The ‘turn to parenting’: paradigm shift or work in progress?

In the field of child and youth support, preventive and parent-oriented support programs are currently emphasized. This is understood as a new paradigm of parenting support, which is sometimes referred to as turn to parenting.” Using Campbell’s (Annual Review of Sociology, 28:21–38, Yee’s (Organization 50(1):69–108, and Hall’s (Comparative 25(3):275–296, 1993) theories concerning the processes of policy making, this study investigates whether such a turn has indeed taken place. We first trace back the concepts of parenting and parenting support in policy documents of the Netherlands. Hereafter, we study the implications for practice by quantitatively analysing developments in the implementation and use of child and youth support interventions from 1990 onwards using a Dutch database of effective youth care interventions. The results suggest a trend towards preventive and targeted preventive interventions. Contrasting the turn to parenting, we find a statistically significant rise in early education programs. The findings suggest that ideas around the importance of parenting and parenting support have been ongoing, and that policy and practice have mutually influenced each other during the time-frame of this research project (1990–2010). International ideas and national actors have changed the policy discourse, but the impact on practice is ambiguous.

In contrast to earlier forms of parenting support-aimed mostly at parents and families experiencing problems in the upbringing of their children-literature suggests a general shift within this field towards preventive and early interventions (Coussée et al. 2011;Lee et al. 2010). In such preventive approaches, interventions tend to target a younger age group of children and their parents; help is offered before serious problems occur and family life is considered to be one of the main sources of influence in the development of children, especially in the early years. The approach is embraced by the Dutch government (e.g., Ministry of Youth and Family [Ministerie van Jeugd en Gezin] 2007. Parents are seen as the linchpin through which children's development can be optimized. Therefore, policy measures and programs should enable and assist parents to organize family life in the most optimal sense (i.e., Council for Social Development [Raad voor Maatschappelijke Ontwikkeling, (RMO)] 2012; Prinsen et al. 2012).
The assumed "turn to parenting" is often explained by two lines of argumentation: (1) Parenting has become more difficult than it used to be, due to societal processes such as fragmented families and women's employment. Parents therefore need support in raising their children (see, e.g., Gillies 2012); (2) Due to external influences (new social media, more complex and competitive societies and migration), parents experience more trouble in finding "the right way" to raise their children, and are therefore actively searching for guidance and support (see, e.g., Asscher et al. 2008;Ivan et al. 2014). However, it is unclear whether such notions and argumentations had their impact on policy measures or whether policy measures have influenced the social perceptions of children and family life (Yee 1996). Also, policy making does not develop in isolation (Wincott 2010;Béland 2009). For instance, the ratification of the United Nations Convention of the Rights of the Child (UNCRC) in 1995 has affected child and family policy throughout the world (Such and Walker 2005). Even in the United States-which has not ratified the UNCRC-influences from the UNCRC can be found in family policy (Levesque 1996). Likewise, social investment approaches-that focus on social investments like lifelong learning and early childhood services as opposed to social protection measures (Jenson and Saint-Martin 2006)-have affected policy measures worldwide thereby drawing the attention towards human capital and skills, towards "an active society" and a concurrent emphasis on the healthy development of children (Geinger et al. 2013;Jenson and Saint-Martin 2006).
Central to this article 1 is the notion that practice is not a one-to-one translation of policy (Jenson and Saint-Martin 2006) but rather that ideas, in this case about the central role of parents in the upbringing of children, may already exist in society. Such ideas influence the process of policy making with many actors influencing and restricting each other in bringing ideas to the fore (Wincott 2010;Béland 2009;Rigby et al. 2007;Campbell 2002;Yee 1996). We will explore this issue by tracing back concepts of parenting and parenting support in Dutch policy reports and combining this with a quantitative analysis of the interventions offered in the Dutch child and youth support system.

