Risk and protective factors of physical violence towards children
-An international systematic review mirrored against Finnish Child Victim study
UNIVERSITY OF EASTERN FINLAND Faculty of Philosophy Department of Educational Sciences and Psychology Subject of Psychology Master’s Thesis Annemari Halme 275147
2
Contents
Abstract ... 3
Tiivistelmä ... 4
1 Introduction... 5
2 Data and research methods ...10
3 Results of the systematic review ...14
4 Results of the Finnish study and comparison to systematic review...18
6 Discussion ...22
References ...25
Appendix 1. Results of systematic review: risk factors ...31
Appendix 2. Results of systematic review: protective factors ...32
3
Abstract
Background: Despite the fact that violence towards children has lessened in Finland, there is a small minority of children who are target of severe violence. It is a very important task of authorities to find and help the children, who are in greatest danger of having severe long-term consequences. These are the children, who are target of continuous, multiple types of violence by various perpetrators. They are called poly-victims, and they also have the least number of protective factors in their lives. The aim of this research was to find risk factors that increase the probability of child experiencing violence. The aim of the study was to review international studies of risk and protective factors from the last 30 years, and compare them with a high-quality Finnish study which is based on self-report of 11 000 Finnish school children. With the help of Bronfenbrenner’s bioecological theory, this study brought together a wide range of risk and protective factors for violence against children.
Methods: Two methods of research were used. First, a systematic review was performed to summarize the international research on risk and protective factors of violence against children done during the last 30 years.
The results were then divided into different systems of Bronfenbrenner’s theory. Then, a quantitative analysis was conducted on Finnish Child Victim Study data to state which protective and risk factors could be found in the Finnish data and which of them were the most significant. Quantitative analysis consisted of cross tabulation, correlations and linear regression analysis.
Results: According to systematic review, significant part of risk and protective factors concerned home and family: Parents may suffer stress from parenting, ill health, substance abuse, poor finances, great number of children or their special needs. These factors increased the risk of being violent towards one’s own children.
Culture of violence was also an important risk factor, and it showed itself as a permissive attitude in society and families, as well as use of violence in intimate partnerships or in parents’ childhood. The most important protective factors were related to safe and nurturing relationship to parents and support from friends, child - oriented routines at home (eating, sleeping etc.), resilience of the child and help of authorities.
Despite the fact that only a small minority of Finnish children suffer from severe violence, almost half of the children have at least one experience of it. Accumulation of violence and social problems surfaced clearly in cross tabulations of the Finnish data, where the children with greatest number of violent experiences reported seeing greatest number of intimate partner violence incidents, binge drinking of parents and a poor or very poor financial situation of the family. The Finnish data also showed, that these children also had the least number of protective factors, i.e. they were without a trusted friend or safe relationship to parents significantly more often than the children with no violent experiences. They also ate together as a family much less frequently compared to children without violent experiences.
Conclusions: Violence towards children is a complex problem. To locate and help the children in greatest danger, the officials must combine several information sources concerning the family. Despite the fact that many of the risk and protective factors are related to the child himself or his parents, many can be influenced
4
from outside: by kindergarten, school, child health center, child welfare or by police or legislation. Protective factors may be built and strengthened, and risk factors can be weakened by supporting children and parent in the way that they need. It is the children with the greatest amount of risk factors and least number of protective factors that are in danger of suffering serious long-term consequences, both mental and physical. The Finnish data is fairly in line with the results of international systematic review, the only difference being than unemployment of parents or child having a trusted friend did not exhibit significant differences in Finnish data.
Tiivistelmä
Tutkimuksen tausta: Huolimatta lapsiin kohdistuvan väkivallan vähentymisestä Suomessa, pieni osa lapsista altistuu vakavalle väkivallalle. Viranomaisten tärkeänä tehtävänä on löytää ja tehokkaasti auttaa niitä lapsia , joiden riskit saada vakavia pysyviä seurauksia ovat suurimmat. Näitä ovat lapset, joihin kohdistuu jatkuvaa, usean tyyppistä väkivaltaa eri tekijöiden taholta (poly-victims), ja joilla on vähiten väkivallalta suojaavia tekijöitä elämässään. Tutkimuksen tarkoituksena oli koota yhteen kansainvälisiä tutkimustuloksia väkivallan riski- ja suojatekijöistä 30 vuoden ajalta, ja verrata niitä laadukkaaseen suomalaisaineistoon, joka koostui 11 000 suomalaiskoululaisen vastauksista. Bronfenbrennerin bioekologisen teorian avulla tämä tutkimus kokosi laajan katsauksen lapsiin kohdistuvan väkivallan riski- ja suojatekijöistä.
Tutkimusmetodit: Tutkimusmetodeja oli kaksi. Ensin tehtiin kirjallisuuskatsaus kansainväliseen tutkimukseen 30 vuoden ajalta käsittäen riski- ja suojatekijät, jotka jaettiin kuuluviksi eri ympäristöihin Bronfenbrennerin bioekologisen teorian mukaisesti. Tämän jälkeen tutkittiin tilastollisin analyysein, mitkä riski- ja suojatekijöistä todentuivat suomalaisessa Lapsiuhritutkimuksen aineistossa, ja mitkä kirjallisuuskatsauksesta löytyneet suoja- ja riskitekijät ovat merkittävimmät tässä suomalaisessa aineistossa. Menetelminä käytettiin ristiintaulukointia, korrelaatioita ja lineaarista regressioanalyysia.
Tulokset: Kirjallisuuskatsauksen mukaan merkittävä osa riski- ja suojatekijöistä liittyy kotiin ja perheeseen:
erilaiset vanhempia kuormittavat tekijät, kuten stressi vanhemmuudesta, terveyspulmat, päihteiden käyttö, taloushuolet, suuri lapsimäärä tai lasten erityistarpeet lisäävät riskiä tehdä väkivaltaa omille lapsille. Tärkeä riskitekijä on myös väkivallan kulttuuri, joka näkyy joko sallivana asenteena yhteiskunnassa ja perheessä, väkivallan käyttönä parisuhteessa tai vanhemman lapsuudessa. Suojatekijöistä tärkeimmät liittyvät lapsen hyviin ihmissuhteisiin, rutiineihin ja elämän säännöllisyyteen, lapsen resilienssiin sekä viranomaisapuun eri muodoissaan.
