Introduction

Fathers play an important role in the lives of their children (Cabrera and Tamis-LeMonda 2013; Lamb and Lewis 2013). The involvement of fathers in the family has been connected to positive outcomes for children including social, emotional, educational, cognitive development and adjustment (Baker et al. 2018; Cabrera and Tamis-LeMonda 2013; Lamb and Lewis 2013).

It is unclear whether the association between fathering and positive child outcomes applies to high-risk families (Zanoni et al. 2013). Fathers, like mothers, also pose risks when they are more negatively involved with the child. There is a gap in the research regarding the father’s role in child maltreatment cases (Brewsaugh et al. 2018; Dubowitz 2006, 2009; Wingrove et al. 2016) and, specifically, regarding the protective role or risk posed by fathers who are not identified as perpetrators in cases of child maltreatment (Shapiro et al. 2011; Storhaug and Øien 2012). Since the mothers, who are the primary caregiver of the child, are often also the maltreating parent, existing child maltreatment research tends to be more representative of mothers in general.

The current study explores whether households with a resident non-maltreating biological father are associated with an increased or decreased risk to the child compared to households with a non-resident biological father or resident substitute father, based on an analysis of child maltreatment court cases. In these cases, parental rights were terminated due to parental incapability to meet the child’s needs. The study therefore focused on the most severe child protection cases in Israel.

Literature review

The role of fathers in child maltreatment cases is a growing research area (Brewsaugh et al. 2018; Wingrove et al. 2016). The interaction between fathers and a family’s characteristics in promoting a child’s welfare or protecting a child from the risk of maltreatment is unclear (Dubowitz 2006, 2009; Guterman and Lee 2005; Lee et al. 2009; Risley-Curtiss and Heffernan 2003; Strega et al. 2009; Zanoni et al. 2013). It has been documented, in the literature, that fathers are not typically included in child welfare services and research (Brewsaugh et al. 2018). Lack of involvement in services may stem from various reasons such as sexism on the part of child welfare workers (Brewsaugh et al. 2018), negative attitudes toward fathers (Clapton 2009) and discomfort in building a trust relationship or working with fathers (Coakley et al. 2014). Similarly, studies on child welfare tend not to include fathers or exclude them from their analysis for various reasons (Lee et al. 2009) ranging from the perception of the father as uninvolved or dangerous (Bellamy 2009), and the perception of the mother as the main caretaker of the child, to social expectations regarding the role of fathers and fathering (O’Donnell et al. 2005), and the lack of direct data about fathers and reliance on information provided by mothers (Dubowitz 2006).

Studies that investigated the risk factors as well as the protective role played by fathers in child maltreatment cases (Dufour et al. 2008; Lee et al. 2008) show that fathers contribute directly and indirectly to the risk of maltreatment. The absence of the father from a household, indicated by the family’s single-parent status, has been associated with a higher risk of poverty, child neglect and physical abuse (Berger 2004; Sedlak and Broadhurst 1996).

The involvement of fathers in their children’s lives and in child welfare services has been shown to lower the likelihood of children being involved in the child protection system, reduce the length of time children remain in out-of-home care, increase children’s reunification with a parent or other relative, and protect against future child maltreatment (Bellamy 2009; Berger et al. 2009; Burrus et al. 2012; Coakley 2013; Guterman et al. 2009; Malm and Zielewski 2009; Proctor et al. 2011; Wingrove et al. 2016). Father involvement with the child is associated with lower maternal child physical abuse risk (Guterman et al. 2009) and with cognitive and social-emotional benefits to children at-risk (Dubowitz et al. 2001; Lee et al. 2009).

Less desirable findings were also reported in the literature. It was found that fathers are overrepresented as perpetrators of physical child abuse (Lee et al. 2009) and that co-occurrence of intimate partner violence and child abuse is more prevalent among fathers than mothers (Dixon et al. 2007). In addition, several risk factors of poor paternal practices have been noted. The father’s mental health problems can impact a parental relationship, successful fathering and child outcomes (Callender et al. 2012; Cummings et al. 2005; Kane and Garber 2004). Depression, in particular, has been closely associated with risk of child maltreatment by fathers (Cummings et al. 2005; Lee 2012; Lee et al. 2011, 2012; Sidebotham and Golding 2001). Living with a substance abusing father is associated with greater risk of physical abuse (Lee et al. 2011) and negative outcomes for children (Osborne and Berger 2009). Poverty is considered a risk factor for child abuse by fathers (Guterman and Lee 2005). Greater family stress due to economic challenges may lead to harsher and hostile parenting and greater risk to the child including a higher incidence of child abuse and neglect (Magnuson and Duncan 2002; Slack et al. 2004).