Ideas and policy making
Explanations for continuity and change in the policy process can be found in different forms of institutionalism (i.e., historical, rational choice or sociological institutionalism). This article perceives the policy process as a dynamic interactive process, in which the role of ideas and discourse should be put central, and which suits discursive institutionalism theories (Schmidt 2010). According to Campbell for instance, ideas can be differentiated into "cognitive paradigms" or "world views" (taken-for-granted descriptions and theoretical analyses) and "normative frameworks" (values, attitudes, identities) that can explain (cross-national) differences in the policy-making process. In contrast, "world cultures" are ideas which explain cross-national similarities. Next to this, ideas can also reflect the so-called "frames"-the formulation policy makers use in order to sell their ideas-and "programmatic ideas, " reflecting actual political programs (Campbell 2002). For instance, cognitive paradigms represent theoretical ideas about the role of parents ('cause') and the healthy development of children ('effect'). Normative frameworks express values such as the central role of parents or the positioning of children within society. World cultures are found in the (worldwide) emphasis on the rights and well-being of the child (e.g., the UNCRC), or the importance of children for the future of society (e.g., social investment approaches). An emphasis on parenting support may also reflect programmatic ideas of a specific political party, or it might be framed in such a way for other purposes but to suit already existing normative frameworks or cognitive paradigms in society.
Several kinds of actors play a guiding role in transferring ideas between institutions and in putting ideas into practice, for example through policy measures (Campbell 2002). First, experts and intellectuals play a key role as their claims of knowledge and expertise are highly likely to be heard. Second, national and international epistemic communities of professionals and experts may play an important role in mobilizing and framing information. Third, existing institutions affect the influence ideas may have by filtering what information can and cannot enter the political arena. Once ideas are embedded in laws and policy measures, and thereby given form in everyday life, their influence is stabilized and thereby prolonged (Campbell 2002;Yee 1996). Change is most likely when important officials are able to translate changes in such way that it suits with a culture's core values. This holds for national ideas as well as for international ideas (Béland 2009). Choices that are made in establishing specific policy measures and instruments may also serve to emphasize a specific discourse, thereby playing into the dominance of a specific policy paradigm (Rigby et al. 2007). Different actors can encourage or restrict the policymaking process and its translation into practice (Yee 1996). For instance, policy makers can legislate certain ideas, but professionals and professional institutions can object to actually implementing these ideas or find ways to restrict beforehand the spectrum of legislation (Yee 1996).
Campbell's and Yee's theories fit Hall's theory (1993) on paradigm shifts. According to Hall, change can come about via three important variables: (1) a change in settings (first-order change); (2) a change in instruments and settings (second-order change); and (3) a change in goals, instruments and settings (third-order change). Hall considers only a third-order setting as a true paradigm shift, as this reflects an actual change in terms and "taken for granted"-assumptions (normative or cognitive frameworks). Other changes are "normal policymaking" (Hall 1993, p. 279). The suggested turn to parenting implies a third-order change in which the goals of child and family policy change, with a shift towards preventive support of parents (as opposed to curative support), and in which newly developed instruments (parenting support programs) play a key role (see also Daly 2010). A paradigm shift as described by Hall (1993), however, implies a rather sudden change, whereas approaches such as Campbell's (2002) and Yee's (1996) indicate a more gradual continuous process, reflecting "work in progress".
Evidently, the process of policy making and policy implementation is an entangled process in which a multitude of actors and factors play their part (Wincott 2010;Rigby et al. 2007). In what follows, we will endeavour to disentangle this process in relation to Dutch child and family support, in an effort to gain insight into underlying mechanisms. We will do so by first describing international developments related to this field. We will then turn to the national context of the Netherlands. Finally, we will quantitatively analyse the practice of child and youth support in the Netherlands. Findings will be taken together in the concluding section of this article.

International context: Children's Rights and the Social Investment Perspective
The UNCRC was adopted in 1989 and subsequently ratified by the Netherlands in 1995 (http://www.kinderrechten.nl). By ratifying, the Netherlands (like other countries accepting the Convention) obligated itself to adapt and develop laws to ensure that the rights of the child are safeguarded (http://www.rijksoverheid.nl). The UNCRC thereby had a direct effect on the development and implementation of child and youth policies. Concurrently, investments in human capital for instance lifelong learning and early childhood services have become focal points of policy measures in welfare states, to address new social risks such as unemployment and single parenthood (Jenson and Saint-Martin 2006). Family policies in this perspective are geared towards helping children to get the best start in life, and towards enabling children to find their place in a knowledge-based society (Jenson and Saint-Martin 2006;Kutscher 2012;Lewis 2006). In line with social investment ideas, child support services are developed for a variety of aims, for instance in overcoming (child) poverty, enhancing social cohesion or encouraging the skills and capacities of children to become flexible citizens. From this viewpoint, the family is seen as a mechanism for tackling social issues, and policy measures taken from this perspective underscore support and guidance for families (Gillies 2012).
