Valuing vulnerable children’s voices in educational research

ABSTRACT

Changing views on childhood and children’s rights entail an increasing demand for listening to children's voices, even in research. All children are, in principle, seen as vulnerable participants in research, but our concern is the particularly vulnerable children. By listening to them, researchers increase the chances of contributing to the improvement of the children’s total situations based on their own experiences.  In this article, after discussing why vulnerable children should take part in research and exemplifying how they should do it, we consider advice for enabling particularly vulnerable students’ voices to be heard in research in ethically justifiable ways. Recommendations for considering vulnerable children’s participation in research are proposed. While children may be regarded particularly vulnerable for a range of reasons, we draw attention to the student group in regular classes identified as having special educational needs by referring to examples from a recent research project in Norwegian schools.

KEYWORDS:

• Vulnerable children’s voices
• ethical considerations
• inclusive research
• special educational research

Introduction

Generally, children are increasingly seen as people with rights relating to who they are now, not only who they later become (Sommer 2006; Bae 2007; Frønes 2007; Nordahl 2010; Tangen 2011). They are no longer regarded primarily in terms of becoming but of being and as competent agents in their own lives (James and Prout 1997; James, Jenks, and Prout 1998; Frønes 2007). The perception of children as competent agents is a central feature of the changing paradigm (James and Prout 1997; Frønes 2007). Theoretically, this understanding relates to ‘the new sociology of childhood’ (James, Jenks, and Prout 1998) and manifests in research theoretically, methodologically, and in the choice of research topics and research questions. But, when it comes to research, there has been a long tradition of research on children which has treated children as objects rather than subjects (Bae 2004; Tangen 2011; Black-Hawkins and Amrhein 2014). From a psychological point of view, there is a growing understanding that children are, in fact, experts on being children (for instance Morrow 1998). Their world differs from the childhood of today’s adults, and besides, children perceive the world differently from grown-ups and communicate differently about their experiences. Therefore, ‘[t]o optimizes our learning, we have to find ways to listen to childen in their own time on their own terms’ (Dahl 2014, 595). At the same time, one should bear in mind that research is always situated and embedded; for instance, issues of gender, class and ethnicity are as relevant for children as for others (Richards, Clark, and Boggis 2015).

Children's generally increasing participation in research is one of the perspectives on which we base this article. Nind (2011) claims that there is a related paradigm shift in disability research regarding participation. Since we in our article are looking into vulnerable children’s participation in research, we include, as another perspective, research on and with particularly vulnerable groups, such as disabled people.

‘Researchers have a special responsibility to respect the interests of vulnerable groups throughout the entire research process’, according to ethical guidelines for research, as here from the Norwegian National Committee for Research Ethics in the Social Sciences and the Humanities (NESH 2016, item 21). Due to their ages, children are generally considered such a vulnerable group, and NESH emphasises that the researcher ‘must know enough about children to be able to adapt both their methods and the direction of their research to the ages of the participants’ (NESH 2016, item 14). When we in this article refer to vulnerable children, other vulnerability factors than age are also present. We understand ‘children’ as learners in primary, secondary and upper secondary education. (The United Nations’ Convention on the Rights of the Child defines every person under the age of 18 as a child).

Research with children must consider ethical issues throughout the whole research process. Perceptions of children affect all the research choices that are made and, even when children are respected as legitimate, important participants in research, researchers must also recognise that they may be more vulnerable, and have less institutional power, than adults (Hill 2005). Bae (2007) shows that relationships between adults and children always bear the mark of asymmetry, adults having the power of definition. Researchers have responsibilities to act with humility and respect in the face of children's statements and to highlight children's experiences and information without compromising them (Bae 2007). At the same time, research with children rather than on them presents major ethical problems (Solbakk 2014), as discussed below. Ethical research ensures the informed consent of both children and their parents NESH (2016). In practice, consent is informed when children understand the main aspects of the research process: the purpose of the research, the main features of the project plan, what they will be asked to do, and the possible advantages and disadvantages of participating (Mukherji and Alborn 2010, cited in Staxrud 2013).

