|Home | About | Journals | Submit | Contact Us | Français|
This study investigates how unemployment, traumatic sexual experiences, substance use, intimate partner violence, and parental involvement collectively contribute to involvement with child protective system (CPS) and court-restricted access to children among low-income, ethnically diverse fathers. Participants were 164 fathers involved in a statewide fatherhood program. The majority of the fathers in the program were unemployed (76%) and ethnic minorities (66%). Logistic regression revealed that traumatic sexual experiences and number of children were significant predictors of CPS involvement, whereas employment and traumatic sexual experience were associated with court-restricted access to their children. The results elucidate that clinicians and father-hood programs may need to attend to the history of traumatic experiences, as well as other contextual factors, of fathers and identify how, through trauma-focused interventions, to positively affect them and their children.
It has been observed that fathers disproportionally represent the majority of perpetrators of sexual and physical child abuse (Guterman & Lee, 2005), yet scant research has examined the role of fathers in children's involvement in maltreatment services (Behl, Conyngham, & May, 2003; Dubowitz, 2006; Francis & Wolfe, 2008; Haskett, Marziano, & Dover, 1996; Scott, Francis, Crooks, Paddon, & Wolfe, 2006). This gap in the literature raises critical questions about current strategies to support the healthy involvement of fathers in the lives of children, especially when they are economically disadvantaged. These issues are also important as fatherhood programs across the country struggle to assist underemployed and unemployed fathers who are economically disadvantaged in meeting the economic and emotional responsibilities of caring for their children.
As fatherhood programs continue to engage with economically challenged men, the programs face the difficulties of simultaneously attempting to increase the earnings and financial abilities of fathers, while fostering their emotional attachment to their children and ensuring that they are not increasing the risk to children and others. To address the gaps in the literature and inform fatherhood programs, we have attempted to investigate the factors that may confer risk and or serve as protective factors of child maltreatment among socioeconomically, and ethnically diverse fathers. The developmental ecological transactional model of child abuse is used to guide our research (Cicchetti & Toth, 1995; Francis & Wolfe, 2008; National Research Council, 1993). The developmental ecological transactional model of child maltreatment describes how abuse may be a function of multiple ecological systems: ontogenetic (individual), microsystems (family), exosystems (community and environment), and macrosystems (societal/culture). This theoretical model was selected because its framework aligns well with the factors included in the current study. In the current study, the ontogenetic (i.e., individual) factors are criminal history, substance use, and history of sexual trauma. The microsystem (i.e., family) indicators are violence toward partner and paternal involvement, whereas the exosystem (i.e., societal/cultural) factor is employment. The current study examined how these factors collectively affected child protective system (CPS) involvement and court-restricted access to children among an ethnically diverse group of fathers involved in a statewide fatherhood initiative. The current study also seeks to document differences in treatment and resource needs between fathers with criminal histories and those without.
Men in record numbers are returning to their families after incarceration (Day, Acock, Bahr, & Arditti, 2005). Travis and Wahl (2005) reported that more than 600,000 men are released from prison annually, and many will attempt to reconnect with their children and partners. These men, on returning from incarceration, face significant challenges and needs. These challenges include a lack of specialized job skills, higher substance use and abuse problems, violent histories, less effective communication skills, and lower educational attainment—all of which are factors that have been linked to maltreatment (Carlson & McLanahan, 2002; Council of State Governments, 2003; Murphy et al., 1991; Wolfner & Gelles, 1993). However, the direct relationship between criminal history and child maltreatment has not been fully explored. Some researchers have shown that fathers with criminal histories can positively affect their child's development (McKeown, Ferguson, & Rooney, 1998; Yogman, Kiindlon, & Earls, 1995). There is research that suggests that formerly incarcerated men want to be involved in the lives of their children and families and, because of this motivation, they are participating in parenting programs to facilitate this involvement (Mendez, 2000). Petersilia (2003) noted that individuals who have been successfully reinte-grated back into their family have lower recidivism rates compared to the individuals who are disconnected from their families. Evidence suggests that paternal involvement postcriminal justice involvement has positive effects for the parent and child (Day, Acock, Bahr, & Arditti, 2005). We were interested in understanding if there is a cumulative effect for criminal justice involvement (more types of offenses charged with) and CPS involvement risk or court-restricted access. In the current study we hypothesized that once other risk factors (i.e., substance use, violence toward romantic partner, etc.) that co-occur with criminal history are accounted for, the main effects of criminal history will be negligible.