The role of fathers in child maltreatment cases in which the maltreatment was perpetrated by the mothers still remains to be explored. There is little research on fathers, in child welfare families, who are not violent, abusive or inadequate parents (Storhaug and Øien 2012). The key question is what role fathers play in families involved in maltreatment cases. It is unclear if non-maltreating/non-offending fathers play a protective role or pose a risk to the child. Generally, we know that a father’s positive impact on a child is dependent on the context and circumstance of his parenting (Doherty et al. 1998; Flouri and Buchanan 2002). It has been documented that families involved in child maltreatment cases usually suffer from a multitude of problems (Slack et al. 2004) and therefore the parenting context is likely to be complex and negatively impact on the father’s parenting. In contrast, a few studies have suggested that fathers can play a protective role in lowering maternal child abuse risk (Guterman et al. 2009) and reducing the time spent by a child in protective custody (Coakley 2013; Coakley et al. 2014). To the best of our knowledge, no study has, to date, investigated the role of non-maltreating fathers in maltreatment cases in which the mother was adjudicated for abuse and/or neglect of the child and her parental rights terminated. The present study was designed to augment the knowledge we possess on fathers involved in child maltreatment cases (Zanoni et al. 2013).

A review of the literature also demonstrates the need to explore the protective role played by a substitute father (also termed “social” or “surrogate father”), defined as a non-biological parental male figure involved with the child. The mother’s resident partner may constitute an important figure in the family life and contribute both risks and resources (Berger et al. 2009). In high-risk families, it is unclear if a resident non-biological partner is associated with lower or increased risk to the child. This study addresses the risk to children in these cases as well. It will help us to understand if the presence of a male figure in a household, regardless of blood ties, is associated with more/less risk and whether specific risks are associated with biological fathers as opposed to surrogate fathers.

It should be noted that fifty percent of families in Israel are nuclear families, that is, consist of two parents and at least one child under 17, while the average number of children per family is 3.7. The percentage of children under 17 who live with their parents is 91%, which is high compared to most OECD countries (Israeli Central Bureau of Statistics 2020). 12% of families with children under 17 are single families, mostly headed by a woman. 95% of couples in Israel are married. The characteristics of families involved with the child protection system mirror this picture: 52% of at-risk children live in a nuclear family with four or more children, and 19% in single families (Sabo-Lal 2017).

The present study

Drawing on a sample of maltreatment court cases in which the mother was reported for child abuse and/or neglect, the study had three objectives:

  1. (1)

    To describe differences in parent and child characteristics between families with resident biological fathers, non-resident biological fathers and resident non-biological fathers. The inclusion of subgroups of paternal figures enabled comparisons between the fathers’ roles and children’s outcomes across cases. Our hypothesis was that there would be differences between resident fathers, biological and non-biological, and non-resident biological fathers, especially with regard to the overall level of risk to the child as indicated, for example, by the number of child maltreatment reports, number of out-of-home placement, types of maltreatment, family poverty and the number of risk factors in a case.

  2. (2)

    To determine whether the association between risk to a child and a resident biological father, as a few studies indicated, persists even with controlled family and individual characteristics. Since abuse and neglect often occur in multi-problem homes and various family characteristics are associated with increased risk to the child, our hypothesis was that a relationship between a resident biological father and risk to a child persists even when controlled for individual and child risk factors.

The study was based on child maltreatment cases in Israel in which the mother was adjudicated for abuse and/or neglect of the child and her parental rights terminated. In order to put the study in context we should note that in Israel, in 2015, 43,971 children (i.e., comprising a rate of around 16 per 1000 children) were referred to designated child protection officers due to neglect (33.5%), physical abuse (26.1%), sexual abuse (11.0%), emotional abuse (6.5%), or other reasons (Gottfried and Ben-Arieh 2019; Israel National Council for the Child 2016). Public records indicate that 92.7% of investigated referrals were substantiated (INCC 2016).

