Defining, assessing and minimising potential harm to children from research involvement is a critical and, at times, difficult task. Harm may be invisible and elusive, subjectively defined, and complicated by differing viewpoints and short- and longer-term outcomes (Alderson & Morrow, 2011 (new edition of the book, 2020)), with tension arising between the underlying ethical considerations. In some studies direct harm may not be anticipated, however, there may be some risk and/or discomfort for participants, whether the project concerns the testing of a new drug or interviewing children about traumatic experiences (H. Fossheim, personal communication, December 14, 2011). Similarly, potential benefits are also hard to define, uncertain and difficult to accurately assess as being a result of research participation (Alderson & Morrow, 2011 (new edition of the book, 2020)).

A best-outcomes, consequentialist approach can be applied to assessing and reducing harm and promoting benefits. In this approach actions are considered right or wrong depending on the nature of their consequences (Gallagher, 2009 (chapter in Tisdall, Davis & Gallagher, 2011). Accordingly, decisions in research may be guided by what are the best expected outcomes, with the anticipated benefits needing to be bigger, and more likely to occur, than any possible negative consequences. This can be interpreted as the right actions being those that result in the greatest overall good for the greatest number of people. An additional factor in the equation is the harm that may occur if the research is not done.

However, a challenge arises using this approach in the potential conflict between the ‘greatest good for the greatest number’ and the consequences for individual children participating in the research. From a rights-based perspective, drawing on the general principles expressed in the UNCRC of best interests of the child (Article 3.1) and protection from all forms of violence (Article 19), researchers have a responsibility to ensure that no harm is done to individual children participating in the research, as well as attending to the long-term and/or larger good for children as social groups.

A further challenge to applying the best-outcomes approach lies in the reality that there may be several outcomes to an action, which include both positive and negative consequences for the individual child. For example, the means utilised to achieve an outcome, such as an interview, may have retaliatory and negative effects for the child, while the future intended outcome, such as access to a service or heightened attention to a severe social problem, could be deemed positive. When the focus is on research outcomes, this can mask the potential for more complicated and heterogeneous consequences related to the tools and processes of data collection. This strongly suggests a need for the means of the research, that is, the research process, to be the key aspect under scrutiny. This topic is addressed in Mary Catherine Maternowska’s case study of a survey on violence against children, undertaken with participants aged 13 to 17 years.

Individual children may be best placed to assess any risks to themselves, which is one of the reasons why informed consent is so important (Laws & Mann, 2004). However, researchers are likely to have a greater appreciation of the potential wider and long-term implications of research participation, and therefore increased responsibility to assess risks posed to children through their involvement in research. Seeking local or specialised advice is important for researchers, when they are in cultural contexts other than their own or working with particular groups of children, in order to have a better understanding of the potential risks. Researcher responsibility is further emphasised when the children involved are particularly vulnerable, such as children separated from family in emergency and transition settings.

Researchers are responsible for protecting research participants from any physical, emotional or social harm that might result from the research (Ennew & Plateau, 2004; Laws & Mann, 2004) and must do everything possible to anticipate any potential adverse consequences. The minimisation and, in some instances, elimination of potential harm can be achieved by ensuring that it is identified and that ethical issues are reflected on from the very outset of the project, as well as throughout its implementation. Once identified, researchers can make efforts to minimise or eliminate any potential risks of harm, distress or discomfort, including incorporating practical measures. For example, certain geographical locations or socio-political contexts may carry specific risks which researchers can develop strategies to manage. In areas where armed conflict or peace-building are occurring, this might include researchers ensuring that children do not travel alone, late at night or in unsafe areas where there are landmines or other dangers (Feinstein & O’Kane, 2008).

Other areas for consideration include those in which the potential harm may be less clearly obvious or easily resolved. Reflection on the multiple possibilities requires knowledge of the local context in relation to the research process and/or topic. There may be consequences for children when they are involved in research which is consciousness-raising. Such consequences may occur, for example, where the research raises awareness of labour rights for children in situations where the child’s employment is a critical source of sustenance for the family. In another example, there may be opportunity costs if children are removed from class to be involved in school-based research.

