The Child’s Best Interest Principle across Child Protection Jurisdictions

Governments have a range of steering mechanisms and incentives to guide and rein in professional use of discretion. In this study, we examine the contested ‘best interest’ principle. We study child protection legislation, providing information about how governments set a standard for decision-making about the best interests of the child in intrusive interventions. The empirical focus is the formulation of the principle of the child’s best interest as formulated in child protection legislation in 14 countries. The results reveal significant differences between countries with regard to the degree of discretionary power vested in decision-makers, the factors considered and the strength of the principle compared with other principles. The countries’ wording of the best interest principle provides direct information about government views of their responsibilities for children in need of protection and the ways in which decisions about their interests should be made.


Introduction
In social science, a core challenge is to understand the uses of government power in professional practices in which exercise of discretion is essential. The aim of this chapter is to examine how governments in 14 highincome countries simultaneously interpret the vague and indeterminate principle of the best interest of the child in legislation and instruct professional decision-makers. Legislation is an important mechanism by which governments state their goals and ambitions and signal and instruct professionals on how they wish their goals to be implemented in the various institutions of a welfare state. Such signalling is typically followed by delegation of authority to exercise discretion, that is 'when someone is in general charged with making a decision subject to standards set by a particular authority' (Dworkin 1967, p. 32). The empirical focus of this chapter is the state's responsibility for children that are at risk of harm and intrusive measures such as removing children into state care. Child protection is a surprisingly understudied area of the welfare state, given the power that is vested in the decision-makers in a very difficult and highly sensitive area of intrusive state interventions into individuals' private spheres (Burns et al. 2017). A key standard in this area is the Convention on the Rights of the Child (CRC) of 1989. This convention has almost universal, global support, 1 and several countries have made it national law. The CRC gives children strong rights, and a major article is the best interest principle that 'In all actions concerning children, whether undertaken by public or private, social welfare institutions, courts of law, administrative authorities or legislative bodies, the best interests of the child shall be a primary consideration' (Article 3). 2 The ratification of the CRC obligates countries to give the best interest principle primary consideration in decisions concerning individual children, and this entails a clear shift away from the traditional relationship between the family and the state (Gilbert et al. 2011;Skivenes 2002). However, the child's best interest principle as outlined in Article 3 in the CRC is ambiguous, and its application as a guideline for decision-making is not straightforward. It allows huge leeway for a variety of interpretations. Thus, an interesting and important question concerns how governments have applied this principle in their legislation. Do they fill the principle with material content or procedural directions? There are numerous dimensions in relation to a child's best interest that a decisionmaker can (and must) consider, including expert knowledge about nutrition, attachment, education, brain development and the normative and cultural values for a good and meaningful life (cf. Skivenes and Pösö 2017). However, 'what is best for any child or even children in general is often indeterminate and speculative and requires a highly individualized choice between alternatives' (Mnookin and Szwed 1983, p. 8; see also Breen 2002;Elster 1989;Freeman 2005). In addition, these decisions involve complex predictions about the consequences of choices and future outcomes. Altogether, the best interest principle in its current state in the CRC offers little guidance for decision-makers; nonetheless, it is the guiding principle for decisions that have a tremendous impact on the lives of children and adults (Breen 2002;Elster 1989;Freeman 2007;Skivenes 2002).
We set out to examine how states formulate the best interest of the child in their child protection legislation in 14 high-income countries, and in this way, we examine how states interpret and implement the principle, whether they differ and how this may regulate, instruct and guide the discretionary authority delegated to decision-makers. The countries we examine in this chapter are Australia, Austria, Canada, Denmark, England, Estonia, Finland, Germany, Ireland, Norway, Spain, Sweden, Switzerland and the United States of America (USA). 3 In the next section, we present the best interest principle and the concept of discretion, followed by an outline of the various countries' systems on child protection and law. Methods and then findings sections follow. We end with a discussion and a final section with concluding remarks.