Background

Vulnerability

The concept of vulnerability is widely used in educational research, as well as in education, special education and policy making, but it has rarely been theoretically defined or analysed. The same holds for the social sciences, social care services and in social work (Virokannas, Liuski, and Kuronen 2020). From a legal point of view, Fineman (2013) argues that vulnerability arises from our dependency and is both universal and particular. It is universal in being inherent to our human embodiment which means that we are always at risk of harm due to, for instance, injury or illness. At the same time, vulnerability is particular since some individuals face a higher risk of harm than others. Coyle (2013) has argued that the notion of vulnerability can beppropriated for different (and potentially contradicting) purposes, and that this can lead to an implementation problem: how exactly should vulnerability be identified and who should determine these factors of identification? Coyle claims that if any identification of vulnerability does not identify vulnerability on the basis of several dimensions, the notion risks becoming essentializing, paternalizing and/or victimising.

We agree with those who think one should be careful not to refer to the term vulnerability without giving it actual content, but at the same time we believe that overall, the notion seems to constitute a potentially useful tool for assessing the origin of disadvantage, exclusion or marginalisation at three levels: the individual (embodied factors), the immediate surroundings (situational factors) and the society as a whole (structural factors) (Flegar 2021). We are in this article unable to cover the underlying reasoning regarding this notion in-depth and will therefore focus on a few key areas related to vulnerable children's participation in research.

‘Special ed. students’ and vulnerability

Potential vulnerability factors in schools may depend on a range of circumstances, for example: being under the care of the child welfare service; having disabilities or physical or mental illness; having had traumatic experiences or different social or cultural backgrounds than the majority. Here, as examples, we focus on recipients of special education in inclusive settings in Norwegian schools, claiming that they are in risk of being particularly vulnerable. The special education system in Norway refers to the Education Act stating that children who do not, or cannot, benefit satisfactorily from ordinary education are – after expert assessment – entitled to special education. According to this formal procedure, barely 8% of students in primary and lower secondary school receive special education, some of whom are diagnosed with disabilities but the majority of whom are described as having specific learning difficulties, especially in reading and writing, or behavioural problems (Norwegian Directorate 2020). Except from a few in separate settings, most students who receive special education receive it in parallel with ordinary education for five to seven hours per week on average, preferably within class, but in reality, quite often outside. During the remainder of the school week, they follow the same programmes as their classmates.

The reason for regarding the ‘special ed. students’ as potentially vulnerable, is that Norwegian studies, as well as international ones, have shown that students who receive special education are more marginalised, both socially and academically, than other students (e.g. Messiou 2006; Haug 2017a; The Ombudsman for Children 2017; Buchner 2017; Nordahl 2018). For Norway it is stated in several studies that the selection of students for special education often appears as random (Haug 2017a). For instance, comparable students without special education are just as low achieving or struggling as the ‘statemented’ ones. Further, findings have revealed weak learning outcomes and determined that teacher competence in special education is often inadequate, and many teachers have low expectations of regular class students who receive special education (e.g. The Ombudsman for Children 2017). For these reasons, we argue that, as a group, children and young people who receive special education in inclusive settings are in – or risk finding themselves in – vulnerable situations. 

Vulnerable persons’ voices in research

The general demand ‘nothing about us without us’ has been voiced for years by disabled adults (e.g. Charlton 2000). Marginalised young people are also now raising their voices, asserting ‘everything about us, with us’ (EADSNE 2016). However, the notion of ‘voice’ is not unambiguous (e.g. Richard, Clark, and Boggis 2015). Conceptions of an authentic, autonomous children’s voice are contested (Mayes 2016). For one, there is not only one children’s voice in research, there is a diversity of voices, so Messiou (2019) insists on using the plural ‘voices’. To her, voices refer to ‘students’ thoughts and emotions, as well as their actions for bringing about change’ (769). This understanding comprises more than the linguistically/verbally expressed views, including not only the most academically successful students (Mayes 2016).

But, even if student voices are listened to, there may be barriers in taking them into account in schools. In their studies of student voices in New Zealand about what learning means to the young participants, Bourke and Loveridge (2016) also included teachers’ interpretations of what the children had said. Interestingly, most of the teachers ‘explored student voice responses in relation to their understanding of the New Zealand framework, not from a child’s frame of reference’ (Bourke and Loveridge 2016, 65). In order to translate children’s ideas back into teaching and learning, they need to be taken seriously, perhaps easier achieved if further research worked with students and teachers together, as suggested by the authors.