Vaikka vain pieni vähemmistö suomalaislapsista kärsii vakavasta väkivallasta, melkein puolella lapsista on vähintään yksi väkivaltainen kokemus. Suomalaisessa aineistossa väkivallan ja muiden sosiaalisten ongelmien, kuten alkoholinkäytön, kumuloituminen näkyi erityisesti ristiintaulukoinnissa, jossa eniten väkivaltaa kokeneet suomalaislapset raportoivat nähneensä eniten parisuhdeväkivaltaa ja vanhempien humalajuomista. He myös kokivat perheen taloustilanteen useammin huonoksi tai erittäin huonoksi. Näillä
5
samoilla suomalaislapsilla oli myös vähiten suojatekijöitä, eli he olivat muita useammin vailla luotettua ystävää tai hyvää suhdetta vanhempiin. Heidän perheensä myös söivät harvemmin yhdessä, kuin niiden lasten perheet, joissa lapset eivät kokeneet väkivaltaa.
Johtopäätökset: Lapsiin kohdistuva väkivalta on monimutkainen ilmiö, joten eniten vaarassa olevien lasten löytämiseksi viranomaisten on yhdistettävä paljon erilaisia perhettä koskevia tietoja. Vaikka monet riski- ja suojatekijät liittyvät lapseen itseensä tai vanhempiin, voidaan hyvin moniin näistä vaikuttaa kodin ulkopuolelta; esimerkiksi päiväkodin, koulun, neuvolan, lastensuojelun, poliisin tai lainsäädännön toimesta.
Suojatekijöitä voidaan rakentaa ja vahvistaa, ja riskitekijöitä voidaan pienentää tukemalla vanhempia ja lasta heidän tarvitsemallaan tavalla. Ne lapset, joihin kohdistuu eniten riskitekijöitä, ja joilla on vähiten suojatekijöitä, ovat suurimmassa vaarassa saada väkivallasta pitkälle aikuisuuteen vaikuttavia vakavia henkisiä ja fyysisiä seurauksia. Suomalaisaineisto on melko yhteneväinen kansainvälisen kirjallisuuskatsauksen tulosten kanssa, joskaan vanhempien työllisyystilanteen, tai sen, onko lapsella luotettu ystävä vai ei, ja väkivallan kokemusten määrän välillä ei löytynyt merkitseviä eroja suomalaisaineistossa.
1 Introduction
The great majority of Finnish children are doing well and are not suffering from violence. However, the number of Finnish children who experience serious violence has stayed the same in 1988, 2008 and 2013 when Child Victim Study was done. (Fagerlund, Peltola, Kääriäinen, Ellonen & Sariola, 2014, 54-55) These children typically face multiple types of violence, and they have the highest risk of adverse consequences. International research and discussion of violence against children typically concentrates on one type of violence (e.g. sexual violence). This study aims to form a comprehensive picture of the risk and protective factors of violence.
Violence is a societal problem: laws, practices, resources and attitudes can be guided and directed so, that risks are lessened and protective factors are supported. Before doing this, it is wise to point out how the international study results apply to Finnish country and society. The study was conducted mainly from the viewpoint of Finnish officials, who decide on allocation of scarce resources of child welfare or police: They need to know which children are high risk and which ones can wait.
There is a multitude of research on variety of types of violence and its potential outcomes for children. Children can be physically or sexually abused by parents, caretakers or siblings, or somebody outside their family, such as a hobby tutor or peers. Physical violence towards children includes for example shaking, spanking, kicking, pinching, hitting, hitting with an object, slapping, beating up, grabbing and choking (Clement & Chamberlain, 2007, 1005). Multitude of immediate and long -term impacts have been linked to childhood violence experience, such as problematic substance abuse (Levenson & Grady, 2015, 36; Sartor, Waldron, Duncan, Grant, McCutcheon, Nelson, Madden, Bucholz & Heath, 2013, 998), delinquent behavior (Watts & McNulty, 2013), mental health problems such as PSTD (Kulkarni, Graham-Bermann, Rauch & Seng, 2011), depression
6
(Kessler & Magee, 1994), anxiety (Cambron, Gringeri & Vogel-Ferguson, 2014) , eating disorders (Rayworth, Wise & Harlow, 2004), suicidality (Hardt, Bernert, Matschinger, Angermeier, Vilagut, Bruffaerts, Girolamo, Graaf, Haro, Kovess & Alonso, 2015) and risk for revictimization (Cuevas, Finkelhor, Clifford, Ormrod &
Turner, 2010).
Children can have single experience of violence (called “mono-victims”), repeated experiences of one type of violence, called “repeat victims”, or several types of violent exposure, called “poly-victims”, e.g. sexual abuse and physical abuse (Hamby & Grych, 2013, 18-19). We know that violence tends to accumulate to same children (Saunders 2003, 359-360). Poly-victimized children have significantly higher risk of mental health problems compared to non-victimized children, and their risk is higher than that of repeat victims (Finkelhor, Ormrod, Turner & Hamby, 2005). Thus, the more various types of violence the child experiences, the worse the prognosis for their mental health.
Despite several risk factors detected for violence against children, the research findings are scattered and the big picture is unclear (Peltonen, Ellonen, Pösö & Lucas, 2014). This is due to several reasons; such as the specialization of research, the disciplinary silos that mostly study sexual abuse, bullying, teen dating violence and punitative violence separately. The same applies to the risk and protective factors (Hamby & Grynch, 2013, 1). There has not been much sharing of knowledge, methods or research approaches. This creates many difficulties, one of the biggest being limited ability to fully understand violence towards children, and overlapping nature of different types of violence. (Saunders, 2003, 368-369) Also, not all violence is always reported, and when it is, different authorities use different reporting methods. Results also depend on who was interviewed; child, parents or authorities.
The aim of the study was to review international studies of risk and protective factors from the last 30 years, and compare them with a high-quality Finnish study which is based on self-report of 11 000 Finnish school children. With the help of Bronfenbrenner’s bioecological theory, this study brought together a wide range of risk and protective factors for violence against children. It looked into the family, school, community and culture for sources of risks and protection, and it included the main forms of physical and sexual violence.
Ideally, this and future research will be of assistance to police and child welfare workers estimating the risk of violence of an individual child.
In this study, risk factors were individual and contextual variables that increase the likelihood of violence towards children. Indeed, some characteristics and environmental conditions may not be causally related to violence, instead they correlate with it. When these correlative factors accumulate, they increase the likeliho od of a particular child experiencing violence. A large number of risk factors predict a high probability of violence.
(Tajima, 2002; Hampton, 1999, 70-71) To discriminate risk factors that have a causal relationship from those which are correlative, some researchers use term risk markers when relationship is correlative, and term causal risk factors when the relationship is causal (Lösel & Farrington, 2012, 10).