The Principle of the Best Interest of the Child and Discretion
An overlooked part of discussions on democratic legitimacy is the role of implementation of policy goals, as political scientist Bo Rothstein (1998) points out in the book Just Institutions Matter. Professional decision-makers in child protection systems are obligated to make decisions according to national legislation and the CRC. Child protection interventions with families based on the best interests of the child and the legitimacy of decisions and interventions are regularly questioned in political and public debates, often in relation to terms such as 'draconian'. Typically, discretion is categorized as weak or strong (Dworkin 1967). Strong discretion concerns decisions that are not 'bound by any standards set by an authority' (Schneider 1992, p. 33). The judges in the European Court of Human Rights can be classified as having strong discretion as they only use the European Convention of Human Rights as their standard for decision-making (Skivenes and Søvig 2016). 4 Weak discretion is when the authority to use judgement is limited by instructions to apply specified aspects or considerations, or to prioritize between different considerations (Dworkin 1967, p. 32;Archard and Skivenes 2009). Dworkin (1967, p. 32) uses an example of a sergeant that has been told to find five experienced people to execute a mission. If 'experienced' is the only criterion, it is clearly a broad instruction, and there is quite a wide array of options for discretion. However, if the sergeant has also been told how experience is defined, for example by criteria such as being over 50 years of age, having five years of practice and self-perceived calmness in stressful situations, the scope for selection is narrower and allows fewer choices. Furthermore, there may be detailed instructions in relation to understanding an order in terms of its material content and collecting information. Thus, in our analysis, a professional decisionmaker will have strong discretion when given the authority to make decisions that are in the best interest of the child, that is Article 3 of the CRC. Professionals have weak discretion when instructions are given concerning the 'best interest principle', and with more instructions, their authority to exercise discretion is weaker. However, if instructions are contradictory, decision-makers must exercise discretion regarding which instruction to follow and how to weigh the instructions against each other, whereby seemingly weak discretion is stronger.
In summary, depending on the type of discretion that legislators set for professional decision-makers in child protection agencies and courts, there will be a continuum of strong to weak discretion for lawful interpretations of the best interest of the child. Before we move on to the broader background of the study, one important point to underscore is that this chapter focuses on legislation as a regulative mechanism, and we are well aware that governments have a range of steering mechanisms and incentives to guide and rein in decision-makers' exercise of discretion, such as guidelines, directives, organizational forms, auditing agencies and choice of profession to implement policy.
The criticism of child protection systems may be because of the strong discretion it gives professional decision-makers in determining the best interests of a child that needs protection. In May 2013, the Committee on the Rights of the Child published a comment on the best interest article (number 3) intended to accommodate the lack of common understanding of the principle (GC No. 14 2013). Part V, Implementation: assessing and determining the child's best interests, outlines seven elements that should be considered when a decision about the child's best interests is to be made (pp. 7ff.): (a) the child's views; (b) the child's identity; (c) preservation of the family environment and maintaining relations; (d) care, protection, and safety of the child; (e) situation of vulnerability; (f ) the child's right to health, and (g) the child's right to education. Each of these elements is laid out in detail in the comment and should provide clear instructions for professionals across countries. However, because this principle covers all areas of a child's life, the comment also underscores that the principle remains ambiguous, and that these seven elements are not relevant to every case, and different elements can be used in different ways in different cases. The content of each element will necessarily vary from child to child and from case to case, depending on the type of decision and the concrete circumstances, as will the importance of each element in the overall assessment. (p. 9) Our focus is on child protection care order decisions, and crudely speaking, the determination of the principle should consist of two parts . One part is based on the scientific knowledge of a child's development and needs. The empirical evidence and documentation provided by the research community establish the ground rules and the important arguments and considerations of valid determinations of what is best for the child. The other part involves normative ethical-cultural considerations about a good life and a good childhood. There is variation across cultures, religions and states, and between individuals and groups about meaningful and good ways of life (Rawls 1971;Shapiro 1999;Skivenes 2002). There are many competing and legitimate ways of bringing up children and as such defining what is good or best for them. Thus, there is not one 'best interest value' that can be expected to be valid and accepted as right for all children. For the former expertbased dimension, there is an argument based on the strength of the evidence and the validity of knowledge, and to some degree, there will be consensus on what has been established as solid knowledge and what is less solid. For the latter value-based dimension, by definition, when there is disagreement and plurality, ethical discussions and interpretations must consider what might be good for a particular individual, family or community. Thus, decisions about the best interests of a child cannot be based solely on expert evaluations, but also on values and norms that hold meaning for human beings. In this landscape, we analyse the instructions made by legislators in 14 countries as expressed in the formation of the best interest principle in child protection.