Several studies have documented that fathers' substance use increases the risk for physical abuse and neglect (Ammerman, Kolko, Kirisci, Blackson, & Dawes, 1999; Terling, 1999). In a national sample of 6,002 households, Wolfner and Gelles (1993) observed that drug users reported 46% more severe acts of violence and 20% more minor acts of violence when compared to nonusers. The increased risks observed by Wolfner and Gelles were also documented in a sample of 206 serious child maltreatment cases examined by Murphy and colleagues (1991). Murphy et al. reported that in 43% of the cases of child maltreatment, one or more of the parents had a documented substance abuse problem. Moreover, parents who had a documented substance abuse problem were more likely to be previously referred to child protective agencies than non-substance-abusing parents. We hypothesize that the more drugs fathers indicated that they use, the greater the likelihood of their reporting CPS involvement with their children and court-restricted access.
Personal sexual victimization, specifically the experience of child sexual abuse, has also been described as an enduring or long-lasting vulnerability risk factor for child maltreatment (Cicchetti & Toth, 1995). This observation raises questions about how traumatic experiences may negatively affect a father's ability to parent and how this increases the intergenerational transmission of risk (Simons, Wurtele, & Heil, 2002). There is a body of research (Beitchman, Zucker, Hood, DaCosta, Akman, & Cassavia, 1992; Browne & Finkelhor, 1986; Finkelhor, 1990; Putnam, 2003) with women and some men that documents the negative effects of child sexual abuse on the intergenerational risk to children. However, few studies have examined this for men involved in fatherhood programs. We hypothesize that men who report childhood sexual trauma will report increased CPS involvement with their children and court-restricted access to their children.
One microsystem risk factor for child maltreatment is intimate partner violence (IPV) or male violence against an intimate female partner (Herron & Holtzworth-Munro, 2002). The literature suggests that child maltreatment and IPV often co-occur (Herron & Holtzworth-Munroe, 2002). Ross's (1996) examination of the data from the National Family Violence Survey found that martial violence was a robust predictor of physical child abuse. Ross's study documented how men who engaged in more aggressive acts toward their spouses were also more violent and abusive with their children. Furthermore, it has been estimated that children are 15 times more likely to experience neglect and abuse when IPV is experienced (Osofsky, 1998). There is additional evidence linking growing up in a home with IPV and the future perpetration of sexual assault (Raj et al., 2006). This association demonstrates the deleterious effects that IPV can have on the developmental trajectories of the children involved in these homes (Choate, 2003; Hong, 2000; Loh, 2003). We hypothesize that men who report using IPV in their relationship will be more likely to have CPS involvement with their children and court-restricted access to their children.
Positive parental involvement has been identified as a protective microsystem factor against child maltreatment (Gorvine, 2003). Positive parental involvement includes behaviors like effectively communicating with children, reinforcing positive behavior, and validating the child's accomplishments (M. Sanders, 1999). These behaviors, when fully integrated into a consistent pattern of interactions between a father and a child, have been shown to support their healthy development. Positive parental involvement and communication are of particular interest given that they are modifiable behaviors that can be integrated into intervention programs. As such, some studies have suggested that more positive communication and involvement with one's child decreases the risk of maltreatment (Harrison, 1997; M. R. Sanders, Cann, & Markie-Dadds, 2003). We hypothesize that men who report positive parenting behaviors will be less likely to have CPS involvement with their children and court-restricted access to their children.