In Israel, a termination of parental rights (TPR) petition is filed in a family court by the State when child protection workers assess that parents are incapable of meeting their child’s needs and when other solutions, such as foster care, have failed to ensure the child’s well-being. The TPR proceeding is based on the Children Adoption Act (1980) which lists several causes as grounds for TPR, the most common of which is parental incapacity, defined as a situation in which a parent is deemed incapable of taking due care of the child because of the parent's behavior or circumstance, and there is no prospect of change in either domain in the foreseeable future, even with reasonable aid from social services. Prior to TPR, state intervention in the family begins following a report of child maltreatment. Initial assessments are made by designated social workers from the Social Services Department. Substantiated cases are then brought before an interdisciplinary committee which assesses the risks, family preservation options and alternative care solutions, in particular out-of-home placements (Oppenheim-Weller et al. 2017). In cases where various previous placement options failed to secure a child’s well-being and placement stability, and the parents were deemed parentally incapable, the committee can recommend filing a TPR petition. It should be noted that, in Israel, in contrast to other countries, most foster care placements tend to be relatively stable. Children do not change foster families often and placements tend to be longer than in other countries. TPR is, thus, less common than out-of-home placements. The courts usually appoint mental health experts such as psychiatrists and psychologists to provide professional assessments of the child’s and parents’ mental health and parental capability. Social workers also submit reports which constitute a comprehensive psychosocial assessment of a family’s circumstances. The rich picture presented in the courts of families enabled the analysis of multiple factors relating to fathers and children’s risk in this study.

Method

Sample

The sample consisted of 237 court cases of child maltreatment in which the mother was adjudicated for abuse and/or neglect of the child and her parental rights terminated. These cases were retrieved by the first author from the Israeli Public Record of court cases and were originally identified through several relevant key terms such as child maltreatment, child abuse, child neglect and parental capacity. The cases were judged by the Supreme Court, District Courts and/or Family Courts and covered every geographic district in Israel. Decisions were selected from different regions in order to control for possible residence cluster. Cases included published and unpublished rulings issued by the courts in the period 2015–2018. Inclusion criteria were as follows: (1) the cases dealt with TPR; (2) the court ruled in favor of TPR; (3) TPR was based on parental incapacity due to abuse and/or neglect of the child; (4) the mother was reported as the perpetrator of the abuse and/or neglect. A total of 360 cases were originally identified but a number of cases were removed from the sample (n = 123) when one of the following conditions occurred: the father was reported as the main or co-perpetrator of the abuse and/or neglect of the child; the father was not alive at the time of the legal proceedings; the family status of the mother and father was unclear; the identity of the biological father was unknown.

The sample was divided into three types of non-offending fathers based on the father’s living arrangement at the time of TPR: (1) resident biological fathers, defined as cases where the biological father of the child resided with the biological mother in the same household; (2) single households, defined as cases where the biological father of the child did not live with the mother and child in the same household. Another requirement for this sub-group was that the mother did not reside with any other partner; (3) resident non-biological fathers defined as cases where the biological mother cohabited with a male partner who was not the biological father of the child.

Procedure

The sample of cases was comprehensively analyzed by means of a questionnaire specifically developed for the study, and was based on main coding categories that were found to be repetitive and key across different cases. It should be noted that categories relating to the child’s or parents’ mental health and cognitive condition, type of maltreatment, health problems and drug abuse (by parents) were based on professional assessments and reports provided by mental health experts to the courts during the TPR proceedings. Content analysis of the cases was undertaken according to the following categories:

Child characteristics

Child characteristics included the child’s age, mental health and/or cognitive impairment, history of out-of-home placements, type of maltreatment [the following categories were identified: physical (14.76%), emotional (9.2%), sexual abuse (2.5%) and/or physical (76.3%), emotional (86.1%), educational neglect (36.4%), medical neglect (22%), lack of parental supervision (8.9%), or abandonment of the child (19%)], health problems, number of children in the family and whether an emergency protection order to remove the child from the biological family due to immediate risk had been issued.