Child Protection Systems, Welfare States and Jurisdictions
The discretion and the standards that governments delegate to decisionmakers are influenced by and dependent on factors such as type of child protection system, welfare state model, legal system and political order, and power in society. We do not have solid knowledge about the relationship between various system features and the child protection area. The countries differ in how they regulate and organize the institutional settings for child protection and decision-making processes (Burns et al. 2017;Gilbert et al. 2011;Hetherington et al. 1997). The prevailing typology of child protection systems distinguishes between risk-oriented and family service-oriented systems (Gilbert 1997;Gilbert et al. 2011). Risk-oriented systems are based on a high threshold for intervention in the private sphere and intervention when serious risk of harm to a child exists. These systems are focused on mitigating serious risks to children's health and safety (Gilbert et al. 2011). Family service-oriented systems have a low threshold for intervention and provide services for families based on a therapeutic view of rehabilitation, in which the state makes it possible for people to revise and improve their lifestyles and behaviour. The major differences between these two systems are their underlying ideologies, degree of solidarity and the ways in which they manage children at risk. This also includes variations between systems in how they regulate and set standards for children's upbringing and where they set the border between private and public responsibility for children in need of assistance. Our expectation is that risk-oriented systems will have a fact-based understanding of the principle, and with clear instructions on how to handle serious risk, whereas the family service-oriented systems will both include fact-based as well as diffuse values-based-needs instructions. Table 4.1 presents an overview of key characteristics of the 14 countries. 5 Our expectation is that countries scoring high on the child rights index and the child well-being index will also have engaged have included what is recommended by the CRC committee, GC 14.

Data and Methods
The data material for the study consist of the formulations of the best interest principle in child protection legislation in 14 high-income countries. These countries were strategically selected on the basis of prior research on child protection systems and a previous comparative analysis, as we had gained knowledge about these countries' child protection systems and established a network of national experts on child protection in each country. The countries in the study are Australia, Austria, Canada, England would often be characterized as a system that is in between the risk-oriented and family-service oriented systems

The Child's Best Interest Principle across Child Protection…
Denmark, England, Estonia, Finland, Germany, Ireland, Norway, Spain, Sweden, Switzerland and the USA. The various pieces of national legislation were collected by systematically examining the countries' chapters from previous book projects (Burns et al. 2017;Skivenes et al. 2015;Gilbert et al. 2011), reading the national CRC reports and doing specific literature searches. This material was collected in 2015-2016 with support and guidance from national experts. 6 To ensure that we had found the right child protection legislation used for care order removals, we asked child protection experts in each of the countries to inspect and confirm our selected law.
Of the 14 countries under study, five (Australia, Austria, Canada, Switzerland and the USA) have a two-layered structure with overarching legislation for the whole country (such as federal law) and then state legislation. In these countries, we first established an understanding of the relationship between the law on different levels, and then selected one state/canton/province/territory to represent the child protection legislation of the whole country in our analysis. We are clearly aware that this could not be regarded as representative of the whole country. 7 The states we selected in these countries were the state/canton/province/territory that our expert(s) were from in the Australian Capital Territory (Australia), the province of Ontario (Canada), the canton of Basel-Stadt (Switzerland) and the state of Massachusetts (USA). This way of selecting laws provided us with manageable material, and we had the advantage of our expert's knowledge of that specific child protection legislation. For simplicity, we use the term 'country/national legislation' even when we only have legislation from a state/canton/province/territory. The data material for the analysis consisted of the sections of the laws that outline the best interest of the child. Typically, the legislation would have sections along the following lines: 'in deciding what is in the best interest of a child or a young person, a decision-maker must …' (Australia: Children and Young People Act 2008). The length of the texts from each country varied, from a few lines in Switzerland to over a half page of law from Australia. Table 4.2 provides an overview of the title of the child protection law in the 14 countries included in the study, as well as the specific section that outlines the best interest of the child. 8 All texts we analysed were written in or translated into English. Five countries have English as the original language (Australia, Canada, England, Ireland and the USA) and five have an official English translation of the law (Denmark, Estonia, Finland, Germany and Norway). For Austria, Spain, Sweden and Switzerland we used a translation bureau to translate the laws into English. Translation is a critical process, as the English concepts and system terminology may not capture or be compatible with those of other countries' ways of organizing child protection systems. For the non-English-speaking countries, we checked the translations with the experts to make sure that the meaning was not lost in the translation process. The translation of the legal texts and the potential loss of meaning from or mismatch with the original language of the law is a validity problem for most cross-country studies. 9 To analyse the laws, both researchers first read the texts and identified themes, and then discussed the themes and made a coding scheme of mutually exclusive codes (cf. Table 4.3). The analysis was driven by two questions: First, what is the material content of the best interest principle, as formulated by the legislators? Second, what type of instructions and guidance does the legislation provide? The texts were coded manually by the researchers in several rounds, and then the reliability of the coding was tested by two project associates. In the few cases of discrepancies, the authors discussed the deviations and agreed upon a mutual interpretation. Includes factors about protection of the child against harm or risk of harm, and/or considerations of prior experiences of harm or potential future risk situations Permanency Includes text related to the importance of permanency or stability of the emotional and/or physical living conditions and upbringing of the child Weight and procedures Includes text that states how the best interests of the child should be weighed against other principles or rights, and/or how material factors should be ranked and/or whether a time frame is mentioned