Understanding macrosystem risk factors posed to children's safety by their economically challenged fathers is an important first step for all fatherhood programs. Over the last decade, the discussions around the economic demands of fatherhood have begun to include how the financial responsibilities assigned to fatherhood are inextricably tied to the emotional responsibilities of the role (Lamb, 2001; Palkovitz, 2002). In other words, men who are more emotionally attached to their children are more likely to provide for them financially. If the economic demands and emotional responsibilities are linked, it raises questions about the impact a father's involvement has on his children and how his employment history introduces risk. Unemployed and low-income men enrolled in fatherhood programs tend to require more assistance with basic skills to assume fully the demands of the father's role (Jarrett, Roy, & Burton, 2002). This difference is important because it underscores a challenge not often observed in middle-class fathers who may choose to disengage with their children not because of lack of resources, but rather as a reaction to the dissolution of their marital and or intimate relationship with the mother of their children (Feldman, Nash, & Aschenbrenner, 1983). Succinctly, this discussion highlights the “dead-beat” versus “dead-broke” exchanges in the literature that seeks to characterize the differences in ability to pay child support between low-, middle-, and high-income fathers (Maldonado, 2006). Understanding how employment status introduces risk that uniquely contributes to maltreatment is important as programs seek to develop and implement effective strategies that work with fathers. This is especially important given the voluntary involvement of these men in services that seeks to increase their parenting skills. We hypothesize that the economic status, as measured by employment status, is not be related to the CPS involvement of their children or their court-restricted access to their children.
The literature reviewed elucidates how ecological factors such as ontogenetic (criminal history, substance use, and past traumatic sexual experience), microsystem (violence toward partner and parental involvement), and macrosystem (unemployment) factors together may be associated with child maltreatment. However, no study to date has examined the collective effects of these identified protective and risk factors on child maltreatment as measured by involvement in child protective services and court-restricted access to children for fathers involved in fatherhood programs. The current study begins to do so using data collected from fathers involved in a statewide fatherhood initiative program in Connecticut. The fathers included in the study were deemed low-income because of the communities they came from and the disproportionate number of them that were unemployed. It is hypothesized that individuals who have higher substance use, had a traumatic sexual experience, and report engaging in IPV will be positively associated to CPS involvement and court-restricted access to their children. Although individuals who report higher positive paternal parental involvement will have lower probability of CPS involvement and court-restricted access to their children. Finally, we hypothesize that the history of offending and low-income status will not be predictive of the outcomes.
We also explored the differences in treatment and resource needs between individuals who have a criminal history and those who do not. Petersilia (2003) reported the men with criminal histories often have unmet mental and physical health treatment and job placement (resource) needs when they reenter the community. We postulate that individuals who have committed more criminal offenses might have fewer resources available. We hypothesize that fathers with more history of criminal offending will express greater parent training, job/education training, and treatment (i.e., substance use treatment and health services) needs than men who do not have criminal histories.
A power analysis using G*Power (Erdfelder, Faul, & Buchner, 1996) was conducted prior to starting the study. The power was set at f2 = .85 and the α at .05. Based on the analysis, to detect a medium effect size (Cohen, 1988) a sample size of 120 was needed, F(8, 111) = 2.02, λ = 18. Therefore, according to the analysis the current study is capable of detecting a medium and large effect.
The principal investigator is the evaluator of an ongoing U.S. Department of Health and Human Services–funded, Connecticut Department of Social Services–administered fatherhood initiative aimed at promoting responsible parenting, economic stability and, where indicated, healthy relationship skills for low-income, noncustodial fathers. The major factors of relevance to the current study were drawn from the larger evaluation that assessed participants on entry into the program. Institutional Review Board (IRB) approval was obtained before initiation of the evaluation. To further protect the information collected from the participants, a certificate of confidentiality was obtained from the U.S. Department of Health and Human Services. The assessment evaluated program participants across demographic, relationship satisfaction and conflict, parenting, mental health, physical health, substance use, employment status, economic stability, criminal involvement, and resource needs. Items from the larger evaluation were included in the current study if they measured the risk and protective factors for involvement in CPS and court-restricted access identified in the current study. The variables in the current study were selected from a larger battery of measures used in the outcome evaluation of this fatherhood initiative. Measures selected for inclusion have demonstrated some effects on CPS involvement in the past with mothers and some father groups. To date, few studies have examined these factors in a sample of low-income, noncustodial fathers.