Parental characteristics

This category included those of the reported maltreatment perpetrator (biological mother, biological father or other), family status of the mother and father (married, single, divorced, separated, widow, not alive), whether the identity of the biological father was known, whether the mother lived with the biological father in the same household at the time of TPR (termed in the current study as “resident father” if affirmative or “single household” if negative), whether the mother lived with another partner, who was not the biological father of the child, in the same household at the time of TPR (termed in the current study as “resident non-biological father”); additional parameters that were categorized related to the age of the mother and father, their mental health and/or cognitive impairment, physical disability, drug abuse problem, criminal records (e.g., criminal conviction, incarceration), and family poverty (e.g., recipients of social security allowances, lack of financial means and/or financial deprivation, unemployment, employment in low wage jobs, work instability).

A composite variable of child risk factors was formulated, indicating the number of affirmative codings of the following variables representing risk to the child’s emotional and/or physical wellbeing: child’s mental health, cognitive impairment, number of out-of-home placements, medical condition, physical, sexual or emotional abuse, physical, emotional or educational neglect, lack of supervision and abandonment.

This composite variable was used as a dependent variable in the logistic regression and was dichotomized into two categories: the first category included cases with up to three risk factors per case (“low risk children”), while the second category included cases with four or more risk factors per case (“high risk children”). A cut-off score of 3 risk factors was chosen since this was the median value of risk factors in the cases examined, meaning that 51% of the cases included three or less risk factors relating to the child, and 49% of the cases included four or more such risk factors.

The rationale for constructing this variable was that, since the sample consisted of high risk families with multiple risk factors, we did not expect to find a “no risk factor” group and dividing the children into “no risk factor” or “risk factor” groups would not reflect the true condition of the children. However, based on the existing literature, differences between cases based on the father’s involvement in the household were expected. We thus decided to compare cases based on the number of risk factors in a case (rather than on whether there were or were not risk factors). Cases were split in the median, providing an accurate divide between a lower and higher number of risk factors cases. This gave us an accurate and more nuanced picture of the risks characterizing the children in the sample and enabled a comparison between the groups of children on a variety of variables.

Analysis

The three types of fathers in the study were compared using chi-square tests on categorical variables and one-way ANOVA tests on continuous variables. For a multivariate analysis, a series of logistic regressions were carried out. The aim was to predict cases with high or low risk children, that is, cases with four or more child risk factors or cases with less than four risk factors, based on the residence of the biological father in the family, and controlling for child and parent characteristics. Variables were entered in hierarchical fashion in order to track the connection between resident biological fathers and child risk factors, as other child and parent characteristics were added to the model. Variables which were significant were retained in the final model.

Findings

Differences on children, parents and characteristics based on three types of fathers

Table 1 presents significant differences between cases involving a resident biological father, single households or a resident non-biological father with respect to children and case characteristics.

Table 1 Prevalence of children and parent characteristics among three types of fathers
Full size table

An overview of the findings indicates that cases where the biological father resided with the mother had the highest rate of risk in respect to almost all indicators presented in Table 1, followed by single households while cases with resident non-biological fathers demonstrated the lowest rate of risk compared to the other two types.

Cases with resident biological fathers included more reports of child abuse, specifically physical and emotional abuse, as well as physical and medical neglect, compared to single households or cases with resident non-biological fathers. Poverty, mental health diagnoses for the child and the father, and paternal and maternal cognitive impairment were more prevalent in these cases. Emergency protection orders were issued at a higher rate. Only two variables were found at a higher degree in cases with resident non-biological fathers: criminal records and maternal drug abuse.

No differences between the three types of fathers were found in relation to the following variables: emotional neglect, sexual abuse, educational neglect, parent’s disability, father’s drug abuse, abandonment, lack of supervision, mother’s mental health and father’s criminal record.

In order to check differences between the three types of fathers in relation to continuous variables, a series of one-way ANOVA tests was conducted. A post-hoc Tukey analysis was used to identify differences between the three types of fathers. Table 2 presents the significant differences that were found.