7
Protective factors can lessen the effect of risk factors by disrupting the causal chain or by inhibiting its onset (Tajima, 2002, 123-124). Existence of protective factors predict a lower probability of violence. Risk and protective factors are linked to each other and they interact with each other; they often are the two sides of the same coin. For example, disability is a risk factor and absence of disability, i.e. good physical and mental health, is a protective factor. Protective factors are often discussed together with the concept of resilience, which means person’s elastic biopsychosocial adaptation to stressful situations and difficulties in life (Lösel
& Farrington, 2012, 9). Resiliency defines psychological attributes of an individual which may protect against negative outcomes (Edwards, Probst, Rodenhizer-Stämpfli, Gidycz & Tansill, 2014) or ability to bounce back from adversity (Lowenthal, 2001, 169). In case of violence, resilience could mean ability to defend oneself before and when attacked, ability to get help from parents, teachers and officials, being aware of and avoiding potentially dangerous situations, ability to cope better with the violent experiences and take the stress involved.
There exists no consensus regarding the term protective factor. Lösel & Farrington divide protective factors into direct protective factors and buffering protective factors. Direct protective factors function as preventing factors per se – without taking other factors into account. Buffering protective factors work only when risk factors are present: they function as moderating factors, and attenuate the effect of the risk factor. (Lösel &
Farrington, 2012, 9) In the systematic review here, a factor was called protective when it helped children cope with violent experience, measured by eg. readiness for school, less externalizing or internalizing problems, better school performance. It didn’t necessary lessen the amount of violence. Thus, in this study, protective factors also altered child’s responses to adverse events, so that potential negative outcomes could be avoided.
Child maltreatment and child abuse are overlapping concepts, both including the element of violence. Child maltreatment is defined by World Health Organization as “all forms of physical and/or emotional ill-treatment, sexual abuse, neglect or negligent treatment or commercial or other exploitation, resulting in actual or potential harm to child’s health, survival development or dignity.” (World Health Organization, 1999). Other definitions of maltreatment include physical, emotional and sexual abuse, abandonment and neglect committed against child particularly by caregiver (Hong, Lee, Park & Faller, 2011, 1058). Child abuse is usually referring to sexual abuse or physical abuse. The term physical child abuse is defined by the Consultation on Child Abuse Prevention as “…results in actual or potential physical harm from an interaction or lack of it, which is within the control of parent or caretaker…” (World Health Organization, 1999, 15). Child physical abuse term stresses the position of the responsible adult, and the dependency of the child of the caretaker.
The concept of physical violence was used here to emphasize the physical harm done by violence and the criminal nature of the violent act. Physical violence has been defined as being “carried out with intention, or perceived intention, of causing physical pain or injury to another person” (Wolfner & Gelles, 1993). Physical violence toward children typically include punching, kicking, hitting, slapping, beating up, grabbing and choking. Finnish child victim survey also included hitting with an object, attacking with knife or gun (Fagerlund et al. 2014, 135).
8
Sexual violence is also included into concept of physical violence. Child sexual abuse has not been unequivocally defined in international research. Laws are varying in different countries; Finnish law currently defines child sexual abuse, as sexual act towards a child (e.g. touching, masturbating when is child watching, showing porn movies to a child, sexual messages or stories, showing genitals to a child, also via webcam), buying sexual services from a minor, persuading a child for sexual purposes (child grooming) (http://www.finlex.fi/fi/laki/ajantasa/1889/18890039001#L20).
Bronfenbrenner’s bioecological model has been widely used to understand the etiology of child abuse (Zigler
& Hall, 1989; Hampton, 1999, 128). This study used bioecological theory to ensure all the causal agents of violence were taken into account. As we know from previous research, some risk factors are historical (e.g.
current laws regarding corporal punishment), some are contemporaneous (e.g. socio-economic status, poverty), some are cultural (e.g. tolerance of violence) and some are situational (e.g. colic). Some factors are parental (e.g. wife abuse, substance abuse) and some are characteristic of children (e.g. temperament). If a too narrow thinking is applied – for example looking for the risks only in the vulnerabilities of the child alone – it will promote a view where children are seen as victimized, passive and themselves the cause of the violence. It would also not increase the efforts of making the society safer for children, but would increase thinking where vulnerable children need to be protected by isolating them - not by changing the environment or teaching children how to protect themselves. (Hollomotz, 2009)
Bronfenbrenner’s PPCT -model developed in the 90s is a part of the bioecological theory, and it consists of four concepts: Process, person characteristics, context and time. Proximal process, the first P of the PPCT - model, refers to merged and dynamic relations between individual and his context. Bronfenbrenner called reciprocal interaction between human, other persons, objects and symbols proximal processes. Regular interaction between these over time is the catalyst of human development. Examples of proximal processes are playing with a young child, child-child activities, reading and learning new skills. (Bronfenbrenner & Morris, 1998, 996) By taking part in these activities, young child understands his place in the world, and can make sense of it.
What happens in the proximal processes, the engines of development, depends partly on types of person characteristics, the second P of the PPCT-model. Bronfenbrenner referred to these as demand, resource and force characteristics. He understood that biological and genetic aspects of the person are significant. He looked at them as personal characteristics that person brings into social situations. Demand characteristics act as an immediate stimulus to another person, such as age, gender, skin color and physical appearance. People form expectations based on these, and they may influence interactions - including violence. (Tudge, J., Mokrova, I., Hatfield, B., & Karnik, R., 2009, 200). For example, physical disability, if immediately visible, is a demand characteristic that is linked to heightened risk of violence (Jones, Bellis, Wood, Hughes, McCoy, Eckley, Bates, Mikton, Shakespeare & Officer, 2012, 899). Resource characteristics are not immediately visible. They include mental and emotional resources, such as past experiences, skills, intelligence and social and material
9
resources. (Tudge et al. 2009, 200) Also intellectual and physical impairments can be resource characteristics, and they also form a known risk factor for violence (Jones et al. 2012, 899). For example, lack of caring parents, lack of awareness of one’s physical boundaries and ability to defend them as well as past experiences of abuse form risk factors for violence. On the other hand, a warm relationship to one’s parents, contact to helpful child welfare personnel or access to high quality health services may serve as protective factors. Force characteristics include differences in temperament, motivation, persistence and similar factors (Rosa & Tudge, 2013, 253). For example, two children growing up in identical surroundings may have different risk potential if one is excessively crying as a baby – a risk factor for physical abuse (Hurme et al. 2008), and the other one has a calm temperament. Bronfenbrenner also noted that change can sometimes be quite passive; person changes environment simply by being in it, and others react to him based on his demand characteristic. Change can also be active, as when person has a drive and desire to change his life. (Tudge et al. 2009, 201) Resilience can be seen as a force characteristic.
Context, the C of the PPCT-model, means environment, and it includes four interrelated systems: microsystem, mesosystem, exosystem and macrosystem. Microsystems include any environments, where child spends great deal of time in activities and interactions: home, school, day care and neighborhood, for example.
“Microsystem is a pattern of activities, roles, and interpersonal relations experienced by developing person in a given face-to-face setting with particular physical features, and containing other persons with their temperament, personality and systems of belief.” (Bronfenbrenner, 2005, 148). This renewed definition of microsystem takes into account the intergenerational transmission of behavior patterns, such as violence: If a child experiences violence at home, one microsystem, the role of school or day care - other microsystem - becomes crucial in child’s welfare.
