to health, the body, and physical education [9,10,17,19]. Previous research regarding peoples' experiences and beliefs related to health and physical activity has largely failed to consider the heterogeneity within certain groups of people [10][11][12]19]. Measured against implicit standards embedded in Eurocentrism and ableism, certain bodies or groups of children have been objects of stigmatization, marginalization, and exclusion due to their backgrounds or (dis)abilities [12][13][14][15][16]. As such, there has been a call for more research that challenges deficit discourses and negative stereotypes and explores the variety of stories among students of diverse backgrounds [12,19]. Yet, this call (and responsibility) raises ethical challenges.
Both of the projects presented in this paper involved challenges regarding the recruitment of children and ensuring their right to make their own informed decisions regarding participation. In Project B, these challenges related to how parents acted as gatekeepers. Participation in research exploring inclusion and exclusion of specific groups of children depended on participants' inclinations toward collective action to improve knowledge within that specific field. This again depended upon identification with the wider group in focus [51]. Some parents tried to limit the focus on their child's disability, thus resulting in them not discussing the disability with their child. One reason could be that the parents sought to reduce the social stigma coming with a disability label [8]. Another reason might be that the parents did not identify with that label of their child. Similarly, some children may have refused to take part if they felt singled out because of their impairment or if they themselves did not identify with the disability label [15]. Recognizing some of the reasons why parents might refrain from allowing their child to participate could help researchers to attentively design the study and inform parents about the intentions with the research, which in this case was to illuminate some of the concerns the parents expressed.
In Project A, the researcher experienced ethical dilemmas of entering a field where not all children gave consent for participation. Scholars have raised concerns regarding how negative representation in research might be reinforced as a result of some groups not wanting to take part in studies they perceive as reinforcing their otherness, that are on the premises of the majority population, and in which they do not recognize themselves [1,8,52]. In regard to Project A, it could then be asked if participation should always be voluntary as a starting point [1]. According to Ellis [24], researchers must ask themselves what the greater purpose of their research is and consider whether it justifies the potential risk to others. As defenders of social justice, researchers have a responsibility to challenge marginalizing discourses, such as bodies at risk, or normalized absence/pathologized presence [7,10,15]. As such, the authors agree with Boddy [1] that there is no single best approach; rather, from a relational ethics perspective, it might be necessary to challenge how researchers consider voluntary participation in some situations. However, it is important that in such cases, researchers spend time building trusting relationships with participants and practice consent as a process [30].
In addition to providing researchers with tools to improve their sensitivity toward how children's everyday life experiences are situated in relationships of power, intersectionality can also inform how relationships of power are negotiated in interview situations [5]. Relational and mindful ethics advocate that researchers consider the complex stories of the people in their studies, as these stories and lives are important to understand the relationship in the interview [33]. In both studies, the authors experienced how social categories framed the stories of the participants, how stories were "clustered around some hegemonic constructions of boundaries between 'self' and 'other' and between 'us' and 'them,'" and that these relationships were "closely related to political processes" [50] (p. 2). Researchers and participants need to continuously attempt to resolve misunderstandings that might appear during interviews [31,33]. Reading emotions and relational cues in interviews and acting upon them might mean that researchers cannot push through with certain questions, even though the answers would be valuable to illuminate the research question at hand [33]. Furthermore, intersectionality rejects the binary thinking that one is only or always included or excluded. Applying multiple interpretations within an intersectional framework can redirect the researcher's focus toward children as active agents and how they resist and negotiate power relationships while still recognizing their vulnerability.
Scholars in the field of physical education have called for more research regarding inclusion that extends beyond single issue approaches and illuminates the richness of children's stories, however, this creates great challenges in regard to anonymity. To navigate this challenge, it is crucial that researchers base their studies around ethics of care. Researchers should seek to handle data in ways that keep stories nuanced and truthful. However, while it may be appropriate to write the stories, not all aspects of data can or should be published [30].
Scholars have a responsibility to challenge homogenous and essentialist understandings of categories [10] and how these understandings influence people's lives. As Gunaratnam [3] stated, there is a need to work both with and against social categories [3]. Discussing racism and ableism necessitates categories, yet categories (re)produce lines of difference. By negotiating and navigating the use of categories in this research, the authors experienced the power relationships entangled within categories and how categories can be used for political means. While procedural ethics might lull researchers into thinking that their studies are ethical, working within a relational ethic perspective calls attention to the need for self-reflection on the researchers' roles, motives, and feelings [33]. At the heart of these reflections regarding the ethical dilemma of categorization was the fear that this research would reproduce rather than confront and challenge marginalization and social inequality. Our article contributes to the literature on ethics in qualitative research. The study has limitations in that it was constructed after the initiation of the two projects; as such we did not generate data specifically for the questions raised in the article. Yet, seeking to make the right choices, we agree with Ellis [30] that researchers need to engage with each other's stories from the field. By sharing stories from research projects, these collective experiences might help researchers to reflect critically on how to use categories ethically in their studies.