Data presented in the current study come from the first wave of participants who have completed the assessment and enrolled in the ongoing program evaluation of a statewide fatherhood initiative. The statewide fatherhood program's goal is to help men become positive and healthy role models by increasing their attachment to their children and families. Activities of interest in this evaluation are case management and outreach services, parent training, activities that support the economic health of men, and, when requested by the men and their partners, healthy relationship skills training. Participants were generally referred to the fatherhood programs through a variety of means including self-referral, the Connecticut Department of Social Services area offices, community programs, support groups, and the Department of Children and Families area offices. All of the men referred to the program were eligible for inclusion. Participants were informed about the voluntary nature of participation in the intervention and the evaluation at the point of contact/referral, as well as during the intake process. Specifically, in these procedures, the voluntary nature of their participation in the evaluation was acknowledged, and their option to withdraw at anytime was explained. During the consent procedures they were informed that their responses to questions would remain anonymous and would not negatively affect their participation in the program offered. All enrolled fathers in the fatherhood initiative program also enrolled in the program evaluation.
The demographic information gathered included age, number of children, ethnicity, employment status, and education level.
Criminal history was measured on a 4-point ordinal categorical scale. Individuals were asked to self-report if they were ever convicted of a crime or multiple independent crimes including felony, misdemeanor, and/or violent offense. For example, if the individual indicated that he had never been convicted of a crime, he was given a “0,” while an individual who indicated that he had committed a felony and a violent offense was given a score of “2,” and so on, up to 3.
Substance use was assessed using three items, which asked the participant whether he used (1) alcoholic beverages, (2) marijuana, and (3) other drugs (i.e., cocaine, heroin, or anything else) in the past 30 days. The participant responded by answering yes or no, and the substance use score was obtained by summing up the three items. We used the cumulative score of the various substances use items, because we believed that would be a better indicator of general substance use, as opposed to examining each item separately.
Sexually traumatic experience was assessed using a dichotomous item, “At any time as a child, have you ever been involved in a sexual experience that was traumatic for you?” This question was asked in this manner to capture the participant's subjective experience and retrospective assessment of the negative experience (B. Sanders & Becker-Lausen, 1995)
The IPV scale is a 4-item measure that assesses verbal and physical violence inflicted on female partners. Sample items include, “Have you pushed, slapped, grabbed, punched, kicked, beat up, burned or choked your partner?” and “Have you engaged in any of the following behaviors: insult your partner or put them down, swear at them or threaten them?” These items were based on items from the Conflict Tactics Scale–Revised (CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996). However, because many of the participants (first 50) and case managers found the items confusing and redundant, the principal investigator merged the items from the CTS2 together. For example, two separate items on the CTS2, “have you grabbed …” and “have you pushed or shoved your partner” were combined into one item. The authors realized that combining the items alters the psychometric properties of the item; however, given the strong program and participant feedback, we made the changes accordingly. The reliability estimate of this measure was α = .66.
The parental involvement scale was derived from the Alabama Parenting Questionnaire (APQ; Shelton, Frick, & Wooton, 1996). For the current study we created an empirically derived scale from the APQ, which we termed Parental Involvement. We first conducted a principle component factor analysis using a varimax rotation with three of the APQ scales: Involvement, Positive Parenting, and Inconsistent Parenting (Sheldon et. al., 1996). The results revealed a three-factor solution, which accounted for 46%, 13%, and 8.4%, respectively, of the variance respectfully. Five items in the second factor fit the construct of interest, Parental Involvement, well. These items included, “you volunteer to help with special activities,” “you help your child with his/her homework,” “you talk to your child about his/her friends,” and “you attend PTA meetings, parent/teacher conferences at your child's school.” The internal consistency estimate for this scale was high (α = .88).