The mesosystem is a system of microsystems: it comprises the processes and linkages taking place between two or more microsystems, the interrelations between them (Rosa & Tudge, 2013 246). For example, if day care teacher pays attention to bruises of the child and asks parents about them, that might have several possible consequences for the child: the child gets help, parents never take him to the daycare again, he does not get help and violence continues, the child is told not to tell the truth about the bruises and abuse continues, violence takes a form that does not leave visible signs. Much of the parent’s reaction and thus consequences to the child, depend on the nature of relationship that has formed between day care teachers and parents.
Also, settings that children are not actually part of, have an impact on them: if parents are particularly stressed at work, or if they lose their job, stress of the parent will reflect itself into the parent-child relation. Parent’s work is an exosystem to the child, since he is not part of it, but it affects his life indirectly. Authorities represent another exosystem, which may influence child’s life: police, health care and child welfare, for example.
Finally, macrosystem means any group whose members share values, belief systems, hazards, resources, lifestyles, opportunity structures, life course options and patterns of social interchange. (Bronfenbrenner 2005, 148-150) Attitudes towards violence and corporal punishment in society are example of this.
10
PPCT-model completes Bronfenbrenner’s earlier work in many ways, most importantly by adding the element of Time, the T of the PPCT-model: A theory concerning children’s social contexts should acknowledge their dynamic, changing nature over time (Henrich, 2006, 1083-1085). These changes often trigger developmental processes; siblings are born, parents may divorce, school starts, health problems may occur. Changes can occur in the child himself, as biological changes, such as puberty or illness, or in the surroundings, like change of school, or best friend moving to another city, death of grandparents. All these factors play their part in the mosaic which determines child’s risks for violence at any one moment, and also the number of protective factors in his life at that moment. When circumstances change, the risk and protective factors change.
No one theory is enough to explain the dynamics of violence (Hamby & Grynch, 2013, 29-30). It is clear, however, that violence happens as a sum of stressors and supports, or potentiating and compensatory factors.
Simple causal links are hard to find; e.g. under some conditions parent’s substance abuse will contribute to violence and under other conditions it will not. There are contributing factors to violence rather than determining ones. (Belsky, 1993, 413-417; Saunders, 2003, 358-359) Often it is the balance between factors that increase the risk and factors that protect the child, that will determine if violence will occur, or how serious and on-going it will be (Belsky, 1993, 413-417; Saunders, 2003, 358-359).
In this study, the first research question was to systematically review the known risk factors for physical violence towards children. Perpetrator may by a parent, caregiver or other person, such as other child, sibling or hobby tutor. ‘Violence’ meaning here physical or sexual violence, but not emotional violence, threats or intentions to attack. This definition was made to limit the scope of the study. The second research question was to systematically review the known protective factors that lessen children’s risk for physical and sexual violence, and help him cope with its consequences. The third research question was to find out if some of the same risk and protective factors are found in the Finnish Child Victim Survey data gathered by Fagerlund, Peltola, Kääriäinen, Ellonen & Sariola (2014).
The current study applies systematic review to answer following research questions:
1) What are the known risk factors for physical violence against children?
2) What are the known protective factors that lessen physical violence against children?
3) Are some of the same risk and protective factors found in the Finnish Child Victim Survey data?
2 Data and research methods
Two research methods were used to achieve the aims of the research. First, a systematic review was conducted to manage the large amounts of international research that has been done in the field of violence towards children during the last 30 years. This helped forming an overview of study results of risk and protective factors. Then, a quantitative analysis was conducted to determine if the same risk and protective factors are found in the Finnish Child Victim Study. Some risk factors are society bound; for example, differences in
11
income are not very big in Finland compared to for example USA, where poverty is a significant risk factor for violence. The Finnish study also introduced the standpoint of the children, since the children themselves answered the questionnaire. It also added the hidden violence that has not been reported to adults or police.
A systematic review answers a defined research question by collecting and summarizing all empirical evidence (i.e. primary research) that fits the pre-specified eligibility criteria. It should give as complete and clear picture as possible of the phenomena. Systematic review starts with a research plan defining research questions, search terms and databases to be used, as well as inclusion and quality criteria. The reasons for a certain search strategy and inclusion criteria must be made explicit, so that the review can be replicated by a third party.
Failure to include all appropriate studies can result in biased view and incorrect interpretations. (Torgerson 2003, 6-11)
To search for the articles, uef.finna.fi was used (international e-articles) during the period of time between 11.1.-22.2.2017. Finna includes dozens of databases from different fields of science, e.g. Scopus, MEDLINE/PubMed, Social Sciences Citation Index and OneFile (GALE). Systematic review progressed as demonstrated by Figure 1: Two separate searches were done, one for risk factors and one for protective factors.
Peer reviewed, full text articles written in English between 1986 and 2017 were searched using Finna databases with the key words a) “violence toward children” AND “risk factors” b) “violence toward children” AND
“protective factors” in the abstract of the article. The 341 full text, peer reviewed research articles were screened to establish that they answered the research question and were quantitative, primary research. Of these 341 articles, 177 did not answer the research question(s). They dealt with e.g. with violent children and childrens’ reactions towards IPV (intimate partner violence), and were therefore excluded. Of the 341 screened articles, 133 were secondary research, qualitative research or duplicate articles (same study published in several journals), and thus were excluded.
12
Figure 1. Diagram illustrating study selection process January-February 2017.
There were 18 eligible studies mapping the risk factors and 15 studies mapping the protective factors. These research articles were read through carefully. Then the contents of the studies were organized into two charts:
risk and protective factors. The charts list authors, country of the study, year the study was published, size of the sample, age of the children studied, forms of violence researched, aim of the research and main results of the study. These can be found in Appendix 1 and 2. Then the risk and protective factors were selected from the both charts, classified according to subject, and arranged into one picture representing the bioecological model: microsystem (home and family, community), macrosystem (culture, values, law), exosystem (parents’
work, officials contacting the family) and mesosystem (social networks). Risk factors were marked with minus sign and protective factors with a plus sign. When the same risk or protective factor was found in several studies, this was marked with a number.
The data used for the second research method consists of the Finnish Child Victim Survey (CVS). It is an extensive questionnaire answered by 11 364 Finnish sixth- and ninth graders in 483 schools. Year’s 2013 CVS was a repetition of 2008 CVS, and the two are almost identical. Clustered sampling was used, and four cluster samples were selected from all over Finland. The sampling was done on class level, not on individual level.