Both the children's rights approach and the social investment approach consider the child as human capital, for which specific input and care is required to retrieve specific output. Some even say child rearing in this perspective can be seen as "working on a product" (Geinger et al. 2013;Hermanns 1992). In this perspective, the role of parents is to raise their children properly both on behalf of the children and on behalf of society. This job should be done adequately, and it is considered to be an individual failure if parents are unable to live up to the requirements (Gillies 2012;Kutscher 2012;Lister 2006). Such ideas are also reflected in parental worries about children's future in a more complex and competitive world. By consequence, child centeredness is growing in European countries, including the Netherlands. Interestingly, it is mainly middle class and nonreligious parents-mostly female-who increasingly share the attitude that children's interests prevail over their own (Ivan et al. 2014).
These policy developments also bear with them some other (unforeseen) consequences: First, emphasizing the rights of the child often indicates an emphasis on the duty of parents to ensure the health, well-being and development of their children, and to ensure that children's rights are being served (Bennet 1996). When seen in this light, parenting support interventions not only serve the well-being of parents and children, but are also a mechanism for guiding and controlling parents (Reynaert et al. 2009).
Second, a strict emphasis on the rights of the child and parental duties often coincides with a neglect of the role of the context and life circumstances: A children's rights perspective considers all children to be equal. Yet, the circumstances in which these rights have to be realized can differ economically, socially and historically (Bennet 1996;Reynaert et al. 2009). Parenting support interventions encourage a familialistic approach, as it centres on the family as the main source for child well-being and less on conditions such as poverty, lack of social support networks and other matters related to the social context of the child (Geinger et al. 2013;Gillies 2011;Ramaekers and Vandezande 2012;RMO 2012).

National policy developments: parenting support services in the Netherlands
In the Netherlands, the governmental care for children and parents can be divided into two specific professional fields: On the one hand, via a system of social security and tax reductions national social policy regulates child allowances and child care arrangements that-directly or indirectly-influence family life and parents' possibilities. On the other hand, a professional system of psychologists, pedagogues and social workers offer specific interventions to parents and children in overcoming problems in child rearing or in treating (possible) problematic behaviour of children.
Child and family policy does not exist as a clearly defined policy field in the Netherlands, but is the responsibility of a variety of Ministries: The Ministry of Education, Culture and Science is responsible for early education programmes for children, whereas the Ministry of Social Affairs and Employment bears responsibility for work-care arrangements. The Ministry of Health, Welfare and Sports is responsible for overall youth care and support services. Only in the period 2007-2010, the Netherlands had a specific Ministry of Youth and Family. Headed by the leader of the moderate Christian political party Christian Union, this ministry expressed a specific family-oriented ideology and emphasized the need for preventive care and support for parents.
However, parenting support in itself was not a new idea and was already mentioned in earlier governmental reports as a solution to societal concerns such as the diminishing child-raising experience of parents, increased labour participation of women and increased awareness of child abuse. Next to this, new scientific insights of that time also encouraged a focus on parenting support (Wilbrink-Griffioen 1992). In 1986, the Council of Youth Policy [Raad voor Jeugdbeleid] wrote a report "Supporting Child Rearing" [Opvoeding Ondersteund] on this subject, followed by two reports of the Ministry of Health, Welfare andSports in 1991 (Wilbrink-Griffioen 1992). From 1995 up to 2003, governmental reports addressed the role of youth specifically. In 1995 the Dutch government organized a working group to give form to the importance of preventive youth policy [Regie in de Jeugdzorg], thus centring on youth themselves as opposed to parents. Comparably, reports of the national government like Youth in Balance I (1999) and II (2000) [Jeugd in Balans I en II] and municipality-based reports like Project Municipal Youth Policy (1999)(2000)(2001)(2002)(2003) [Lokaal Jeugdbeleid], prevention was mostly concerned with youth and youth-related issues and the focal points were the (political and social) participation of youth, collaboration between youth services and the educational system and safety-related matters (Netherlands Youth Institute 2015b; Staatscourant 1995).