Limitations
We have only looked at the formulation of the principle in the selective legislation for child protection; we have not examined other legislation or other mechanisms that may instruct professionals in the front-line services and the courts, such as political-administrative directives or case laws. We do not consider how the principle is actually applied and reasoned in courts or child protection agencies. The focus is solely on the legislative formulation of the best interest principle in the selected laws in these countries. The limitation with this approach is that owing to various ways of constructing legislation and their legal order, we cannot exclude the possibility that some countries include the general principles of their child protection systems within the best interest principle, whereas others have laid out principles and important considerations elsewhere in legislation and/or their administrative/legal systems.

Findings
The analysis of the legislation showed that the best interest principle is worded in various ways. Nine of the 14 countries used the term 'best interest' in the text, whereas England and Ireland used the wording 'the welfare of the child'. Similarly, Austria and Switzerland refer to 'the child's well-being' in a literal sense, and Finland uses the term 'the interest of the child'. Each country has chosen different strategies for approaching the principle, with eight of the 14 countries providing rather broad descriptions of the material content on how the principle is to be understood. Overall, eight main material considerations are represented in the legislation analysed, as well as one category for the weight of the principle and procedural aspects. Table 4.4 presents an overview of the findings and shows how some themes are present in almost all legislation, whereas others are only of concern for a few countries.

Child's Participation
Twelve of the 14 countries (the exceptions are Germany and Switzerland) require the child's view, opinion, wishes, feelings or meanings to be considered (cf.

The Child's Needs
Ten of the 14 countries include some wording regarding the needs of the child, whereas the legislation in Estonia, Ireland, Norway and Switzerland does not explicitly mention needs. The child's need is a contested concept. Although most of the considerations around best interest are directly or indirectly about the needs and care of the child in the wider sense, here we focus on direct mention of particular needs in relation to the child, such as physical, emotional, intellectual and educational needs. Thus, we do not include statements that focus on aspects such as the need for parental care or protection from harm, or any other consideration that is included in the other seven categories that we identified in the material. Although the legislation from Australia, Austria, Canada, Denmark, England, Finland, Spain and the USA varies in its characterizations of needs, most of the countries cite both emotional and basic physical needs. Finland serves as an example of the latter, with an extensive description of needs on both general and specific levels: (1) Child welfare must promote the favourable development and well-being of the child. … 1) balanced development and well-being, and close and continuing human relationships; 2) the opportunity to be given understanding and affection, as well as supervision and care that accord with the child's age and level of development … (#1, 2.1, 2.2) The USA and Denmark emphasize both care and affection as well as the impact of upbringing on the child's adult life: 'The health and safety of the child shall be of paramount concern, and shall include the long-term wellbeing of the child' (USA #1, para. 3).