Of the schools contacted, 78% took part in the study. The sample is representative of all Finland. (Fagerlund et al., 2014, 27) Parents were not informed of the study, nor were they asked for a permission, so the conflict of interest between them and the children was minimized. CVS data comprises of mostly unreported cases (Fagerlund et al., 2013, 9). CVS asks of child’s experience of violence according to Parent-Child Conflict Tactics Scale, and it covers both the last 12 months and the whole lifetime of the child. The first CVS was conducted in 1980, and it has been repeated in 2008 and in 2013. (Fagerlund et al., 2014, 25-26)
For purposes of this study, two new variables were formed on the CVS data and two already existing variables in CVS were modified. 1) The dependent variable measuring amount and variety of violence experienced by the child was formed. 2) An explanatory variable measuring the amount of intimate partner violence in the family was formed. The remaining explanatory variables that were modified, were: 1) child having a disease, illness or special need 2) employment status of parents. Other variables used here were in the same as in CVS;
they have not been modified or recoded.
The dependent variable measuring amount of violent experiences was formed by summing up 71 questions regarding child’s experiences of violence during his entire life, including last 12 months. The forms of physical and sexual violence included in this variable were: pushing, shaking, pulling hair, hitting, slapping, hitting with an object, punching with fist, grabbing, attacking, kicking, spanking, threatening with a knife or gun, using knife or gun, telling sexual jokes and stories, sexual touching, trying to kiss or kissing, suggesting intercourse or oral sex, forcing to intercourse or oral sex. Different surroundings were included: foster care or institution, hobbies, home or elsewhere. Perpetrators could be parents, siblings, hobby tutors, other children or other people. Questions regarding bullying, threats, emotional violence or stalking via phone or internet were
13
not included into the variable. Amount of violent experiences -variable classifies children according to exposure to multiple types of violence and multiple perpetrators they had experienced during their life, i.e.
amount of re-victimization. The formation of the variable was as follows: Each mention of different kind of violence was given one point, no experience was given zero points. There are 5 classes: “No violent experiences”, “1-2 violent experiences”, “3-4 violent experiences”, “5-6 violent experiences” and “Over 6 violent experiences”. This variable gave highest scores to children who have experienced violence by many perpetrators and the forms of violence have been diverse. Cronbach’s alpha for the variable was .819.
A variable measuring the number of intimate partner violence incidents witnessed by child towards mother was formed by summing up the answers of 11 questions concerning incidents of IPV witnessed by the child;
both during last 12 months and his lifetime: The forms of physical violence towards mother included in this variable were: pushing, shaking, pulling hair, hitting, slapping, hitting with an object, punching with fist, grabbing, attacking, kicking, spanking, threatening with a knife or gun, using knife or gun, other form of violence. The same questions regarding violence towards father were left out, since studies only show a significant relation between IPV towards mother and violence toward child. In IPV seen by the child-variable , each mention of different kind of violence was given one point, no experience was given zero points. A child who has never seen his mother being target of IPV in his life, got zero point. There are three classes for answers:
“No IPV experiences”, “1-2 IPV experiences”, “3 or more IPV experiences”.
A variable measuring disability, special needs and illnesses was modified into a dichotomic form. It discriminated children who have no such conditions (0 = “No disability, illness or special need reported”) from children who have one or several disabilities (1 = “One or more disability, illness or special need reported”).
A variable measuring unemployment of parents was formed by combining answers of two questions asking if mother is employed and if father is employed. If the answer was no, this was defined as unemployed. This may include retired because of sickness or other reason for not working. The answers regarding both parents were combined into variable Employment status of parents. Answers were classified into four classes 1 = “Both parents unemployed”, 2 = “One of the parents unemployed, 3 = “Both parents working” and 4 = “No info of employment data”.
A variable measuring binge drinking of parents was chosen to measure alcohol abuse of parents. Question was
“How often, during last 12 months, have you seen your parents clearly drunk?”. Answers were “Never = 1”,
“Once = 2”, “Couple of times = 3”, “Once a month = 4”, “Once a week = 5” or “Several times per week = 6”.
It was assumed here, that if child reports seeing parents drunk once a week or several times per week, alcohol abuse is probable. In CVS, there was also a question asking “How often alcohol is used in your family?”, but this was not considered relevant, since it is more the amounts that matter when abuse is diagnosed, not how often alcohol is consumed.
Financial problems of the family were measured by question “What do you think of the financial situation of your family?”. Answers were “Very good = 1”, “Fairly good = 2”, “Some financial difficulties = 3”, “Lots of
14
financial difficulties = 4” or “I don’t know = 5”. Support from friends was measured by question “Do you have at least one trusted friend, with whom you can talk about anything that worries you?”, answers “Yes = 1” and
“No = 2”. Relationship with parents was measured by question “Are you able to talk about your issues with your parents?”, answers “Yes = 1”, “To some extent = 2”, “Hardly at all = 3”, “Not at all = 4”. Child-oriented daily and weekly routines were measured by the question “How often do you eat together with one or both your parents?”, answers “Every day = 1”, “Several times per week = 2”, “Once a week = 3”, “Once a month
= 4”, “Less than once a month = 5”
After forming and modifying the needed variables, statistical analysis was conducted using IBM SPSS Statistics 24. Simple descriptive statistics were calculated using frequencies and percentages. The analysis consisted of crosstabulation, Chi Square tests and linear regression analysis. Because of the large sample size, Cramer’s V was calculated. Correlation was measured for Amount of violent experiences and Amount of Intimate Partner Violence seen by the child.
Crosstabulation was chosen as one method of analysis, because variables were nominal or ordinal, and also because the groups could be compared with help of adjusted standardized residuals: eg. how do the children who see their parents drunk several times per week differ in their experiences of violence from the children who have never seen their parents drunk? It was hypothesized that the variables produce significant differences in violence experience only in the extremities (seeing a lot of IPV, seeing parents drunk often, having lots of financial difficulties in the family), and this would show in crosstabulation.
Linear regression analysis was used to determine to what extent the independent variables explained of the violence experiences, and which variables were the most important explaining factors. The number of n’s was acceptable, and variables were not multicollinear (correlation under .7). Tolerance -values were between .845 - .995. VIF -values were 1.005-1.183. Cook’s Distance was .000 - .007. Normal P-P Plot of regression standardized residuals of the dependent variable was acceptable.
3 Results of the systematic review
The results of systematic review showed that both risk and protective factors can be found from all system levels. The results are gathered into Figure 2. In all the ecological systems, there are several factors changing constantly, both increasing and decreasing risk.
15
Figure 2. Risk and protective factors in micro-, meso-, exo- and macrosystems.
According to this systematic review, risk factors were very concentrated around violence happening at home, done by parents: All studies, except one, which dealt with convicted sexual abusers working with children . Majority of the violence experienced by children happens at home, but the result is still rather disturbing: we don’t know much about violence towards children outside home. Community violence was much less researched subject with five studies out of all 33, all mapping protective factors. Two of these studies used same data.