Studies claiming to be inclusive have the stated intent to improve the situation of vulnerable groups, for example Allan and Slee (2008), or Tangen (2009) who highlights that vulnerable school children’s research participation should help improve their ‘quality of school life’. Overall, the intention of participatory research is ‘to contribute to social change, that helps to create a society, in which excluded groups belong, and which aims to improve the quality of their lives’ (Walmsley, Strnadová, and Johnson 2018, 758). An example is a study aiming at optimising daily practices in educational settings (Messiou 2019): An inquiry conducted in many European countries looked at an innovative strategy for helping teachers respond positively to learner diversity. The strategy merges the idea of lesson study with an emphasis on listening to the views of students. The research suggests that it is precisely the involvement of students’ voices that makes the difference as to advice for improving practice.

Walmsley and Johnson (2003) proposed the following requirements for an inclusive research process regarding vulnerable groups, in this case disabled persons: relevance of the research; potentially improved life situation for the group in question; involvement of the target group in every stage of the research. The authors later claimed that inclusive research should bring added value which other research methods could not offer (Walmsley, Strnadová, and Johnson 2018) or, as put by Nind and Vinha (2012), answer questions that cannot otherwise be answered.

Universal Research Design (URD)

Researching with children who have a range of capabilities and skills, and sometimes impairments, requires special attention to access and usability. Generally, Universal Design (UD) refers to environments that ‘can be accessed, understood and used to the greatest extent possible by all people, regardless of their age, […] ability or disability.’¹ (Also see for instance Meyer, Rose, and Gordon (2014)). Universal physical access, perceptible information, for instance optional audio or written information, and simple and intuitive use are among the demands of universal design that are frequently mentioned. In a framework for Universal Design for Learning (UDL) the universal principles are applied in instructional environments. The aims are to provide multiple means of representation and expression as well as engagement in the classroom (Rose and Gravel 2009).

When dealing with the field of research, the following guidelines for a Universal Research Design (URD) were introduced by Williams and Moore (2011, 4):

(I) plan multiple options for people to learn about, respond to, and arrive at opportunities to participate in research; (II) provide multiple means to communicate the information in research instruments and instructions for participants; and (III) provide multiple means of responding to research instruments and interventions.

Even if Williams and Moore and other authors mostly have had young persons or adults in mind when URD is concerned, we think that the guidelines above make up useful advice when vulnerable children of varying age are concerned as well. Adaptations might for instance include individual dialogues with children during the research process, related to age and known learning obstacles, in order to ensure optimal understanding. Several examples in the discussion below illustrate how facilitation might be done. According to Peña et al. (2018) researchers should strive to apply principles of universal research design to all stages of inquiries, not only as in our case when vulnerable children are concerned, but in all research.

‘The SPEED project’ – an example of children’s (more or less) participation in a special education research project

The project ‘The function of special education’ (2012–2017) – SPEED, in short – was a national research project investigating special education in Norwegian primary and lower secondary schools. Several qualitative and quantitative methods were used in the SPEED project, the central one being a large survey in which children, teachers and parents participated. The authors of this article took part in different stages of the project. Methods and findings will not be reported in detail here; that has been done elsewhere, see for instance Haug 2017b; Haug 2020; Festøy and Haug 2018; Mølster and Nes 2018; Opsal and Tonheim 2018. See also information in the endnotes of the article.

In the aftermath of this research, we have become increasingly aware of the importance of children’s participation in the whole research process. The examples regarding process and results from SPEED in the following are included, not because they are necessarily admirable regarding the presence of children’s voices, but because they illuminate relevant issues of participation in a fairly typical conventional inquiry.

In the main part of SPEED, digital questionnaires were used² One strength of the project was that many children answered, commenting on a wide range of topics and enabling comparison between groups and with the adult respondents. All students of the class took part. They completed the questionnaire during a school lesson with a teacher present. The guardians had given their written consent to the research participation of their children, but it is not described how or by whom children themselves were informed and asked. Access was facilitated by translations of the questionnaire into 15 languages, including Arabic, Tamil and Thai. The extent to which students were given assistance in answering the questions is unknown. In addition to linguistic minority students, other students, such as those with intellectual, sensory or concentration difficulties or who struggle with reading, might have needed help too, or perhaps additional time, but we do not know if that occurred.