As well as knowledge of the local context, it is critically important that researchers are knowledgeable in relation to the characteristics of the sample population with whom they are conducting the research. For example, if children from minority groups are involved in the research then researchers (depending on their own associations) may need to consult appropriately within those communities in order to identify potential harms or adverse consequences. Similarly, if the research involves children with particular disabilities it would be essential to identify potential harms to ensure children’s equitable access to research participation. Reflecting on the possibilities can help prevent or mitigate damaging follow-on consequences and ensure that researchers have information or referral sources on hand if required.

Research methods can contribute to minimising or exacerbating harm. In order to do no harm it is vital that the research methods are scientific and reliable. This underscores researchers’ obligation to ensure that children’s views are accurately reported. It can be argued that unsound designs or inadequate sampling methods are unethical, since they take up time and effort but will not produce satisfactory results or because they may unintentionally lead to the exclusion of groups of children. Reliable methods include the use of several well-validated research tools that have been adapted to match the capacities of the children involved as well as the issues being researched; at least two steps in data collection with adjustment of the tools and/or the research plans in between; and triangulation. Large scale surveys, especially when children are not involved in the development of the survey questionnaire, and methods such as direct interviews and focus group discussions, may not be tools that are the most adequate or appropriate in research with young children.

The predictability of risks may differ in different research paradigms. For example, in biomedical research, it may be necessary to develop medicines that are trialled on child samples, as extrapolating adult data to children is inappropriate (Yeung, 2007). It can be difficult to quantify the risk for participating children and consider these in relation to the benefits of developing medicines for children more generally. If the nature of risks is unable to be predicted reliably, then providing accurate information to children and their caregivers to enable informed consent for clinical trials becomes an essential ethical challenge.

A research situation in which it is challenging to predict the risks and likely outcomes is one in which there is a possibility that indications of previously undetected conditions or abnormalities may be discovered incidentally during the research process. The revelation of such chances and consequences would cause added concern to the child and the parents, and therefore requires great sensitivity. A number of risk-related issues raised in such situations, requiring sensitive consideration, are discussed in the ERIC case study by Sebastián Lipina.

The use of new technologies in research, such as the internet, adds another dimension to conceptualising and minimising harm. Children may have unrealistic expectations of privacy, individual information may be more readily accessible, participants can provide false details or maintain fake identities (disguising adult voyeuristic or harmful intent) and the distinction between public and private domains is blurred (Alderson & Morrow, 2011 (now updated to a 2020 edition); Lobe, Livingstone & Haddon, 2007). Certain safety features, as used on websites, can be built into internet-based studies. However, it is essential that these are appropriate to children’s evolving capacities and/or that such built in safety features are enabled by default as evidence shows that children struggle with user tools, safety devices, privacy settings and reporting mechanisms (O’Neill, Livingstone & McLaughlin, 2011). In addition, new modes of access to the internet are increasingly mobile, with children going online via their own laptop, mobile phone, ipod, ipad or handheld device, and thus less open to adult supervision. While this may contribute to children’s privacy from parents or others in the environment during research participation, it also underscores the need for children to be aware and empowered with regard to internet safety risks (O’Neill et al., 2011).

In addition to the importance of identifying potential harm, to aid efforts in minimising it, a further key aspect is attending to the resource aspects of research initiatives. It is absolutely critical that ethical considerations are taken into account in the planning stage of research, itemised in the project budget and factored into the timeline, in order for ethical principles and practices to be implemented throughout each phase.