Other main finding was that many protective factors can be supported from outside the family, even if they exist in the family microsystem: health- and maternity services are essential in both mapping the risk factors and providing help and support before violence escalates. School and law are also important players in preventive work. Interaction between systems was clearly visible: e.g. child’s strong commitment to school is clearly his attitude, but it can be strengthened or weakened by actions in other systems; school, teachers, community as bioecological theory suggests. The detailed results of the studies including risk and protective factors are found in the Appendix 1 and 2.
As stated, most of the studied risk and protective factors were gathered around the microsystem, mostly home, family, child himself and his parents. Risks in microsystem could be classified into parental risks, parental attitudes and behavior and features of children. Parental risks were (maternal) parenting stress, young age,
16
alcohol and drug problem, intimate partner violence, no significant other at home and health problems of caretakers (mental illness, depression, PSTD). The strongest risks found were maternal parenting stress, intimate partner violence and alcohol and drug related problems in guardian. Parental attitudes and behavior included mother’s belief in necessity of corporal punishment, previous spanking of child and lower authoritarian parenting attitudes. The attitudes of parents are affected by general positive attitudes towards corporal punishment of children, which is a macrosystem factor. Features of children include number of children in the family (the more children, the higher risk for violence), young and premature children, children who react strongly when upset, demand more attention, have special needs or use alcohol. In general, the more attention and care the children need, the more likely is that they will face violence from parents.
Protective factors found in the microsystem of home and family could be divided into relationship between child and parent(s), living skills and routines, resilience and (emotional) intelligence. Relationship between child and parents and felt acceptance from parents were mentioned in eight studies, which made them an important protective factor. Another important protective factor were child-oriented routines at home and child’s living skills. This included regular routines in eating, sleeping and activities, which provide predictability in child’s life, and are especially important for maltreated children (Lowenthal, 2001, 169). It was also protective, if child could take care of these himself; daily living skills correlated with better performance in school despite maltreatment (Coohey, Renner, Hua, Zhang and Whitney, 2011). Many results indicated towards resiliency of the child: child’s emotional intelligence and IQ, child’s level of self-control and emotion regulation skills, child’s positive emotions and future orientation. High level of resiliency was a protective factor for girls (Tlapek, Auslander, Edmond, Gerke, Schrag and Threlfall, 2016). Protective factors that did not clearly fall into any of these categories were spirituality, which was however associated with greater resiliency (Howell & Miller-Graf, 2014, 1991-1992) and child’s strong commitment to school and viewing education important. Thus, school and daycare are places where routines and schedules may increase feelings of safety and security, which in turn may increase commitment, forming a positive circle. Avoiding revictimization was a protective factor, which can be viewed as an end result rather than protective factor on its own.
In community microsystem, only two risk factors were found; mother’s high exposure to community violence , which increased the probability of intergenerational continuity of abuse (Valentino, Nuttall, Comas, Borkowski and Akai, 2012) and child sexual abusers working with children. In Austrian study, 15% of convicted child sexual abusers had worked with children, teacher and priest being most common occupations.
Voluntary workers in youth organization (e.g. boy scouts) or sports coaches were also a common workplace for abusers wanting to get in touch with children. (Bennett & O’Donohue, 2014; Turner et al. 2014, 333) Risk factors in the macrosystem – culture, values and law – included being an immigrant child and positive attitudes towards corporal punishment. They were the sides of the same coin: an immigrant child from a culture with more violence favorable attitudes stick out from the statistics. Immigrant children in Sweden were
17
eight times more likely to be physically abused by their parents than Swedish children. The longer immigrants had been in Sweden, the less abuse was reported. Notably, 60% of these immigrant children were abused by somebody else than parent or caretaker, often by other children. This may imply increased bullying. The link between immigrant parents' unemployment and abusing their children was significant. (Lindell & Svedin, 2000) The study made in Vietnam showed that use of violent disciplinary practices on children was strongly associated with positive attitudes towards corporal punishment (Cappa & Dam, 2014) and a study in 25 countries showed that acceptability of violence makes it more usual (Lansford, Deater-Deckard, Bornstein, Putnick & Bradley, 2014).
Mesosystem, the system of microsystems, showed one important risk factor; diminishing social networks (Christoffersen, 2000), and on the other side, protective nature of friends. Friends were sometimes even more protective than close ties with the family (Howell and Miller-Graff, 2014; Saltzinger, Feldman, Rosario and Ng-Mak, 2010). There is no lack of evidence linking social isolation and lack of social support with elevated risk of child abuse and neglect. Moving often is one indicator of this. It could be that maltreating parents isolate the family on purpose by moving often, not inviting friends as well as not using the community resources that are available. (Belsky, 1993, 422) Being involved in a religious community was found to be a protective factor.
It can be seen as a part of larger context of social support in micro- and mesosystem; close friends and warm relations to family.
Risks and protective factors in exosystem typically include contact with authorities (e.g. police, prosecutors, social workers, psychologist, doctors, unemployment officers). Low income in the family and unemployment as well as financial worries of the parents were risk factors (Peltonen, Ellonen, Pösö & Lucas, 2014; Bardi and Borgognini-Tarli, 2000; Lindell and Svedin, 2000; Christoffersen, 2000). Parental criminality was found to be a risk factor in two studies (Christoffersen, 2000 and Crandall, Chiu and Sheenan, 2006). Research shows that crime is heavily concentrated in families, and that family violence is a correlate of maternal and paternal arrests. So, there is a relationship between family violence and criminal behavior. (Junger, Greene, Schipper, Hesper & Estourgie, 2013, 127) The systematic review also shows strong evidence of intergenerational transmission of violence: parent’s own experience of violence as a child strongly increases the likehood of physically abusing his own child (Peltonen, Ellonen, Pösö & Lucas, 2014; Saile, Ertl, Neuner and Catani, 2013; Tajima, 2000; Ellonen, Peltonen, Pösö and Janson, 2016).
The presence of authorities in the child’s life can be protective: Multiple reports from child protective services decreased the likehood that child will die as a result of abuse (Douglas, 2015) and a diagnosed behavior problem of the child was protective factor (Coohey, Renner, Hua, Zhang and Whitney, 2011). It was speculated that a diagnose of any kind will increase attendance of professionals, which could serve as a protective factor.
In the same vein, children remaining home after first time investigation for suspected maltreatment live with persistent risk of repeat maltreatment (Campbell, Thomas, Cook and Keenan, 2012).
18
It is worth mentioning that in risk factor studies, 11 out of 18 were answered by parents or caregivers, usually the mother. Four were based on reports (forensic or police) or population based register. In protective factor studies, 9 out of 15 were answered by parent and child, usually mother. None of the studies interviewed children without consent of his parents or caretaker. Only one study had fathers as sole participants (Ellonen, Peltonen, Pösö and Janson, 2016).