Findings revealed worrying data about the status of ‘special ed. students’ in inclusive schools. When looking into issues of loneliness, wellbeing and behaviour,³ the result was negative for the ‘special ed. students’, here of 5th and 6th grade (age 10–12), compared to their peers without special education (0.5 std. difference⁴). When split on types of registered student difficulties, the difference between students with behavioural difficulties and their peers was particularly large when wellbeing and behaviour were concerned (more than 1.0 std difference) (Haug 2020). The other difficulty categories used were specific learning difficulties, general learning difficulties and other difficulties.

Students’ attitudes to school subjects are here exemplified by the factor ‘My opinions on the subject Norwegian.⁵ The results regarding the fifth and sixth graders’ assessments of their relationships to this subject showed that the ‘special ed. students’ clearly valued the subject less than their fellow students (0.5 std difference) (Nes 2017).

For the sake of comparison, we include parts of what the teachers reported about the same fifth and sixth graders – with or without special education – referred to above. On the factor ‘Student’s engagement and work effort⁶ the responses revealed that the class teachers considered the engagement and work effort of ‘special ed. students’ to be much lower than those of other students. In the eyes of the class teachers, the difference between students with and without special education was huge (nearly 1.0 std). The difference between the two student groups’ ‘adaptation to class rules⁷ was equally large, as reported by the teachers (Nes 2017).

An interview and observational study in SPEED involving younger ‘special ed. children’ and their teachers and parents underpins parts of the results from the survey, revealing that teachers’ perceptions of the students’ work efforts were far more negative than the students’ own perceptions (Festøy and Haug 2018). The same qualitative study also showed a discrepancy regarding perceived student participation in decisions about content and methods in the lessons; contrary to what the teacher said, the students did not feel they had a say.

Regarding children's involvement in other stages of the research process than providing answers, their participation was more of an indirect kind. For instance, in the initial stages, the chosen questions and instruments were based on previous research regarding children and young people, some of them in vulnerable situations. The SPEED research group itself was interdisciplinary, and a reference group for the project provided input from various academic fields, but not directly from children, their families or interest organisations. The project description of SPEED referred to ethical challenges but not explicitly to the challenges faced when researching young ‘special ed. Students’. Dissemination of results was done in various traditional ways by the researchers via articles, reports, books, talks, newspaper articles, etc.

To sum up vulnerable children’s participation in the SPEED project, in terms of results we have first learnt how ‘special ed. students’ as respondents provide important messages about their situation in school, which is basically about how they in many situations feel socially and academically marginalised compared to other children. Second, we have shown that children can provide different answers than adults, in our case revealing how students who receive special education in part were regarded very negatively by the classroom teachers in ways that did not match the students’ self-perceptions. Further, the teachers felt that the children had a say concerning the lessons, while the students themselves did not feel it that way. However, in terms of participation in the research process, in several steps of the inquiry the children’s voices are not accounted for.

Discussion

With the SPEED project as a point of departure, we will discuss vulnerable children’s voices in all stages of research. Other examples of participatory research are also included. The SPEED project confirmed what other studies have shown, that students who receive special education are, in their own eyes, more marginalised socially and academically than their peers, and therefore at risk of being particularly vulnerable (cf. Messiou 2006; Haug 2017a; The Ombudsman for Children 2017; Buchner 2017; Nordahl 2018). At the same time, SPEED data indicated that ‘special ed. students’ may be rendered even more vulnerable by subtle markers of exclusion because class teachers view these students as much more marginalised than they do themselves. It seems that entering the special education system in itself contributes to marginalising some students. This occurs in spite of the fact that ‘special ed. students’ are strikingly similar to other students, since in many cases it is random who is stated as having special educational needs (Haug 2017a). That is, it is not the embodied or individual differences between the two student groups that are prominent (cf. Fineman 2013; Flegar 2021). Explanations must be sought on other levels, such as situational factors – for instance the teacher’s practice and views on ‘special ed. students’, or structural/societal factors (Flegar 2021).

When SPEED is concerned, we regard the inquiry as research based and as such relevant for the ‘special ed. students’, in accordance with Walmsley and Johnson’s (2003) requirements, even if the respondents did not have a direct say in choice of research issues or research instruments. But one may wonder what students’ own priorities would have been, if they themselves – or may be a reference group of children – had been asked (cf. Bourke and Loveridge 2016). Regarding informed consent, children’s role in the SPEED project is not quite clear. Bourke (2017, 232) underlines that informed consent is an ongoing process and that ‘reaching informed dissent is as important as consent’. Unconventional ways of consenting may be needed (Richards, Clark, and Boggis 2015).