In 2003, two different governmental approaches dominated the political agenda concerning families and youth, which appear to have given a new impulse to the encouragement of parenting support and its concurrent emphasis on preventive forms of care. First, the "as-as-as-policies" [zo-zo-zo-beleid], meaning "as soon as possible, as fast as possible and as close to home as possible" (Matthijs and Vincken 1997). This resulted among others in giving municipalities the responsibility of public child healthcare for children age 0-4 years. It included tasks such as monitoring and general developmental and risk assessments, vaccination, information and counselling, and influencing health risks. Parenting support services were suggested to be included in this range of duties, mostly geared towards the prevention of child abuse (Hermanns 2000). Second, in 2003 the Dutch government established a working group named Operation Young [Operatie Jong] to investigate possible bottlenecks within the Dutch youth care system. Operation Young centred on the role of Dutch national governance in relation to youth and family policy. In 2004 Operation Young published its findings, and suggested (among others) policy measures related to support structures in and around schools, the use of risk assessments and the establishment of the Youth Monitor, that tracks the development of children on matter such as family life, health and education to provide relevant institutions with the necessary information (Central Bureau of Statistics 2014; Van Eijck 2006).
The parenting support theme thus re-occurs in a variety of policy reports in the 1990s and early 2000s (Hermanns 1992), but it did not really dominate the political agenda up until 2007. In this year, a new governmental coalition was formed, consisting of Christian Democrats, Social Democrats and the Christian Party, which established the earliermentioned Ministry of Youth and Family. The Ministry was also a direct result of the findings of Operation Young, and was highly influential in the transformation of the Dutch youth care system. With it came a strong emphasis on both preventive interventions and the role of the family (Ministry of Youth and Family 2007). 2 The Ministry developed municipality-based Centres for Youth and Family as easy-access organizations where parents could turn to for help and advice. The core of these Centres was the public child healthcare service (with an extended age range of −9 months to 18 years), which was already strongly embedded in Dutch society since the beginning of the 20th century (Rigter 1996).
As the above suggests, preventive forms of care and parenting support have been on the Dutch political agenda since 1986, but it took until 2007 to come into effect in the form of a Ministry for Youth and Family and municipality-based Centres for Youth and Family. During this time, new policy measures were also implemented in response to UNCRC requirements. This suggests a paradigm shift in which all previous influences and ideas come together to transform the field (Campbell 2002;Hall 1993). The transition to municipality-based Centres for Youth and Family implies a change in settings, but in line with the indicators of a paradigm shift, such a shift would also imply a change in instruments with a growth in targeted and preventive interventions. Also, goals would change by focusing especially on parents as the lynchpin in child well-being, and the number of parenting support interventions should significantly outgrow the number of other interventions (i.e., early learning or regular child-oriented interventions). In what follows, we will analyse whether a shift in policy ideas has also resulted in a change in practice.

Methods
The database of effective interventions of the Netherlands Youth Institute (http://www. nji.nl) has been analysed to investigate the translation of policy change into a change in practice. This database includes a wide variety of child and/or family support programmes, and was initiated to inform the professional field about effective interventions and to encourage them to use these. To be included in the database, program developers can submit their program, which is then assessed on several aspects, for instance theoretical soundness and effectiveness found in national and/or international evaluation studies. Assessment of programs is done by an independent Committee for the Recognition of Interventions [Erkenningscommissie Interventies]. Programs are re-assessed every 5 years to measure their progress in effectiveness (Netherlands Youth Institute 2015a). It is important to note that this database varies from time to time as new programs enter the database while other programs may be excluded from this database for instance due to negative findings in evaluation studies.