Permanency
Nine of the countries (Australia, Austria, Canada, Denmark England, Germany, Norway, Spain and the USA) cite factors related to the importance of permanency and stability of emotional and/or physical living conditions and the upbringing of the child. Permanence is essential for structure, strength and consistency to support children's development (Skivenes and Thoburn 2017), and in the US legislation it is formulated as follows: 'The department's considerations of appropriate services and placement decisions shall be made in a timely manner in order to facilitate permanency planning for the child' (#1, para. 5).
Several of the countries have a focus on stability in relation to the birth family or a change to an established living arrangement. Australia, Denmark, Spain and the USA are all in this category: '(d) the likely effect on the child or young person of changes to the child's or young person's circumstances, including separation from a parent or anyone else with whom the child has been living' (Australia #1.d). Canada, England and Norway have chosen a neutral formulation of stability, here illustrated by Norway: 'When applying the provisions of this chapter, decisive importance shall be attached to finding measures which are in the child's best interests. This includes attaching importance to giving the child stable and good contact with adults and continuity in the care provided' (4-1, para. 1).

Protection
Nine of the 14 countries include consideration of the child's safety or risk factors, and the five countries that do not explicitly mention this are Denmark, Estonia, Ireland, Norway and Switzerland. The focus is on two types of risk: (a) the potential harm unnecessary removal or intervention may have on the child, and/or (b) the risks of abuse, neglect or harm to the child if he or she remains in a potentially dangerous situation. The Austrian, Canadian and Swedish legislation includes both dimensions: 'In the assessment of what is best for the child, particular focus must be placed on-the risk of the child or other family member being subjected to abuse or the child being illegally removed or retained or otherwise treated badly' (Sweden #2.a). The legislation can reference protection broadly or be more detailed (but yet far-reaching), like the English legislation, which includes both past and present risks, in that '(e) any harm which he (the child) has suffered or is at risk of suffering' (#3.e). The legislation from Australia, Austria, Spain and the USA lists both physical and psychological (or emotional or spiritual) harm or abuse. Harm or violence against another family member is mentioned in the legislation from Australia, Austria, Canada and Sweden. The Austrian legislation highlights the negative impact a decision made against the child's wishes may have: '6. the prevention of an adverse effect on the child due to the taking of action against his or her will' (#6).

The Child's Relationships
The legislation in eight of the 14 countries includes various forms of consideration of the child's relationship with a caregiver or other significant others. The legislation in England, Estonia, Finland, Ireland, Switzerland and the USA does not refer to this consideration. This code does not include text that relates to the parent's or caregiver's care of the 'child's needs' intended to conserve 'permanency/stability' for the child. Amongst the eight countries that mention the child's relationships, most emphasize both the relationship to the caregivers or parents and the child's relationship with other family members. This is illustrated by the Canadian legislation, which expresses that the decision-maker shall take into consideration '6. The child's relationships and emotional ties to a parent, sibling, relative, other member of the child's extended family or member of the child's community' (#3.6). The legislation also emphasizes the child's place in the family and the importance of a positive relationship with a parent for development. Only the Swedish legislation in this sample defines the relationship with the natural parents: 'the child's need for close and good contact with both parents' (#2.a).

The Child's Identity
Considerations that focus on the child's individuality in terms of cultural inheritance or other aspects important to the child's identity are mentioned in six of the 14 countries: Australia, Austria, Canada, England, Finland and Spain. The legislation that requires that the child's identity, individuality or culture be considered varies in the comprehensiveness and details of the wording/text. The Austrian, Canadian, English and Finnish legislation are brief, for example: '7) the need to take account of the child's linguistic, cultural and religious background' (Finland #2.7). By contrast, the Australian and Spanish legislation elaborates: 'd) Preservation of the identity, culture, religion, convictions, sexual orientation and identity of the minor, as well as non-discrimination against same for these reasons or any other conditions, including disability, guaranteeing the harmonious development of their personality' (Spain #2.d).
The Australian legislation has a specific focus on Aboriginal or Torres Strait Islander children or young people, stating that it is a high priority to protect and promote the child's cultural and spiritual identity. ' [M] aintaining and building the child's or the young person's connections to the family, community and culture' (#1.g) are also emphasized.

Parents' Perspective
Four countries mention the parents' or caregivers' capacity to care for the child, or their opinion about the child-that is, the legislation from Australia, Austria, England and Ireland. English law illustrates this, referring to '(f ) how capable each of his parents, and any other person in relation to whom the court considers the question to be relevant, is of meeting his needs' (#1.f ). The focus is on parents' abilities to meet the child's needs (Australia and England), their acceptance of the child (Austria) and the 'rights and duties of parents' (Ireland).