While quantitative methods were crucial in SPEED, qualitative methods are often seen as most beneficial when the purpose is examining children's views and experiences, especially qualitative interviews but also a range of creative methods (Olsen 2012; Mossige and Backe-Hansen 2013; Bourke 2017; Haugen 2018; Alerby and Kostenius 2011). Examples include a variety of visual and digital approaches (Alderson and Morrow 2011; Dahl 2014; Cowie and Khoo 2017; Haugen 2018). One example is how children were invited to take photographs based on their understanding of the project research goals. In the interviews afterwards, the children could show and talk about their photos to the researchers (Cowie and Khoo 2017). Richards, Clark, and Boggis (2015) found inventive methods even when children had multiple and complex needs for adaptation.

In the SPEED project children acted directly as sources of data through surveys and interviews. If we had had only the class teachers’ opinions of the students, the study would have presented an even more disadvantageous picture of the children with ‘statemented’ special educational needs.  Regardless of the methodological approach, the ethical challenges when vulnerable children are participants deserve special attention. An interview situation with semi-structured or open questions can create great discomfort for participants if difficult, sensitive experiences are subjects of the interviews. Mossige and Backe-Hansen (2013) argue that, when using anonymous questionnaires, children and young people should be assured that no one will see their answers to individual questions. This is different in an interview situation, since students relate directly to another person who inevitably seeks to understand their answers. Interviews could also move in directions that interviewees do not anticipate, but interviewers can notice if interviewees seem uneasy and adjust the interview situation accordingly (Mossige and Backe-Hansen 2013, 59). In interviews as well questionnaires, questions should be posed about how the young respondents feel about the queries asked, for instance whether they feel relevant or not (Mossige and Backe-Hansen 2013).

In SPEED the young participants were not asked how they felt about the inquiry, but they were involved as important respondents, after all. Many researchers claim that precisely by dealing with children as data sources, their experiences are taken seriously (Gamst and Langballe 2004; cited in Haugen 2018; Lundy 2018). Research participation may be empowering for vulnerable children (Walmsley, Strnadová, and Johnson 2018). Trusting and respecting children's responses and perceptions means considering children as credible sources. The extent to which children – or adults, for that matter – correctly remember events or personal feelings is affected by context (Gamst and Langballe 2004). When you ask for children's subjective experiences, incidents, and perspectives relating to their school realities, their answers will be true at that moment, without this affecting the child's credibility (Bae 2004; Sommer 2006). 

Children’s participation in research as respondents, as in the SPEED project, is one thing, their contribution in collecting the data, as well as analysing them and disseminating the results are others. Schäfer and Yarwood (2008) exemplify how young persons themselves acted as co-researchers, conducting interviews with peers about their daily lives. The young volunteers were trained how to ask questions and follow them up as well as using the technical equipment. The eldest (up to age 16) conducted group interviews alone with a video camera. But the authors strongly warn that the described approach is no guarantee that hierarchical power structures will not hamper the research (Schäfer and Yarwood 2008).

Children’s participation in data analysis is far less widespread than in providing the data (Cowie and Khoo 2017; Nind 2011). But, if young vulnerable persons really are to be seen as co-researchers, taking part as a ‘sense-maker’ in analysis of data is essential (Nind 2011). You may even speak of research not only with children, but by them (Dahl 2014; Cowie and Khoo 2017). This occurs when children initiate projects and/or share power and responsibilities with adults (Hart 1992; Shier 2001). Real participation occurs when children, including the particularly vulnerable, are involved throughout the whole project and take part in the decisions that are made, see for instance an example with use of puppy production by Mayes (2016). Importantly, young persons, and particularly the vulnerable ones, will need training and support to exert such roles; some will need a lot (Mayes 2016).

In the SPEED project dissemination of results was undertaken in traditional ways by the adults. As an example of children’s participation in dissemination of research findings, Cowie and Khoo (2017) explain how results were presented by the primary school-aged participants themselves on a ‘community sharing day’. Family as well as the school and the wider local community could watch and listen to the childreńs own power point presentations from the project. In other of these authors’ projects, results included photo books with children’s texts. However, the authors highlight that ‘[a] key challenge is how to present children’s views and images while attending to their privacy, safety and dignity […] particularly if the child’s identity is clearly visible and they are potentially construed as vulnerable through the accompanying text’ (Cowie and Khoo 2017, 241–242).