The database accredits programs according to an 'effectiveness ladder' (Van Yperen et al. 2010); programs are defined as 'not acknowledged' , 'in preparation' , or 'acknowledged. ' The first two labels mean that the program has not yet shown sufficiently to be effective. The latter group-which has entered the first step on the effectiveness ladderis subsequently defined as 'theoretically effective' , 'first indications' , 'good indications' , and 'strong indications' . The first of these labels means a program is theoretically sound, the other labels reflect an increase in proven effectiveness by evaluation studies (Van Yperen et al. 2010). Any intervention can be submitted for accreditation as long as the standardized guidelines of the Committee are followed. It is not mandatory to submit a program for accreditation (Zwikker et al. 2009).
The database was downloaded on November 20th, 2012 and sampling and analyses are conducted on the interventions that were included in the database at that time. We focused solely on acknowledged interventions (n = 169), as these programs are preferred to be implemented and used by Dutch child and family support organizations (Ministry of Youth and Family 2007). Interventions may exist that are not included in the Institute's database, but due to the sheer size and fragmentation of the Dutch youth care system-a mixture between formal, regular and commercialized care, with each organization being able to develop its own intervention tailored to its specific local context or problems-it would be impossible to define the full scale of youth care interventions within the Netherlands (Van Daalen 2010). Due to matters of financial accountability, many institutions prefer to work with these effective or at least acknowledged interventions. In consideration of the Dutch emphasis on effectiveness, it is reasonable to limit our scope to interventions that are defined as being (theoretically) effective. All of the interventions in the database are used, but may exist alongside other, smaller programs.
The time-frame in which a program is developed (for home grown program) or implemented in the Netherlands (for foreign programs like Triple P) is our independent variable. We have made a distinction between six equal time-frames taking into account the time-frame of the research project (1990-current times): programs developed before 1990, programs developed between 1990 and 1995, between 1996 and 2000, between 2001 and 2005, programs  Our dependent variables are 'type of intervention' , 'type of support' , 'targeted family member' , 'targeted age group' and 'goals of intervention' . The variable 'type of intervention' consists of universally preventive interventions, targeted prevention/early interventions and curative interventions. Universally preventive interventions are those interventions that target the general parent population and which do not specify any "parental conditions. " They are available for all parents. Targeted prevention programs do specify certain conditions-which vary between programs-and are offered to specific groups of parents only. For instance, a program such as the Community Mothers Program (Moeders informeren moeders) is specifically developed for young first-time mothers from a low social-economical background, whereas a program like Pedagogical Advises (Pedagogisch Adviseren) is developed to offer help and assistance to all parents in Dutch society. Curative interventions are interventions for parenting situations in which problems have already occurred, and the interventions specifically aim to tackle these problems, for example, Multisystemic Therapy that helps parents in dealing with children who are exhibiting serious externalizing problem behaviour such as drugs abuse, truancy and aggression.
'Type of support' consists of parenting support programs, early education programs and the so-called general youth care programs. Parenting support programs are programs that, following Daly's (2013) definition, are aimed at supporting parents in the upbringing of their children. Early education programs [Vroeg-en voorschoolse educatie] focus specifically on the (often cognitive) development of children and are usually offered at kindergarten, school or child-and day-care centres. General youth care programs are mostly aimed at children themselves and are offered at professional youth care organization-either residential or ambulant.
The variable 'targeted family member' divides programs into programs aimed at parents, programs aimed at children or programs aimed at the family as a whole as this is not always directly related to the kind of support that is offered. Early education programs for example sometimes also include parents and help parents giving form to specific issues in their family life. Thus, early education programs may target the family as a whole as opposed to children only. The assumption of preventive care is also that help is offered in the early stages of problem behaviour, or before problems even occur. The age-period the interventions target is therefore also included. This division is based on generally acknowledged developmental psychological stages of development: baby (<2 years), toddler/infant (2.5-6 years), primary school (6.5-12 years), adolescence (12.5-18 years) and early adulthood (18 years or older).
Lastly, based on the skill-based approach of the child investment perspective, we have taken the specific goals of the interventions into consideration, with the assumption that the turn to parenting coincides with interventions being aimed towards teaching 'correct' skills and behaviour rather than investing for example in parent-child relations or social support networks. The different goals were structured according to 10 qualitatively different sets: information; skills; cognition/social information processing (SIP); behaviour; abuse/neglect/safety; sensitivity; motivation/attitude; dependent on need; self-efficacy/empowerment and communication.