Any research involving children may incur distress and researchers need to be prepared for this. It is especially likely to be an issue when the research topic may be upsetting for children, perhaps because it is a reminder of painful experiences, or when children are particularly vulnerable. This may occur, for example, when children have been abused, maltreated, or are separated from their parents as a consequence of humanitarian emergencies or death. A fundamental concern in research involving children who have been abused or maltreated is whether they might be seriously disturbed, or even re- traumatised, by the research process. In such studies researchers can incorporate specific measures into the research design to ensure children’s protection and care, such as ensuring the child has the support of the primary carer, using a child-centred methodological approach (Mudaly & Goddard, 2009) and incorporating child-friendly debriefing processes. Where appropriate, some studies may involve recruiting children who have had prior access to therapeutic studies, which may also be a protective factor for children.

In addition to sensitive topics or situations obviously more likely to cause upset, researchers will not always be aware of topics that may be distressing for individual child participants. Seemingly innocuous questions or situations may trigger strong and unexpected reactions. Children may be caught off-guard emotionally by unintentionally disclosing too much. It is therefore important that researchers know how to respond appropriately in the face of children’s anxiety or distress. Researchers should be prepared in advance for the kind of emotional responses that children may have and seek to differentiate harm from discomfort and/or research-engendered distress.

Key considerations in deciding about research activity in the light of risk and harm assessment might include: the expertise and skill of the researcher and members of the research team; the level of supervision and support required from senior researchers and experts in the relevant area; and the availability of follow-up support for children. Considerations in relation to dealing with distress are discussed in the ERIC case study by Elsbeth Robson and Ruth Evans, as they reflect on interviews conducted with young caregivers in the context of the HIV epidemic in Zimbabwe, Tanzania and in the UK.

It is also important to note that some of the worst psychosocial harm from an interview may not be apparent during or immediately after the interview session. One suggestion is to incorporate a post-interview check-in by someone who is a known, trusted support, familiar with the child’s usual behaviour, and can read the signs and find additional help if needed.

Researchers need to ensure that support is available to children, if needed, during and after the research process. An important part of research planning is anticipating potential adverse consequences of involvement in research and ensuring that arrangements are made to address these (Schenk & Williamson, 2005). It may be appropriate, in some studies, for researchers to help equip or support children’s parents or carers to respond to emotional reactions their children may have. Psychosocial support or counselling may be required for children to help them cope with the consequences of recalling distressing experiences or feelings. Other forms of support, such as health or welfare-oriented support services, may be needed to help children address difficulties that become apparent during the research process. Researchers’ responsibilities to do good and to do no harm to research participants requires them to plan for such eventualities. It may be helpful to have explicit protocols for addressing potentially concerning situations, and for making referrals (WHO, 2011) and providing a list of services that can provide support to children and families.

Referring children and families to support services requires an awareness of available resources. Most likely this involves consultation and gathering information within the local community context to ascertain organizations or agencies that are available and suitable to offer skilled support. Researchers can seek to engage with the local support service working with children and children’s issues if any, by seeking their services if required, informing them of the study and getting their local intelligence on the matter. In some contexts identifying or accessing child-focused services may be impossible for a range of reasons – support agencies may not have the capacity, they may be geographically removed or be inadequate and likely to cause more harm than good – in which case alternative strategies will need to be explored.

In the absence of child-focused services it may be appropriate to request help from organizations offering similar support. For example, if researching violence against children there may be local organizations supporting women affected by violence who are able to adapt some services to meet children’s needs (Laws & Mann, 2004). Research participants may also draw on individual resources, local culturally relevant mechanisms of support, and strong local community or faith-based organizations and groups (Ruiz-Casares, 2013). If support services are not identifiable, and the need for support is likely, the decision to undertake the research should be reconsidered.