Table 2 Mean, standard deviation, and significance level for differences regarding children and family characteristics between three types of fathers
Full size table

Households with resident biological fathers were associated overall with more maltreatment reports and a higher number of risk factors to the child compared to single households and households with resident non-biological fathers. Households with resident biological fathers included more children in the family and the child’s average age was almost 5 years old. The mothers in these cases were older than mothers in single households and households with resident non-biological fathers. No differences were found in the father’s age. Single households had a higher number of placements and the child's age at the time of TPR was older compared to households with resident biological and non-biological fathers.

Logistic regression for the prediction of child risk factors

Table 3 presents the results of the hierarchical logistic regression aimed at distinguishing between high-risk children (cases with four or more child risk factors) and low-risk children (cases with three or less child risk factors) based on the type of father in the household and case characteristics.

Table 3 Logistic regression model for high-risk children by residential biological father and other risk factors
Full size table

The variable of the resident biological father remained significant across the three models. In the first stage, when entered solely into the model, it increased by 4.17 times the likelihood of four or more child risk factors per case compared to single households. In the second stage, when variables relating to the father and child were entered, the presence of a resident biological father increased the likelihood of high-risk children by 3.25 times. Drug abuse by the father had a greater impact, increasing the likelihood of high-risk children by 3.73 times. The father’s mental health problems increased the likelihood of high-risk children by 2.71 times. The variables of the number of children in the family and the child’s age were also significant, increasing the likelihood of high-risk children by 1.51 and 1.19 times, respectively. In the third stage, when poverty and the mother’s mental health were entered as variables, the presence of a resident biological father increased the overall likelihood of high-risk children by 2.79 times. Family poverty was the most influential variable, increasing the likelihood of high-risk children by almost 18 times, followed by the father’s drug abuse and the mother’s mental health, which increased the chances by 4.56 and 3.99 times respectively.

Discussion

The study examined the association between non-offending resident biological fathers and children at risk, in cases of TPR. The cases consisted primarily of single-mother households and households with resident biological fathers. The main questions that were explored were whether the presence of a biological father in the home was associated with lower or greater risk to the child and whether this association held after controlling for family and individual characteristics.

Overall, it was found that the presence of a biological father in a household was associated with greater risk and negative outcomes to the child. The findings from the logistic regression indicated that resident biological fathers were associated with an increased risk to the child compared to cases where the mother lived alone or with another partner. Households with resident biological fathers had 2.79 times the risk of a child suffering from at least one or more of the following: mental health problems, cognitive problems, abuse, neglect, and higher out-of-home placements as single households. Biological fathers were connected with an elevated risk to the child even if they did not maltreat the child and after controlling for important factors such as paternal and maternal mental health problems, poverty, substance abuse, the child’s age and number of children in the family. While mental health problems, poverty and substance abuse significantly increased the risk to the child, the presence of a resident biological father made its own unique contribution. Two explanations can be offered in this regard.

The first is that co-occurring problems in a family can often compromise parenting practices, including those of fathers. If a family suffers from poverty, parental mental health issues and substance abuse problems, the cumulative negative stressful effect can make it hard on fathers to adequately parent. Poverty was specifically associated in the research literature with additional stress, mental health problems and greater risk of hostile or harsh parenting practices (Slack et al. 2004). Studies have indicated that economic strain can cause greater parenting stress and negatively impact on a father’s mental health and interactions with his children (Barnett 2008; Conger et al. 1993; Magnuson and Duncan 2002).

A second explanation is that fathers can put a child at risk through their negative impact on the mother’s capacity to parent (Lee et al. 2009). Evidence for this explanation emanates from studies showing that a negative relationship between parents increases the risk of maternal child abuse. For example, economic difficulties have been connected to an increased hostile co-parenting relationship (Barnett 2008; Brody et al. 1994; Bronte-Tinkew et al. 2010). It was also found that decreased social support by the child’s father to the mother was connected with a higher risk of maltreatment (Kotch et al. 1995). The findings of the present study indicate that cases with resident biological fathers have a higher rate of maltreatment, specifically physical and emotional abuse and physical and medical neglect by the mothers.