Analysis
Chi-square calculations and standardized (expected) counts were used to deduce specific trends. Due to the selection criteria for the programs, the number of programs that is included is not large enough to make more sophisticated analyses possible. In consideration of the small (observed and sometimes also expected) counts in some of the cells in the contingency tables, we also used Fisher's exact test to re-analyse the data, which allows for small counts (see, e.g., Mehta and Patel 1986). There are minor differences between Chi-square and Fischer's exact test, but none of these differences affected the underlying hypotheses. Unless otherwise specified, results of the Chi-square analyses are given in the results section.
We first examined all (theoretically) effective interventions to see if and how policy measures are translated into practice and whether a turn to parenting has taken place. Second, we looked at the sample of parenting support interventions explicitly in order to find out whether policy tendencies described in this article can be found within this type of intervention solely (i.e., geared towards preventive forms of care or towards a younger age group).
First analysis showed that only five of the intervention goals are referred to frequently: information (36.9 %), skills (69.2 %), cognition/SIP (13.8 %), behaviour (27.7 %) and empowerment/self-efficacy (24.6 %). The other goals only represent a small percentage and were therefore not included in the analyses. It should be noted that interventions can include multiple goals, and goals are therefore not mutually exclusive. Results also show that 16 interventions were developed from 1990 to 1995 (9.5 %). In the time-frame 1996-2000, the number of programs doubled with 35 interventions developed during that time (20.7 %). The time-frame 2001 to 2005 also saw a rise in support programs, with 55 interventions developed in that time (32.5 %). From 2006 to 2010, 43 interventions were developed (25.4 %). No intervention within this database was developed after 2010, 6 interventions (3.6 %) have been developed before 1990, and of 14 (8.3 %) interventions it was unknown in which year these were developed. These three groups are left out of further analyses, since they, respectively, do not represent any interventions, do not suit the time-frame of this study, or should be regarded as 'missing value' since the independent variable is unknown. An overview of the descriptives is given in Table 1.

Changes in child and family support from 1990 to 2010
Chi-square analysis showed no significant differences between the specific periods and the type of program (p = 0.240, df = 6). However, a slow but certain growth in preventive measures can be found (see Table 2). Although the time-frames 1990-1995 and 1996-2000 show that fewer preventive programs were developed than expected, the last two time-frames (2001-2005 and 2006-2010) show a growth in preventive programs. The time-frame 1996-2000 saw a rise in the developments of curative interventions (compared to the other kinds of interventions) but this growth has slowed down in the years afterwards, whereas both preventive and targeted interventions have seen a growth in development in final the time-frame, 2006-2010.   Regarding the type of support that is offered, Chi-square analysis shows a significant difference (p = 0.042, df = 9), but not in the expected direction towards an increase in parenting support. As Table 2 shows, fewer parenting support interventions have been developed after the year 2000 than would be expected if there had not been any development at all. In contrast, early learning programs and general youth care programs saw a rise after the year 2000 (see Table 2).
Following the notion of the turn to parenting, we also expected interventions to be targeted increasingly towards parents. The results, however, show no significant differences between the time-frames and the targeted family member of the program (p = 105, df = 9: see Table 3) nor can we find a clear trend. In absolute numbers, however, we find that most programs target children themselves, and that these kinds of programs have more than doubled since 2001. Programs targeting parents remain rather constant (see Table 3).
The results also show no significant differences (p = 0.105, df = 9) regarding the targeted age group, nor is any clear trend visible. Although interventions aimed at babies (younger than 2 years) do show an increase in development in the last time-frame Table 2 Results of χ 2 analysis of time-frames compared to kinds of programs and kinds of support * p < 0.05 (χ 2 , p = 0.045; Fisher, p = 0.029)   [2006][2007][2008][2009][2010], fewer of such interventions were developed in the years before (see Table 3). The idea that preventive programs also imply that a younger age group is targeted is not supported by these data. The large number of programs targeting the 6.5-to 12-year age group may reflect the growth in early education programs.

Changes in parenting support from 1990 to 2010
Looking at parenting support programs specifically, targeted preventive interventions dominate the field (50.8 %), although differences are not significant (p = 0.295, df = 6). This is mostly due to the major increase of such programs in the last time-frame, with almost a third of these kinds of programs being developed from 2006 to 2010 (see Table 4). Thus, specific groups of parents are expected to need support instead of a more general idea that parenting (for all parents) has become more difficult and in need of support.