4 Results of the Finnish study and comparison to systematic review
The analysis of the Finnish data showed that 54.8% (n = 6229) of the children reported that they had not had any violent experiences in their life. One to two violent experiences were mentioned by 28.6% (n = 3254).
Almost ten percent, 9.7% (n = 1103), mentioned 3-4 violent experiences, three percent (n = 367) mentioned 5-6 experiences and four percent (n = 411) reported over 6 experiences. This means 45% of children in Finnish data reported having been in one or more situations where they had been target of some forms of physical or sexual violence.
A great majority, 97.8% of the children had not reported seeing any violence between their parents (n = 11 112). Only 1.6% (n = 149) of the children reported seeing 1-2 instances of IPV, and a minimal 0.9% (n = 103) reported seeing 3 or more incidents of IPV. A special need, illness or disability was reported by 24.2% (n = 2754) of the children, and 75.8% said they do not have any special need, illness or disability.
Almost third, 30.1% (n = 3422), had never seen their parents drunk, only 2.5% (n = 289) reported seeing parents drunk once or several times per week. Majority of the Finnish children, 79.6% (n = 9045), reckoned the financial situation of their family good or fairly good, and 11.7% (n = 1329) said there were some of lots of financial difficulties in the family. Only 8.1% (n = 919) of the children felt they cannot answer this question, which indicates that children are fairly well acquainted with the finances of the family.
In 71.4% (n = 8114) of families both parents were working and in 20.3% (n = 2304) both or one of the parents were not working.
A bit less than one in ten children, 7.1% (n = 812), said they don’t have a single trusted friend. A good half of the children, 58.3% (n = 6623), said they are able to talk about their issues with parents, 36.4% (n = 4131) had some difficulties in this and 4.8% (n = 548) said they cannot talk at all or hardly.
Majority of the Finnish children, 85% (n = 9668), ate together as a family every day or several times a week, 14.3% (n = 1621) ate together once a week or less often.
There were statistically significant differences in amount of violent experiences of the child and amount of experienced IPV by the child (X2 = 403.13, df = 8, p<.001, V=.133). Adjusted standardized residuals showed that the children who reported three or more incidents of violence toward themselves had significantly more often also seen their parents being violent towards each other (three or more times), adjusted residual being
19
11.8. Children with no violent experiences toward themselves were over-represented in group with no experiences of parents being violent to each other, adjusted residual being 12.2. Effect size was moderate.
Also, amount of violent experiences was only weakly correlated to amount of intimate partner violence seen by the child (r = .0148, p = .001).
There were also statistically significant differences in Amount of violent experiences and disability, illness or special need of the child (X2 = 196.45, df = 4, p<.001, V=.131). Adjusted standardized residuals showed that the children with one or more disability, illness or special need were under-represented in the group of the children who experienced no violence, adjusted residual being -11.3. Children with one or more disability, illness or special need were over-represented in all the groups who reported experienced violence, especially in a group with over 6 incidents of violence (adjusted residual 8). Effect size was moderate.
Binge drinking of parents and amount of violent experiences of child had statistically significant differences (X2 = 276.16, df = 20, p<.001, V=.105). Children who reported seeing their parents clearly intoxicated/drunken several times per week, were underrepresented in the group zero experienced violence (adjusted residual -4.3) and overrepresented in a group of 6 or more violent experiences (adjusted residual 9.5).
Children who experienced no violence toward themselves were over-represented in a group of never seeing parents clearly drunken (adjusted residual 9,6), even if alcohol was consumed. However, effect size was small.
A significant amount, 5056 children out of 11634, did not answer this question at all.
Child’s assessment of family’s financial situation and amount of violent experiences of child had statistically significant differences (X2 = 348.98, df = 16, p<.001, V=.088). Differences were stronger in the extremities:
excellent financial situation or lot of financial difficulties. Children who assessed family’s financial situation as excellent, were overrepresented in a group of no violent experiences (adjusted residual 11.1). When children reported some or a lot of financial difficulties, they started to appear underrepresented in the zero-violence group (adjusted standardized residuals -7.9 to -9.8). “Fairly good” financial situation was as common in families where children had experienced violence a little bit or a lot. Effect size was small.
Unemployment of parents and amount of violent experiences of child showed small statistical significance (X2
= 25.94, df = 12, p<.011, V=.028) which was supported by adjusted standardized residuals (4.0) when both parents were currently unemployed: the number of children with 6 or more violent experiences was overrepresented in a group with both parents unemployed. Effect size was moderate.
Child having at least one close and trusted friend to talk to made statistically significant difference to amount of violence experienced by the child (X2 = 54.04, df = 4, p< .001, V=.069). Children with no violent experienced were over-represented in a group of having a friend (adjusted residual 6.0), and children with no friend were over-represented in a group of children with 3 or more incidents of violence (adjusted standardized residuals varied from -2.6 to -4.0). Effect size was small.
20
Being able to talk to parents of one’s concerns made even more statistically significant difference to amount of violence experienced (X2 = 795.86, df = 12, p<.001, V=.153). Children with no violent experienced were over-represented in a group who could talk to parents, adjusted residual being 20.5. Children who stated they cannot talk to parents at all, were overrepresented in a group with most versatile violent experiences, adjusted residual being 10.1. Even children who stated they had some difficulties in talking to parents had more violence experiences than the children who had easy communication with parents. Effect size was moderate.
How often the family was eating together and amount of violence experienced by the child showed statistically significant differences (X2 = 331.55, df = 16, p<.001, V=.069). Children with no violent experiences were overrepresented in a group who were eating together every day (adjusted residual 10.5). Children with 6 or more violent experiences were overrepresented in groups eating together once a month or less than onc e a month (adjusted standardized residuals 6.2 and 7.2). Effect size was small.
Using linear regression analysis with Enter -method, the same variables were used in forecasting the number of violent experiences. The model with all the variables was statistically significant (F8,6224 = 100.47, p <.001, R2 = .114) and it explains 11,4% of how much violence the child has experienced. Betas, p- and t-values of the dependent variables are presented in Table 5. All factors except employment status of parents and child having at least one trusted friend to talk to were statistically significant. The most important factor explaining the amount of violence the child had experienced was child’s ability to talk about his issues with his parents (β = .172, t = 13.237, p <.001). The more there were difficulties in talking, the more risk there was that child had experienced violence. The number of intimate partner violence incidents seen by the child was the second most important explaining factor for the amount of violence the child had experienced (β = .123, t = 10.220, p <
.001). Third factor explaining most of the violence was the number of times child had seen his parents drunk during last 12 months (β = .111, t = 8.826, p < .001). Financial situation of the family had the smallest explaining power of all factors (β = .041, t = 3.314, p = .001).