The existence and nature of follow-up services can impact on researchers’ decisions about reporting suspected child abuse. Some researchers argue that the risks to children and their families in reporting maltreatment, such as harm to children, unnecessary humiliation and stigma (if allegations are unsubstantiated) and lack of follow-up services, may outweigh the benefits (Kotch, 2000). The challenge regarding a breach of confidentiality in the light of suspected abuse or safety concerns is further complicated in some contexts where there may be a complete unavailability of services to afford child protection or attend to psychosocial or mental health consequences. For example, in some contexts, particularly in developing countries, there are no government sponsored child protection services, or there is a limited set of social services to which children and families can be referred and children may be mistrustful of social services (Abebe, 2009; Hutz & Koller, 1999; Veena & Chandra, 2007). Furthermore, it may not be in children’s best interests to report abuse, as reporting may increase their vulnerability with corrupt figures in authority (Young & Barrett, 2001).

Local consultation is critically important and researchers need to have a sound understanding of the legal requirements, suitable referral sources and alternative resources within the area in which the research is being conducted prior to commencing it. As noted above, and discussed in the ERIC case study by Mónica Ruiz-Casares, researchers can explore alternative strategies in contexts where identifying child-focused services is unlikely.

In addition, the skill and responses of researchers in managing safety concerns are important, and can help children cope emotionally and facilitate transitions to the necessary services. Ideally, research staff should have training in child protection, particularly with regard to recognising and managing concerns and, where possible, work in partnership with experienced researchers and local child protection experts. Ethical supervision allows researchers to discuss concerns, seek guidance and debrief, as well as providing a forum for accountability [this is discussed further in the Researcher Support page].

Reporting concerns about children’s safety is not always straightforward. As noted above, there may not be suitable or adequate services available. Child protection services may be inadequate, heavy handed or dreaded by children who have had previous experience or knowledge of them. Children may already be accessing such services and may disclose the inadequacy of these to the researcher. Reporting can also be further complicated in situations in which children are living in the care of the state or a non- government organization and disclose institutionalised punishment or violence from their carers. If children are already receiving care, reporting may put them further at risk.

Research in areas where there is an increased likelihood of children being at risk, such as research with children who are sexually exploited, may be compromised by children not engaging in the research process if they expect such safety concerns to be reported by the researcher. An additional challenge for researchers in some contexts are the legal harms that can arise for children if they were brought to the attention of the authorities. For example, children who are stateless or asylum-seeking and those engaged in criminal practices, can be at risk of arrest, deportation, detention, corporal punishment and attainment of criminal records. One way some researchers manage these concerns is by ‘raising the reporting bar’ so that only really extreme cases of immediate risk would be reported.

In many cases, researchers may not be aware of what their actions might complicate or precipitate for a child, especially if there is little recourse to effective protective systems. There are no easy answers to these dilemmas, but it is critical that researchers are aware of the possibilities and plan responses in advance of conducting the research, which take into account strengths and limitations of the local context, the services available, the child’s evolving capacities and individual situation, and the potential impact on them of reporting safety concerns. It is also possible that researchers stand to learn from children themselves about the potential risks and harms of them being involved in a study. A preliminary discussion with a group of children in a safe environment may elucidate ethical issues that only children perceive, such as bullying or stigmatisation or unwanted attention from public security officials.

In some research contexts tension exists between protecting children from harm and respecting their right to participate in research. Researchers’ capacity to include children in research is particularly restricted if the research topic is considered sensitive (Powell et al., 2011) and children are also less likely to be given the opportunity to participate if they are considered particularly vulnerable (Powell & Smith, 2009). In addition, children may not be included in research for pragmatic reasons, such as lack of skilled personnel to undertake sensitive research with children and lack of adequate funding. The importance of protecting children from harm is ethically mandated by the philosophical principles of non-maleficence and justice, and reflects children’s rights as expressed in the UNCRC (specifically Articles 3.1, 3.3 and 36). However, a strong protectionist discourse denies children the right to participate and express their views on matters of concern to them (Powell & Smith, 2009) and children may be ‘gate-kept’ out of research on the basis of potential risk (Graham & Fitzgerald, 2010).