For parenting support programs specifically, more than half of the interventions include the youngest age range (0-2 years) with a large increase in the last time-frame. Yet no significant differences are found (p = 0.227, df = 9) and there are no other clear trends (see Table 4). It should be noted that this analysis is calculated on the starting age of the interventions. The older age groups might therefore be included in these programs as well.
Not surprisingly parenting support interventions target parents mostly (68.8 %), yet some interventions target the family as a whole. No significant differences are found (p = 0.284, df = 3) and the number of programs targeting parents remains stable. However, a surprising shift has taken place in the time-frame 2001-2005 when more interventions have been developed-both compared to the expected count and compared to the time-frames before and after-that target the family as a whole (see Table 5). This coincides with an increase in curative interventions (see Table 4) as well as with the implementation of the as-as-as policies (Matthijs and Vincken 1997) and can be explained by a change in ideas in which treatment for serious problems had to be organized "as close to home as possible", meaning within the family. In line with the social investment idea, the results in Table 5 show that most parenting support interventions  [2006][2007][2008][2009][2010]. We also see a trend indicative of increasing attention for behaviour and decreasing attention for cognitive issues. None of the results, however, are significant (p = 0.444, df = 3).

Discussion
The research described in this paper centred around the translation of policy into practice, and how ideas and actors affect this translation (Béland 2009;Campbell 2002;Yee 1996). In the Netherlands, parenting support was occasionally highlighted as a theme in governmental reports since the late 1980s onwards, but most prominently with the Ministry of Youth andFamily (2007, 2010). This Ministry, alongside international developments such as the ratification of the UNCRC, has been an important actor in dominating the discourse regarding parenting and preventive care, implying a specific 'turn to parenting' . Yet, although our quantitative findings show a growth in the number of targeted and preventive interventions in the 'general population' of interventions, this growth is not significantly different from the growth of other types of programs (curative interventions). Moreover, the data show a significant growth in early education programs, as opposed to other types of support, and also show that interventions target children themselves more than they target parents. These findings contrast the notion of a "turn to parenting" as highlighted in policy, and indicates that the emphasis on the role of parents is policy framing rather than actually taking place in practice. Literature suggests that parents are the linchpin in the healthy development of children (Gillies 2012;Kutscher 2012;Lister 2006). Yet, underlying goals of international developments such as the ratification of the UNCRC and general social investment approaches are oriented towards children themselves. These ideas are mirrored in the findings of this study, which may be indicative of policy not developing in isolation (Wincott 2010;Béland 2009) as well as of the time it takes for policy ideas to change practice. The Netherlands ratified the UNCRC in 1995, but a significant growth in early education programs did not take place until after 2000. Taken together, our findings indicate that change is work in progress, rather than a sudden paradigm shift as a turn to parenting would indicate (Hall 1993). Several national and international ideas are reflected in the way support programs are given form, thereby for instance solidifying child-centred ideas (Béland  Campbell 2002;Yee 1996). Also, the growth of early education programs and of targeted preventive interventions is indicative of the Dutch child wavering between public and private concern (Van Daalen 2010). The respect for the privacy and autonomy of families in Dutch society means that (1) there must be some indication of problems before the government is allowed to interfere, and (2) respecting the privacy of families whilst wanting to ensure the skills and well-being of children, means that the responsibility of that well-being is shifted towards other institutions such as schools and preschools. The normative framework regarding privacy appears to have hindered an actual paradigm shift, or a turn to parenting (Hall 1993). One may conclude that in the Netherlands developments such as the UNCRC and the influence of social investment theories have led to an "educationalization" of childhood-educating children through all kinds of institutions (including the family) to enable them to find their proper position in society (Reynaert et al. 2009;Campbell 2002;Yee 1996).