Both explanations indicate deficient empathy on the part of the fathers in the study. Previous studies have found that parents, who experience a high level of personal distress, often suffer from information-processing difficulties which makes perspective-taking and empathy towards the child more difficult (Kilpatrick and Hine 2005). Due to their inability to engage in perspective-taking, high-risk fathers tend to be physically aggressive towards their child and have decreased capacity to effectively parent. Biological fathers were less emotionally available to their children, and engaged in more detrimental parental practices that pose greater risk to the child (Kilpatrick and Hine 2005). It could also be that lack of empathy towards the mother compromised her ability to parent adequately. Lack of spousal support has been connected to risk of maltreatment (Kotch et al. 1995).

The findings of the present study have several implications. The study adds to the limited knowledge we have on fathers involved in child maltreatment cases, particularly cases that, due to their severity, reach a TPR outcome (Brewsaugh et al. 2018; Dubowitz 2006, 2009; Wingrove et al. 2016; Zanoni et al. 2013). The findings suggest that fathers should be part of a family intervention plan even when they are not identified as the main perpetrator of abuse or neglect, specifically when there is a high rate of risk factors in the family. They also refine our understanding of the risk factors concerning fathers, in particular, economic hardship and the father’s substance abuse. These characteristics can help to identify fathers who pose a risk to the child and formulate preventive plans.

The fact that an association was found between the presence of a resident biological father and increased risk to the child after controlling for other factors suggests that fathers can generate risk, beyond their specific problems as individuals, through the mere fact of being part of the family system, especially in serious cases with multiple risk factors in the family. The additional risk posed by resident biological fathers may reflect a more general familial stress or dysfunction. As documented in previous studies, child welfare families often face multiple problems (Slack et al. 2004) that can impact every member of the family.

There is a need to understand fathers, not just as individuals with problems, but also in the context of family life. While it is vital to address specific issues faced by families such as poverty, substance abuse and mental health, a systemic approach to the family can be beneficial. The family-as-a-system approach posits that, instead of focusing solely on the individual problems of a father or mother, the emphasis should be on the interaction between family members, their relationships and interconnections. The challenges faced by fathers can be viewed not only as reflecting inner, psychological pathologies but also as barriers to effective family interactions and relationships. An integrative approach that aims to ameliorate individual problems and family stress as well as the family system as a whole is thus recommended (Barnett 2008). More research is clearly needed in order to clarify the association between factors relating to the father as an individual and those relating to the family as a system in cases with varying degrees of risk, such as cases within the child protection system that do not involve out-of-home placement or cases that involve the latter but without TPR.

The conclusion from this analysis is that fathers, not only mothers, should be part of any intervention aimed at treating or preventing child maltreatment in severe cases where TPR is considered. Fathers tend to be overlooked in child welfare practice in general (not only in termination cases) to the detriment of the child and the family (Lee et al. 2009). Scholars have called for more father-inclusive practices and services in family treatment plans for child maltreatment cases (Bellamy 2009; Brewsaugh et al. 2018; Coakley et al. 2014; Dubowitz 2006; Wingrove et al. 2016; Zanoni et al. 2013). The findings of the current study strongly support this call, even though some studies suggest difficulties in outreaching fathers in order to engage them in a treatment plan (Duggan et al. 2004; Coakley et al. 2014). It is therefore important to learn how to effectively engage fathers in social services (Bellamy 2009; Lee et al. 2009).

Another barrier that has been noted in the literature and needs to be dealt with is the negative perception of fathers and attitudes of sexism on the part of child welfare workers, which may account for the lack of fathers’ involvement with child welfare services (Brewsaugh 2017). The present study did not address this issue and it is recommended that future studies examine the connection between caseworkers’ attitudes towards the participation of fathers in social services and case outcomes. The inclusion of fathers in services provided to the family is necessary and can be of benefit to both children and parents. A recent study, indicating the positive result of effectively involving fathers, showed that preventive intervention among low-income families which focused on father involvement and the couple relationship had a positive effect on the parents, the parent–child relationship and the child’s outcome (Pruett et al. 2017). Future studies are needed to determine whether such positive effects hold true for families reported for abuse and neglect and considered for TPR.

The findings of the current study also suggest that resident non-biological fathers, that is, the mother’s male live-in partner who is not the biological father, may offer some benefit to a family. Cases including this type of father demonstrated the lowest risk to the child in the key measures examined, the lowest rate of child abuse and neglect, less child mental health problems and family poverty, no cognitive impairment of the father and the lowest rate of paternal mental health problems. These fathers posing the lowest risk to the child compared to the other types of fathers may be due to a lack of engagement with the child, or, conversely, because they were able to offset the mother’s behavior by supporting the child.