Future
Three countries-Denmark, Finland and Spain-include considerations of the child's future life or adulthood, as the legislation from Spain illustrates: 'e) Preparation for transition to adulthood and independence, in accordance with their personal capacities and circumstances' (#3.e).

Weight and Procedures
Although the interpretation of the best interest principle varies considerably, all 14 countries specify how factors should be ranked or the best interest principle weighted against other principles or mention a timeline. The role of the principle in relation to other principles and rights is foregrounded by eight countries-England, Estonia, Finland, Ireland, Norway, Sweden and Switzerland-and they include terms such as the best interest being paramount, a primary consideration or a priority. For example, 'the child's welfare shall be the court's paramount consideration' (England #1.b). Although many countries include specific material factors to be considered, as we have shown above, seven countries-Australia, Canada, Denmark, Estonia, Germany, Spain and the USA-have an unspecified caveat that 'any other fact or circumstance' (Australia #2) considered relevant should be included in the decision-making process. Four countries-Australia, Canada, England and Finland-mention the importance of making the decision without delay, for example, in the English act 'any delay in determining the question is likely to prejudice the welfare of the child' (#2).

Summary Findings
In sum, the findings show that there are different understandings between the 14 countries of the principle of the child's best interest, as well as differences in delegation by the government to the professional decisionmaker of authority to exercise discretion. All countries have some reference to weighting, timelines and/or procedural requirements.
Regarding the material content of the principle, we identify clear distinctions between the legislative interpretations of the child's best interest principle. It is clear that depending on the specific country, these considerations represent zero to seven material themes that professionals are instructed to consider. Ten of the 14 countries have four or more material themes reflected in the child's best interest principle, whereas the other four have zero to three material themes to consider. The principle's foundation as a right for children is evidently a right that is open to interpretation and degrees of implementation by the ratifying states.

Discussion
We learn that from an empirical view, there are eight factors that are important when a child's best interest is determined in child protection situations in these 14 states. This provides us with information about what is perceived to be important for children's upbringing and gives material content to the best interest standard. Some factors are more important than others: Child's participation is regarded as decisive by almost all of the countries (12 of the 14), which indicates that children have a prominent position in child protection legislation and that legislation accords with the fundamental principle of children's participation in the CRC (Art. 12; cf. GC 14 2013). There is also a large amount of the national legislation (in 10 of the 14 countries) that raises the issue of the child's needs as an important consideration. This is a basic premise for a healthy upbringing, and in child protection cases this is likely to have been a neglected factor in the lives of the concerned children. Permanency is also high on the list, and denotes the importance of continuity and the sense of belonging for a child. Protection of children from harm and neglect and providing them with a safe living environment are included by many of the countries (nine of the 14). Preservation of a child's relationship, be it to carers, parents or others in a wider sense, is included by eight of the 14 countries. These five factors centre around the child and his or her viewpoints, relationships, permanency, needs and risks. The remaining three considerations are included by fewer than half of the countries and can be regarded as factors that do not have the same relative importance: child's identity (six of 14); natural parents' perspective (four of 14) and the future of the child (three of 14). Compared with GC 14, which refers to seven elements that should be considered, we find a strong overlap of four elements: (a) the child's views; (b) preservation of the family environment and maintaining relations; (c) care, protection and safety of the child, and (d) situation of vulnerability. However, it is interesting to note what is not overlapping. Only a few countries mention: (e) the child's identity; (f ) the child's right to health and (g) the child's right to education. Most of the strong multicultural countries (in terms, e.g., of population constellation, indigenous people), such as Australia, Canada, England and Spain, have included the child's identity as a factor. However, neither Germany nor the USA mentions this factor, though both are multicultural countries that have received a large influx of migrants. The lack of focus in the legislation on education and health is surprising, as it is well established that these areas are extremely important for children's future adulthood, and to gain employment and independence from the state. Related to this is the curious lack of attention to the future well-being of the child, when the best interest of a child should be considered. Only three of the 14 countries in our sample mention this. The CRC committee does not include this element at all. We would expect the future of the child to be an important consideration for professional decision-makers with regard to measures such as a care order application. Furthermore, there is solid research from Adverse Childhood Experiences 10 studies on the negative consequences of neglect and abuse, and it makes sense for professional decision-makers to ensure that a child in need of protection can achieve a good adult life, as the Danish do.
We cannot identify a pattern between type of child protection system and understanding of the principle, nor between countries' ranking on indexes and the understanding of the principle nor in the form of an overlapping with the CRC committee's recommendations. This may be due to the countries involved, or it may be that the child protection 'typology' and index ranking are too broad and lack sufficient nuances.
All countries mention weighting of the material factors, time limits or relationships between rights or principles. An interesting point here is how the legislation draws upon a broader set of considerations if deemed necessary, just as the CRC committee recommends (cf. GC No. 14). The question of how much weight the decision-maker should attach to the child's best interest is also a way of regulating discretion. CRC Article 3 states that the best interest of the child shall be a primary consideration. A primary consideration does not have an 'absolute priority' over other considerations. 'Primary' means 'first', in that it should be a 'first consideration', but it does not necessarily determine the course of action (Freeman 2007).