This item asked the participant whether he felt that his father was involved in his life when he was growing up (yes = 1, no = 0). The item was included in the analysis as a covariate.
One item assessed the participants' current employment status (yes = 1, or no = 0).
Because the researchers were interested in elucidating the predictors of involvement with child protection services, one item from the assessment battery “Reason for Referral: Department of Children and Families (DCF) involvement” was selected as the primary outcome measure. DCF is the CPS agency for the state where the study was conducted. Referral from DCF signifies the agency's concern about the parenting skills of the target parent. Without this risk, no referral would be made. This is a dichotomous one item measure (1 = yes, DCF involvement [CPS], 0 = no). This strategy has been used by others (Berger, Paxson, & Waldfogel, 2009). In the current study they argued that CPS was unlikely to contact a “family”—in this case a father, regarding a screened-out child maltreatment report.
Another area of interest to the current study was whether the father's access to his child was restricted. Although this variable could be related to CPS involvement, we wanted to examine if the predictors were different from CPS involvement. To measure this outcome, respondents endorsed the prompt “Has the court placed any restrictions on your contact with this child?” A score of “1” indicated that there was court-restricted access or “0” if there were no court restrictions to access their child.
The treatment and resource needs assessment is an 11-item self-report measure that assesses three areas of resource: need for educational training (4 items), parental training (4 items), and treatment (i.e., health, mental health, and substance use; 3 items). An example of a sample item is “Do you think you might want help with any of the following…”, which was followed by the stem in the treatment and resource areas (e.g., finding a job, parenting skills, health services). Participant could respond either yes or no. The internal consistency estimate for the entire needs assessment measure was α. = 70.
One hundred and sixty-four fathers enrolled during the 1st year of the program, and complete assessment data were included in the study. The age of the participants ranged from 17 to 59 years, and the average age was 33.75 (SD = 7.85). On average, the men reported having 2.14 (SD = 1.53) children; the range of number of children was 1 to 10. The racial/ethnic proportion was 43% (n = 71) African Americans, 31% (n = 52) European/White, 23% Latino/Hispanic (n = 38), and 3% (n = 5) did not report their race. With respect to employment status, 76% (n = 124) of the fathers were unemployed, 24% (n = 40) were employed. Fifty-eight percent (57.8%, n = 95) of participants had a valid driver's license. Nearly one half of the participants (48.2%, n = 79) had been incarcerated, imprisoned, or jailed for a non-child-support-related offense. The participants were asked to indicate if there were ever convicted of a misdemeanor, felony, and or violent crime. They could have multiple arrests in each category. Eighteen percent (18%, n = 30) of the participants in the sample reported having been convicted of at least three types of offenses (i.e., misdemeanor, felony, violent crime), 24% (n = 40) were convicted of two types of offenses, 33% (n =55) were convicted of one type of offense, and 25% (n = 41) have never been convicted of any type of offense. In regard to their level of education completed, 32% (n = 53) fathers had some high school education, 2% (n =3) had less than an eighth-grade education, and 64% (n = 106) had a high school diploma or General Equivalency Diploma equivalent. One hundred and forty-six (88%) of the men indicated that they did not reside with their child(ren). Refer to Table 1 for details of the demographics.
To explore the relationship between the variables in the current study we conducted correlation analyses. To examine the predictors of CPS involvement and court-restricted access to child, we performed a logistic regression. Logistic regression was selected because this analysis handles discrete outcomes (yes =1, no = 0), and the impact of the predictor is explained by the odds ratios (ORs) or the odds of the event occurring or not occurring (Garson, 2008). Finally, analysis of variance (ANOVA) was conducted to examine mean differences in resource needs between individuals who had zero to three criminal offenses (Table 2).
Pearson correlations were performed to examine the associations between the variables in the study (see Table 3). Sexually traumatic experiences were associated with IPV (r = .18, p < 0.05), CPS involvement (r = .24, p < 0.05), and court-restricted access to child (r = .28, p < 0.01), whereas number of children was related to CPS involvement (r = −.22, p < 0.05). The strongest magnitude between variables was between the two outcomes of CPS involvement and court-restricted access (r = .43, p < 0.01). Finally, employment status was related to court-restricted access to child (r = .25, p < 0.01), criminal history (r = −.25, p < 0.01), DCF involvement (r = .28, p < 0.01), father involvement (r = .35, p < 0.01), and parental communication (r = .26, p < 0.01).