The findings of the sample of parenting support interventions specifically show that this specific type of program is predominated by targeted preventive interventions, which suits social investment ideas that target specific social risk, and hence specific social risk groups (Jenson and Saint-Martin 2006). These programs are also mostly aimed at the youngest age group (<2 years) and are indicative of a skill-based approach which suits the above-mentioned child investment perspective. Furthermore, the growth in parenting support interventions seems to have taken place mostly before 2007, which may indicate that policy follows practice rather than vice versa. In other words, ideas that already exist in society (e.g., via world-culture ideas) are eventually reified in policy measures, thereby stabilizing their influence (Campbell 2002;Yee 1996). Moreover, during the time-frame of this research , the Netherlands have been governed by seven different governmental coalitions, in which either the Christian Democratic Party [CDA] or the Labour Party [PvdA] dominated. These political changes are not reflected in our data, and may be indicative of the notion that ideas are framed in such a way that it suits existing notions or that existing institutions filter what is put in practice and what not (Campbell 2002;Yee 1996).
It should, however, be noted that parenting support interventions dominate throughout the years of this study, which makes it harder to find a specific turn. The notion that parenting support is given preference over other kinds of interventions still stands, but in the Netherlands this is not a real paradigm shift (Hall, 1993). Also, a program like the Australian Triple P positive parenting program (Sanders 2004: translated for the Dutch setting in 2004) has been extremely popular within the professional field of youth support in the Netherlands. There are currently 500 organizations offering the Triple P Program, in 204 (out of 403) municipalities. Also, 13,500 professionals have been trained in this program (Netherlands Youth Institute 2014). Triple P increasingly dominated the field of parenting support and this popularity may have "suppressed" the development and use of other parenting support programs, representing a certain bias within this field. One program may have taken the place of what could have been several programs, which would have increased the number of parenting support programs after 2005 in this study, thereby influencing our results. Another explanation may be that a turn to parenting did take place before 1990 and is therefore not found in the data we have collected. This, however, seems unlikely as youth policy in the Netherlands was already highly individualized since the 1960s (Timmerman 2009). Again, normative ideas about the privacy of the family seem to result in a focus on the individual (i.e., children) rather than the family.
This study comes with some limitations which imply caution in generalizing our findings. First of all, our sample of interventions is limited, which prevents the use of more elaborate statistical analyses. Latent Growth Modelling for example would be suitable in this regard, but requires longitudinal data which is not at our disposal. Second, our distinction of time-frames in which the interventions are developed is a practical but somewhat arbitrary choice: the year of development for example does not give any information about which programs are mostly used in practice in that time. Usage information is not available but would have given a more realistic picture of what happens in practice. Comparably, and as we have already mentioned before, not all interventions are included in the database of the Netherlands Youth Institute, which means that not all interventions that are used in practice are included in this research. Again, that would have given a more realistic picture of what is happening "on the floor. " Third, except for one, none of our results show any significant differences and findings should thus be interpreted with caution. Further research should address these matters for example by including usage information or by acquiring longitudinal data. Last, we do not have the advantage of having comparative data, for instance, from other European countries or from the USA. A comparative study could have given more insight in how ideas turn into practice (or not), and could be taken up in future studies. Despite these limitations, however, there are certain trends visible and these should be acknowledged.

Conclusions
The study presented in this article shows the complex and dialectical relation between policy and practice. Emphases of the policy process (in this case the emphasis on parenting) are not necessarily or directly related to developments in practice (in this instance the growth in early education programs), and both fields seem to mutually influence each other. In line with discursive-institutional theories, this study empirically shows that a variety of ideas and actors influence this process (Schmidt 2010;Campbell 2002;Yee 1996). To our knowledge this study has been a first attempt to address the translation of policy into professional programs, thereby quantifying data with regard to the public debate about matters of privacy, parental duties and state interference. Although the limitations of this study do warrant caution, we also believe that it opens up new possibilities gaining insight into the complex ways in which policy decisions impact the practice of social support institutions and vice versa. This can be relevant for other policy fields and/or social issues as well, for instance educational policy or policy measures regarding care for the elderly (see for an example Kalis et al. 2005). Also, although this was not the main intention of this article, this kind of research, which endeavours to combine policy developments with changes in practice, also provides valuable information to the debate about the role of ideas and discourse in policy analyses (Schmidt 2010). Such analyses could be detailed in future studies, for example in a case study centring on two or three of such interventions.