Strong and Weak Discretion
Through legislation, governments set standards for decision-makers about interpreting and applying the best interest principle. A distinction between weak and strong discretion can be drawn, and in the material of the 14 countries, there are six where legislation provides decision-makers with strong discretion, including Estonia, Ireland, Norway and Switzerland. Strong discretion is evident, as in these countries. The legislators have only set a few requirements on the considerations that decision-makers must take into account when making the child's best interest decisions in child protection. At the opposite extreme is national legislation that provides decision-makers with weak discretion; the remaining eight countries are in this category. Here, a range of considerations are provided that the decision-makers should consider. For example, the Australian, English and Finnish legislation list seven considerations. The category of strong discretion includes legislation or guidelines with few instructions concerning what professional decision-makers should consider (cf. Table 4.5 below). Of the four countries that we categorize as having strong discretion-Norway, Ireland, Estonia and Switzerland-the Swiss legislation is by far the most general, and professional decision-makers are vested with strong discretion: 'With all state action that affects children and young people, their welfare is to be given priority.' Evident in this analysis is the complete lack of instructions to decision-makers on interpreting the best interest principle. This makes the decision-making situations vulnerable to contingency and personal perceptions and indicates a lack of standards to hold decision-makers accountable. We could have seen a procedural approach, evident in the Spanish and Estonian legislation, in which the decision-makers must collect all relevant information and hear all involved parties to ensure reasoned deliberation: '3) assessing all the relevant circumstances in aggregate, to form a reasoned opinion concerning the best interests of the child with regard to the planned decision' (Estonia #3). The Estonian legislation also displays a strong child-centrism by requiring that the situation and the person of the child are considered, and any decision that is contrary to the child's wishes must be explicitly justified. This child-centrism in the Estonian legislation may be explained by its recent enactment (2016), which is part of a trend of increasing child-centrism throughout Western countries (Gilbert et al. 2011;Skivenes and Søvig 2016).
The category of weak discretion is included in legislation in 10 of the 14 countries, and this category is characterized by substantial instructions for the decision-makers in the form of four or more instructions (cf .  Table 4.6). The Australian legislation illustrates how the delegation of weak discretion takes form in practice: 'What is in the best interest of the child or young person? (1) For the care and protection chapters, in deciding what is in the best interests of a child or young person, a decision-maker must consider each of the following matters that are relevant to the child or young person:…' (#349, 1). This instruction is followed by a list of 11 separate considerations for a decision-maker. Australia, Austria and Spain authorize the weakest discretion, with Germany, Sweden and the USA at the other end of the weak discretion continuum. However, even for the countries that provide the weakest discretion, there are openings for considering additional elements that the situation and the decision-maker see fit, or the possibility to deviate from the considerations. Surely, it is sensible to have flexibility in these complex and sensitive cases and to let decision-makers use their professional competency to assess the situation and the specific child involved in the decision-making. However, as evident in the procedural approach in the Estonian legislation, it makes sense to recommend that whenever a consideration is deemed unnecessary, this should be explicitly explained.