A logistic regression was performed to assess the relationship between criminal history, substance use, traumatic sexual experience, IPV, father involvement, parental involvement, number of children, employment status, with the outcome CPS involvement (0 = no involvement, 1 CPS involvement). The results for both logistic regressions are presented in Table 4.
The logistic regression result assessing the overall model was significant, χ2 = 29.39, p < 0.001, and the Cox and Snell R2 indicate that 18% of the variance was accounted for by the predictors of CPS involvement. The overall model was able to classify 87% of the cases correctly. Specifically, the model was able to predict 96.4% of individuals who were not involved in CPS and 21.1% of CPS-involved participants. The significant predictors in the model were traumatic sexual experience and number of children. Individuals who reported having a traumatic sexual experience were 10 times more likely to be involved with CPS (OR= 9.97, p < 0.01), whereas individuals who had fewer children, reported greater CPS involvement (OR .31, p < 0.05).
A second logistic regression was performed examining the amount of variance observed in court-restricted access to children by the predictor variables. The analysis revealed that the model was significant χ2 16.72, p < 0.01, and the Cox and Snell R2 indicates that the variables in the model accounted for 11% of the variance of court-restricted access to children. The model was able to correctly predict 96% of the individuals who had no court restrictions on access to their child(ren), and 28% of the individuals who had court-restricted access, for an overall classification rate of 87%. In this model, traumatic sexual experience and employment status were predictors of court-restricted access to child. Participants who indicated that they have had a traumatic sexual experience were nearly 6 times more likely to have court-restricted access to their children than those individuals without traumatic sexual experiences. Individuals who were employed were 3.67 times more likely to have court-restricted access.
Of the four parental resource need items, the fathers on average reported that they needed to work on approximately two of these items (M = 1.95, SD = 1.42). With respect to the four educational resource needs, the fathers on average reported that they needed to work on two of these items (M = 2.11, SD = 1.32). When asked about the three treatment needs, the fathers on average reported that they needed to work on fewer than one of these items (M = .4, SD = .72). From these observations, we explored group differences in expressed resource needs.
Our last research question sought to examine differences in resource needs (i.e., parental training, educational training, and health and mental health needs) by their history of criminal offending. The 11 resource needs were categorized across three areas: treatment needs, parental training, and educational training. Three ANOVAs were performed to assess differences in resource needs depending on history of criminal offenses. The first ANOVA revealed that there were differences between groups in resource needs, F(3, 162) = 7.16, p < 0.001. The men who were convicted for more than two types of category of crime reported greater resource needs compared to the men who had no convictions. The overall ANOVA examining educational needs by their history of criminal offending was significant, F(3, 162) = 2.82, p < 0.05. Post hoc analysis revealed no significant differences between the groups but near-significant difference between the group with no history of criminal offending (M = 1.71) and the group convicted of one type of criminal offending (M = 2.36), p < 0.07). The third ANOVA model that examined the differences in parenting needs by history of criminal offending was non-significant, F(3, 162) = 1.35, p < 0.26 (see Table 1).
This investigation sought to investigate how unemployment, traumatic sexual experiences, substance use, IPV, and paternal parental involvement collectively contributed to involvement with CPS and court-restricted access to children among low-income, ethnically diverse fathers.