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t Sig.
B Std. Error Beta
1 (Constant) -.227 .095 -2.391 .017
Number of intimate partner violence incidents w itnessed by child
.550 .054 .123 10.220 .000
Do you have a disability, illness or a special need?
.235 .029 .097 7.987 .000
21
How often, during the last 12 months, have you seen your parents distinctly drunk?
.100 .011 .111 8.826 .000
What do you think of the financial situation of your family?
.042 .013 .041 3.314 .001
Employment status of parents
.016 .024 .008 .665 .506
Do you have at least one trusted friend, w ith w hom you can talk about anything that w orries you?
.082 .051 .020 1.604 .109
Are you able to talk about your issues w ith your parents?
.284 .021 .172 13.237 .000
How often do you eat together w ith one or both your parents?
.072 .014 .065 5.156 .000
a. Dependent Variable: Amount of violent experiences
Figure 4. Coefficients of the linear regression.
The results of quantitative analysis show that the following risk factors that were found in international systematic review, were also found in Finnish data: Intimate partner violence seen by child; disability, special need or illness of the child; binge drinking of parents and poor financial situation of the family affect the probability of child having experienced violence. The following protective factors were also found in the Finnish data: child’s warm and close relationship to parents and child-friendly daily routines protect child against violence. Contrary to systematic review results, the following risk factors; unemployment of parents and child not having a trusted friend, did not affect the probability of violence in the Finnish data.
In sum, the quantitative analysis reaffirmed what we know based on the systematic review: the probability of child becoming target of violence consists of many risk and protective factors acting together. Only some of the findings of systematic review could be investigated quantitatively here. Some of the most important factors that had to be left out of this study were parents’ own experience of violence (intergenerational violence), maternal parenting stress and parent’s mental illness as risk factors and resiliency of the child as protective factor.
22
6 Discussion
According to systematic review, significant part of risk and protective factors had to do with home and family:
Parents may have additional parenting stress because of great number of children or their special needs. They themselves may suffer from ill health, substance abuse or they may worry about poor finances or unemployment. Especially when parents were not getting help in these situations, these factors increased the risk of being violent towards one’s own children. A culture of violence was also an important risk factor: a permissive attitude towards punitive violence in society and family, violence between parents or in their own childhood increased risk of being violent towards children. The most important protective factors were related to child’s safe, stable and nurturing relationship to parents, felt acceptance from parents or caregivers, parental warmth and quality of interaction between child and parents. Child’s high level of self-control and emotion regulation skills were important protective factors, implying that resilience protects children from violence and helps overcome its consequences. Support from friends and intervening authorities lessened violence, as did existence of child-oriented routines at home (eating, sleeping etc.) and child’s daily living skills.
Violence towards children is a serious problem also in Finland: almost half of the children have at least one experience of it. Accumulation of violent experiences and social problems was distinct in the Finnish data, where the children with greatest number of violent experiences also reported seeing greatest number of intimate partner violence incidents, binge drinking of parents and reported more often a poor or very poor financial situation of the family. The Finnish data also showed that children with highest number of violent experiences also had the least amount of protective factors, i.e. they were without a trusted friend or safe relationship to parents significantly more often than the children with no violent experiences. They also ate together as a family much less often than children without violent experiences. Contrary to international study results, unemployment of parents and child not having a trusted friend did not affect the probability of violence in the Finnish data. Despite severe violence towards children being rare in Finland, it should be kept in mind that almost half of Finnish children reported experiencing at least one violent incident, either physical or sexual.
Thus, there is a possibility of it escalating into severe violence at some point of child’s life, if protective factors fail or become missing and risk factors increase. Children are more prone to violence than adults, just because they are children and thus more vulnerable.
However, results of the study are encouraging, because risks of violence usually accumulate before violence becomes a serious problem: this leaves many chances for intervention. It also means there are many ways to intervene - to lessen the risks and strengthen the protective factors. For example, maternity health services survey substance abuse and offer rehabilitation when needed. They also survey mother’s experiences of violence both in her childhood and by her current partner. When the child is born, parents get information of nutrition, parenting, taking care of their relationship and health. Help is offered in form of psychology sessions, sleep schools and home care help, for instance. Some of the services are tied to social welfare to encourage parents to take their children to maternity health services.
23
Daycare may help lessening parenting stress by allowing parents time for work and rest, by delivering info on the development of the child, offering professionals for discussion and ways for parents and children to strengthen their relationship. One of the aims of the early childhood education is to support parents in parenting, and to act together with parents to secure balanced development and holistic wellbeing of the child (Law of the Early Childhood Education, Varhaiskasvatuslaki, 2 a §). Daycare can help child in forming a safe and nurturing relationship with the carers and to find friends. The routines and schedules of daycare help creating feeling of safety for the child. Since resiliency and emotion control are protective factors that can be developed in the individual, teaching resilience related skills (emotional control etc) at school or daycare could work as preventive method. Of course, not all children go to daycare, and if a case of violence is suspected, care must be taken to establish non-judgemental communication with parents.
Maternity health service, daycare and schools can help in recognizing signs of violence in the child, but special expertise is needed. Violence is also difficult subject to bring up. Resources are often not the bottle neck – it is more often a question of will and motivation of the authorities. However, not all parents use services of maternity health center or daycare – perhaps just the families that would need them most. Also, as bioecological model shows us, the study of violence towards children has been very concentrated around home and parents.
This may give us a too narrow view of the problem: Violence in internet and in hobbies should not be forgotten.
Laws concerning new forms of sexual violence, such as grooming, are always one step too late: problem already exists before a law is established. Laws change attitudes, as has happened with corporal punishment, but it takes time.
Study of violence has often been done from disciplinary silos of certain type of violence. This study provides a much broader view of the subject with a comprehensive picture of the risk and protective factors of violence.
A holistic view is essential in understanding the very nature of violence: it is not caused by a certain factor, but by many different risk and protective factors existing at home, social surroundings and in the society. This message showed clearly in this work. This study is also valuable in showing that the same risk and protective factors exist here in Finland as are reported in international reviews: Child Victim Study data is high quality and comprehensive with over 11 000 respondents and versatile questions. An additional merit of this study is its practical implication: the results of this study can also assist authorities in planning preventing methods that produce best results in our country. It may help police in prioritizing when investigating possible violent crimes against children: which child is in greatest danger?
Preventing severe violence towards children demands tight co-operation and information sharing of the authorities. The big picture of the risk and protective factors is only possible to form by co-operating. Any risk assessment tool will need information from different sources; police, teachers, kindergarten, child protection social services, hospitals, and most importantly: the family and the child themselves. Sharing information between officials is known to be very important, but often failed part of the process. In England, a Multi- Agency Safeguarding Hubs (MASH) is a screening mechanism that aims to collate information to form an