Concluding Remarks
Although in this analysis, we have made bold statements about how 'countries' think about the child's best interest principle, we are also aware that child protection systems in most high-income countries are based on the same basic principles, including that the family/parents have the primary responsibility for their children, and the least intrusive principle, that removals should be temporary and that the welfare/best interest of the child should be considered (Burns et al. 2017). Nevertheless, the organization of systems and the removal proceedings differ between jurisdictions (Burns et al. 2017). Our analysis of national child protection legislation of the child's best interest principle shows distinct differences between countries in the interpretation of the principle of best interest. This indicates that what is deemed important for children across child protection systems differs, possibly because of cultural views. What is clearly evident is that the recommendations of the CRC committee are only partially included in child protection legislation that we have analysed; this may be attributable to the committee report only being published in 2013. Nevertheless, it is interesting that the overlap between the committee's recommendations and the countries' legislation is not more extensive as one should expect that standards for children's needs across countries would have included some of the same topics. We should ask whether the indeterminacy and the ambiguity of the best interest principle demands that each individual be considered on his or her own terms, and that there should not be any material content guiding the decision because it can only be made in regard to the unique child it concerns. Following this line of reasoning, the very simple principle formulated in the Swiss child protection legislation may be consistent with this norm. However, this raises the fundamental problem with the exercise of discretion to which decision-makers are not bound by any standards, as Dworkin (1967) already had discussed. We believe that the emphasis on indeterminacy has led to an exaggerated belief in the differences between individuals at the expense of their commonalities. The established knowledge about common needs for people to lead a good life is valid for all children. Shapiro (1999) labels this basic need 'factbased' and suggests that these should be distinguished from best interests (value-based considerations).
A striking finding is the variations in discretionary authority that are displayed. It is a broad continuum that extends from Switzerland at one end with strong discretion, and Australia at the other with weak discretion. An example of the force of clear instructions for professionals is shown in a study of child protection workers in England, California (USA) and Norway on the use of adoption as a child protection measure (Skivenes and Tefre 2012). Only Norway has given professionals strong discretionary authority over this matter. Presented with the same case, close to all staff in England and California (USA) recommended adoption for the child, whereas six out of ten of the Norwegians did the same. The empirical studies that should be pursued are whether and how the various prerogatives of professional decision-makers are exercised. We anticipate that in countries with weak discretion, there will be more similarities between decision-makers on important assessments such as care orders.
The implementation of children's rights across states still have a long way to go, and we are only witnessing an emerging child-centrism placing children on an equal footing to adults in modern societies. Providing instruction on the fact-based elements of the principle is in our opinion an important way to enhance children's position across societies, independent of ethical-cultural values and norms.

Notes
1. All states in the world, except the USA, have ratified the CRC as of 16 May 2017. 2. Cf. the recommendation by the CRC committee on the interpretation of the material content of the best interest principle (GC No. 14, 2013). 3. Although the USA has not ratified the CRC, we include it as several states use the principles from the CRC and the best interest principle in child protection, Gateway, C. W. I. 2016. State statutes: Determining the best interests of the child [online]. Available at: https://www.childwelfare. gov/topics/systemwide/laws-policies/statutes/best-interest/ (accessed 3 June 2016). 4. The Court is not without regulations, as only legal scholars can be judges, and decisions are made in accordance with legal methods and follow legal precedents. 5. For practical reasons, we use the country name in the text, even though for several countries, we only analyse the legislation of one state/canton/ province/territory. 6. We are very grateful for the expert knowledge, guidance and help to include, for example, child protection legislation for the entire USA. 8. An online appendix of all the laws included in the analysis is available on the project website: http://www.uib.no/sites/w3.uib.no/files/attachments/appendix_to_the_best_interests_of_the_child_in_child_protection.pdf. For the sake of simplicity, we reference only the country and the paragraph in the chapter.
9. Even in official translations, the wording of the principle may be different from that in the native tongue. For example, the expert from Estonia noted that the official English translation used the term 'best interest of the child', but the Estonian version uses the phrase 'child's interest'. The expert from Switzerland pointed out that the German word 'Kindeswohl' entails more than the English translation of 'well-being of the child', although 'Kindeswohl' might be understood as equivalent to the best interest of the child. 10. https://en.wikipedia.org/wiki/Adverse_Childhood_Experiences_Study.