Conclusions and recommendations

We sum up our journey regarding vulnerable children in research by suggesting recommendations for researchers aiming to increase vulnerable children's research participation in ethically justifiable ways. In addition to researchers, addresses for the recommendations are also the bodies that exert quality control of, and approve, research proposals. Our concern is to value and hopefully increase children’s voices in research, with an overarching goal of ultimately improving the circumstances of the vulnerable children, for instance by informing teaching practices.

Recommendations

The following recommendations for participatory research with vulnerable children are not intended as a checklist, but as issues for reflection and increasing awareness.

• Research with vulnerable children must aim to answer questions that cannot otherwise be answered.

  A premise that should underpin all research involving vulnerable children is that one should only involve them in the research when the knowledge one seeks does not exist elsewhere.

This recommendation is adapted from Nind and Vinha’s (2012) requirements for researching with intellectually disabled adults. Walmsley, Strnadová, and Johnson (2018, 752) emphasise that these research participants should ‘bring something different and unique which brings added value to a research project’, a premise that in our opinion holds for research involving vulnerable children too (Tangen 2010; Dahl 2014). However, due attention has to be paid to the difficult balance between, on the one hand, the needs of protecting the children and, on the other, the need for research-based knowledge. The more vulnerable the children are regarded; the more a thorough consideration of their protection is required. 

• The research must be relevant and in the best interest of the vulnerable children.

  The research should contribute to providing a better quality of life for children and young people who are particularly vulnerable. It must be considered whether and how vulnerable children themselves can influence the research issues.

One of Walmsley and Johnson’s (2003, 64) research requirements was that research should ‘further the interests of disabled people’. Essential to the approach we have adopted in this article has been to highlight that vulnerable school children’s research participation should help improve their circumstances (Allan and Slee 2008; Dahl 2014; Walmsley, Strnadová, and Johnson 2018; Tangen 2009). The SPEED project, and most other school research relating to vulnerable children, have had the stated intention of contributing to the improvement of the daily lives of the children in question, such as improving the quality of special education and ordinary teaching. 

Furthermore, an underlying assumption in SPEED was that improvements for the most vulnerable would benefit all. This ‘enrichment perspective’ was formulated by Befring (1997), who claimed that schools that are good for disabled and other vulnerable students are good for all students. Another of Walmsley and Johnson’s (2003) requirements for inclusive research with disabled adults was that the research problem must be owned by disabled people. Although this point may have limited relevance for some in our target group, we find that, in principle, vulnerable children should influence the choice of research topics (Dahl 2014; Tangen 2011). In order to have their voices heard, not only children's age but also their specific needs for adaptation may need consideration. In parallel, NESH (2016) requires that the researcher must have sufficient competence about children and vulnerable groups to be able to adapt the research to the participants. 

• Vulnerable children should participate in the whole research process in ethically justifiable ways.

  Children and adolescents who are particularly vulnerable must, when relevant, have the opportunity to participate in the entire research processes of studies that apply to themselves. This has to take place in ways that are ethically justifiable and age-appropriate. The children’s needs for adaptation should be met in accordance with principles of universal research design (URD).

This recommendation is consistent with Walmsley and Johnson (2003), who state in their requirements for inclusive research with disabled people that they should be part of the whole research process, even in designing and managing the research. We recommend that ways of facilitating also the participation of vulnerable children in the whole research process should be considered. In planning a project and applying for funding, as well as when publishing from it, children’s participation in the research ought to be accounted for. However, neither SPEED, nor related research that we know of, have involved children in project design or management, although various interests have been considered in connection with the composition of research groups and steering groups.

When it comes to research methods, they must give vulnerable children opportunities to have their voices represented adequately (Schäfer and Yarwood 2008; Tangen 2009; Mossige and Backe-Hansen 2013; Messiou 2019; Bourke 2017; Haugen 2018). Vulnerable children’s feedback on their experiences of being involved in studies should also be included. The young participants might be asked whether the questions in the study addressed topics they perceived as important, whether other issues should have been included, and whether the questions were unpleasant to answer (Mossige and Backe-Hansen 2013).