The tension between protecting children from harm and respecting their right to participate derives in part from the underlying conceptualisations of children held by different disciplines and across differing contexts. Ethical debates in the clinical sciences have tended to highlight protection of children from risk or direct harm, whereas in the social sciences debates have focused on children’s exclusion from research and argued for greater inclusion as a means of addressing power imbalances and ensuring that children’s voices are heard (Dixon-Woods, Young & Ross, 2006). To the extent that it is possible, children should be consulted about actions that may best balance their participation and protection rights, in accordance with the principles of justice, beneficence and respect.

Discrimination or bias in children’s participation in research may come about as a consequence of the power relationships that exist in society and between children. Researchers are not immune from such power relationships or societal influences. Their own values, beliefs and experiences impact on children’s equitable participation in research. Therefore, gaining insight and understanding into their own assumptions, as well as the social, community and family representations around such characteristics as age and gender, is critical in ethical decision-making and can be greatly facilitated and enhanced through ethical supervision and training.

In some contexts, where there is acute social differentiation, certain children may be silenced by others. Children from linguistic minority groups, for example, may face discrimination from other children. This presents a challenge to researchers during recruitment, identifying children who may potentially be excluded through discriminatory processes, and in data collection, particularly in focus group situations, ensuring that all children’s voices are heard and realities are presented.

The methods used may also make it difficult for some children to participate. For example, children who are hearing-impaired may require additional support, such as the engagement of a skilled interpreter, to facilitate the process of informed consent and ongoing communication through the research process. In another example, online research may be difficult or impossible for children who do not have convenient access to internet. Researchers using online methods are encouraged to consider whether digital exclusion may reinforce or further compound social exclusion. Participatory research should confront and not reinforce existing patterns of discrimination and exclusion, encouraging groups of children who typically suffer discrimination and are often excluded to be involved (Feinstein & O’Kane, 2008). Additional ethical issues are raised when researchers become aware of discrimination, exclusion or unfair social incidents and experiences affecting children while conducting research. Researchers then have to make difficult decisions related to accountability and advocacy, as discussed in Jude MacArthur’s ERIC case study, in which an incident of bullying and exclusion of a child with disability was witnessed.

In focus group research more subtle forms of “inner diversity” (Eurochild, 2011, p.7) may impact on children’s equitable participation, such as how they learn, communicate, solve problems, make sense of the internal and external environments, and manage sensory input. For example, some children listen best when they are engaged in a tactile manner so having objects or material for them to handle will enable them to listen more fully. In another example, “some children need to begin speaking in order to understand what they are thinking; others don’t speak until they have ‘gathered their thoughts’” (Eurochild, 2011, p. 8). Having an awareness of these differences can help researchers to develop and implement ways to address them, thereby facilitating the more equitable involvement of all focus group participants.

In RCTs and other experimental designs it is important to maintain random assignment to conditions, so that all children in the sample have equal opportunity to be selected in all conditions. Randomisation should be purely by chance, such as using computerised allocation or random numbers, and should not use any systematic characteristics, such as date of attendance at the clinic or social security numbers. In RCTs, the control group should be given the current best treatment, rather than a placebo, and be offered to go on a wait-list to receive the new treatment/s after the trial is completed and if the trial shows favourable results.

Children’s involvement in research design, as in child-led research, can be enormously beneficial to the research itself. It does not necessarily resolve issues of power and representation, as research with children is mediated by adults (Clavering & McLaughlin, 2010; Kellett, 2010). However, the meaningful participation of children and young people throughout the design and implementation of research studies can be a critical factor in the success of the project (Radford et al., 2011).

A related area in which children’s opinions have been sought very rarely and which needs to be redressed as a matter of some importance, is children’s views on ethics in research, including debates over ethical considerations such as those discussed in the ERIC compendium. Jurrius and Uzozie (2012) suggest that a good start to finding concrete methods to cope with ethics issues is the conversation about ethics between researchers and, importantly, between researchers and children. Considerations with regard to discussing ethics with children are the focus of the ERIC case study by Kitty Jurrius.