It should be noted that these cases also had the highest rate of maternal drug abuse and criminal record. This suggests that surrogate fathers can have a mitigating effect on the child–mother relationship but, while risk to the child was low, it was nonetheless present. It is widely accepted that resident non-biological fathers pose maltreatment risk to the child, mainly because they do not have an emotional, normative commitment to the child’s welfare in the same way as biological fathers (Berger 2005). However, Berger et al. (2009) found that, when the mother was not the alleged perpetrator, the associations between the presence of social fathers and the involvement of child protection services were somewhat stronger. In the current study the main perpetrator was the mother and cases with resident non-biological fathers were found to pose the least risk to a child. It could be that, in cases where mothers are the main perpetrators, social fathers have mitigating effect and pose less risk to the child, compared to cases where they are the perpetrator of the abuse, and obviously present greater risk to the child. It could also be that, since the number of resident non-biological fathers in the study was small, the study was underpowered to detect more differences for this group than was found. Future studies should include a higher number of cases with resident non-biological fathers as a comparison group. Any effective family intervention plan should, therefore, take into consideration surrogate fathers, since the latter can be beneficial to the child but also pose a risk.

The study has several limitations. The study’s sample included high-risk children and was based on cases where TPR occurred. Due to the limitations of the data, it was impossible to examine the impact of other important factors such as the impact of the biological father’s positive involvement with the child (Guterman et al. 2009) and its association with a lower risk of maltreatment. No information could be gleaned from the court records regarding the father’s involvement with the child or the quality and richness of the father–child relationship. There is evidence that the father’s involvement and relationship with the child is more influential than his presence or absence from the house (Guterman and Lee 2005). Positive paternal involvement with the child has been associated with a lower risk of maternal physical abuse of the child (Guterman et al. 2009). Future studies may benefit from exploring the contribution of fathers to positive outcomes for children depending on the father’s level of responsibility for the child’s care and upbringing.

Similarly, the impact of the co-parenting relationship (Doherty et al. 1998), and male-relative involvement in the family (Bellamy 2009) were not examined. Nor did the study examine the services provided to the family. In order to inform intervention and service efforts for both children and families (Risley-Curtiss and Heffernan 2003), future studies should explore the connection between services provided to fathers and the levels of risk to the child (Bellamy 2009).

The study was conducted in Israel and therefore generalizability of the findings may be limited due to differences in the Israeli social-cultural context and child welfare system compared to other places. For example, the foster care system in Israel is relatively stable and provides children with long-term placement. In consequence, it is the most serious cases of child maltreatment that reach the TPR stage. Replication of the study to other social welfare and legal systems would be needed in order to discern social sensitivity or universality of the findings.

Since the study was retrospective and correlational, and examined the association between risks and various family statuses of fathers, this meant that issues of causation could not be examined. Future studies would benefit from a prospective design that follows families with different types of father involvement, and determines which cases were considered for TPR, as well as other child placement outcomes.

The current study focused on mothers as perpetrators of maltreatment and the risk or protective effect of different types of fathers. It is suggested that future studies should focus on fathers’ involvement in maltreatment and TPR. Another avenue of research would be to test the contribution of fathers to the overall risk of children depending on the father’s living arrangement at different points of time: at the time of the perpetration of the maltreatment, at the time of the child’s removal and at the time of TPR.

Conclusion

The study demonstrated an association between biological fathers residing in a household and increased risk to the child. The presence of a biological father in a household increased the risk of the child suffering from abuse and neglect, mental health problems as well as multiple transitions in the foster care system. In addition, specific problems relating to the biological fathers, such as mental health and drug abuse, increased the risk of poor parenting, compared to single-mother households. The study showed that, in contrast, surrogate fathers had some protective impact on the mother–child relationship but generated their own risk when associated with mental health problems, substance abuse or criminal record. It would be beneficial to explore further the contribution of fathers of different types to the child’s outcome and their impact on the child in order to be able to understand better the risks and contributions of fathers to families involved in maltreatment cases.