The researcher must show respect for human dignity in the choice of topic as well as informed consent towards those who participate in the research (NESH 2016). Regarding data collection Tangen (2014, 686) asks ‘what steps should be taken to secure autonomy, privacy, and wellbeing and minimize risk’ while Nind (2011) discusses how young or disabled persons can participate ethically in data analysis. Likewise, Cowie and Khoo (2017) draws attention to possibilities and challenges regarding vulnerable children’s contribution in the dissemination phase.

Concluding remarks

In this article, we have addressed the necessity of listening to vulnerable children’s voices in research and exemplified how to do it. Young age is seen as a vulnerability factor in itself; here we refer to situations where other vulnerability factors are present as well. The ways to include children are numerous; we have just mentioned a few examples. At the same time, there will be more or less dissonance between the ideals of participatory research and practice, as we have seen. However, we would argue that conventional research with vulnerable children can still be of value, even if it does not follow all advice slavishly. The main point is to enter the path of increasing vulnerable children's contribution in the research process. In parallel, Lundy (2018) claims that children's right to participate in collective decision-making is central, but that it is demanding to do so at all levels, and not always beneficial for the participants. But, she argues that in many cases it may be more important to implement children's participation in parts of the decision-making process; ‘something is better than nothing’ (Lundy 2018, 344).

According to Alderson (2000, 243), ‘[t]o involve all children more directly in research can rescue them from silence and exclusion, and from being represented, by default, as passive objects`. We agree with Fottland (2000), who researched vulnerable children, that whether one should facilitate research that actively includes the children's perspectives is basically a value issue. An alternative question might be, ‘Do we have the moral right to not do so?’ (Fottland 2000, 31, our translation). To increase research participation for vulnerable children, however, further research is needed, not only to widen the range ofmethodological approaches and research participants, but also to develop research ethics and reflections on the problems raised by such research (Tangen 2008; Richards, Clark, and Boggis 2015).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Ann-Cathrin Faldet

Ann-Cathrin Faldet is an associate professor in special needs education at the Inland Norway University of Applied Sciences. Her field of research is youth and behavioural problems, risk- and protective factors, and her theoretical grounding is linked to the socio-cultural theory and agency perspective. She is currently researching how young adults who have exercised violence in adolescence looking at their own school history.

Kari Nes

Kari Nes is a professor emerita in education at the Inland Norway University of Applied Sciences. Her field of research is inclusive education and special education. Current projects include ‘push- and pull-out phenomena in inclusive school systems’, which is a comparative study of teachers’ views of such phenomena in Italy and Norway.

Notes

1 http://universaldesign.ie/Built-Environment/Building-for-Everyone/9-Planning.pdf.

2 The number of respondents and response rates varied in different parts of the project. The response rate for students (N = 16,282) was 80%, and 73% for their class teachers in the survey to which we refer (Topphol, Haug, and Nordahl 2017). All translations of instruments and results in the SPEED project are ours.

3 Survey data from a total of 5100 students from the fifth and sixth grades have been analysed. Factor scores refer to groups of questions. The factor ‘social isolation’ had two statements: ‘I feel depressed at school’ and ‘I feel lonely at school’. The factor ‘wellbeing’ had seven statements, such as ‘I enjoy going to school’ and ‘I like it in my class’. The factor ‘behaviour’ had eight statements, such as ‘I quarrel with other students’, ‘I protest if the teacher annoys me’ and ‘I fight with other students’. These factors were entirely or partially based on Sørlie and Nordahl (1998), and Rutter and Maughan (2002). The students ticked off options on a four- or five-part Likert scale. Differences between groups are expressed in effect size, i.e. differences between standard deviations.

4 The statistical analyses used were factor and reliability analyses with satisfactory results, as well as frequency and variance analyses (Topphol, Haug, and Nordahl 2017).

5 The questions about the subject Norwegian are based on the project ‘Quality in education’ (Haug 2012).

6 The questions for the factor ‘engagement and work effort’ are based on Skaalvik's engagement scale (Skaalvik 1993). Teachers were asked to rate each student regarding the following issues: ‘The student's engagement for success at school is … ’, ‘The student's work effort at the school is … ’, and ‘The student’s interest in learning during classes is … ’.

7 Questions for this factor include ‘pays attention to teacher’, ‘is tidy’, ‘is not distracted’, etc.