Traumatic sexual experience and number of children were significant predictors of CPS involvement. Traumatic sexual experience and employment status were associated with court-restricted access to their children. Consistent with theory and previous research (Simons et al., 2002; Widom & Ames, 1994), traumatic sexual experience was a robust predictor of CPS involvement and court-restricted child access. This observation is consistent with other research that shows that traumatic sexual victimization is related to CPS involvement and other negative consequences for the adult victims and their children (Cicchetti, & Toth, 1995; Francis & Wolfe, 2008; Wolfner & Gelles, 1993). This raises questions about the role of trauma on the ability of men and fathers to protect their children. Further, might focused interventions and initiatives in this area help mitigate the negative effects not only for the men and fathers but also their children and other family members. From a stress-coping framework, Weiler and Widom (1996) discussed how specific coping and impulsive behavior styles may develop in response to the victimization experience. They indicate that behaviors that are displayed by the victimizer such as lying, abusive behavior, manipulation, and disconnection with feelings are sometimes internalized and adopted by the victim. Although some of these strategies might be adaptive during the victimization experience, they become maladaptive patterns of coping when the individual becomes an adult and parent. Observations from this study call for fatherhood programs to explore with men how their past sexually traumatic experience impacts their ability to parent and coparent. Given the links between past sexual trauma and risks for child maltreatment, programs may consider engaging men and fathers in proven psychoeducational trainings to help them mitigate the effects of their past trauma.
The findings additionally highlighted that fathers who indicated that they had fewer children were more likely to be at increased risk for CPS involvement. This observation contradicts what we hypothesized. It may be that as men have more children they learn positive techniques to engage in healthy ways with their children or they have less contact with these children because of their limited resources to meet the children's needs, all resulting in less risk. However, this explanation is speculative and thus warrants further investigation. These investigations should focus on the nature, quality, and experience of the fathers and their children to be better able to describe the differences present and their relationship to risk and protection.
Unexpectedly, it appeared that men who were employed were more at risk of having court-restricted access to their child. Reasons for this finding may be related to these employed fathers adhering to traditional gender roles and struggling to maintain them (O'Neil, Helms, Gable, David, & Wrightsman, 1986). Alternatively, because of these fathers' employed status, they may feel entitled to unfettered access to their children and as the relationship either ends between them and the mother they may have more negative interactions that undermine their access with them the judicial system. For these men, balancing family life, work, and the social expectations of their role as fathers might tax their resources, thus increasing the probability of engaging in maladaptive behaviors that results in court-restricted access to their child (Brunswick, Lewis, & Messeri, 1992). Because we did not measure gender-role conflict, these hypotheses should be investigated further. Understanding the unique coparenting patterns between employed and unemployed, low-income fathers would begin to elucidate this observation.
The findings suggested that criminal history was not related to CPS involvement or court-restricted access to a child. These results help to support our hypothesis that having a criminal history does not automatically increase the risk of maltreatment. As Mendez (2000) in his study illustrated, men who were incarcerated wanted to foster a relationship with their children and were willing to seek training to do so. Hence, it appears that other factors, for example, past sexual traumatic experience, or as other studies have indicated, substance abuse (Ammerman et al., 1999; Terling, 1999), may increase the risk posed by men with criminal histories. These findings warrant further examination into the factors that facilitate men with criminal histories maltreating children. Additionally, research needs to examine and isolate what are the implications of these findings for CPS agencies.
IPV was not associated with CPS involvement or court-restricted access. These findings are in contrast to previous studies that suggested that as IPV increases so does physical child abuse (Herron & Holtzworth-Munroe, 2002; Ross, 1996). Given that these fathers do not reside with their children, the relationship between IPV and maltreatment may be reduced because of the distance. The results from the current study that suggest no relationship between self-report of IPV and CPS involvement of their children and court-restricted access to their children may also be an artifact of the observation that men have been shown to underreport their use of violence in their intimate relationships when compared to their female counterparts (Armstrong, Wernke, Medina, & Schafer, 2002; Caetano, Schafer, Field, & Nelson, 2002).
Substance use and parental involvement (communication) were not associated with the outcomes. The lack of relationship between substance use and CPS involvement is surprising given that previous research has showed a strong link. Because these are noncustodial fathers, this lack of effect might be reflective of the distal relationships these men may have with their children. In addition, the men might be hesitant to disclose substance use issues on entry into a program due to the legal and programmatic ramifications. Underreporting of substance use has received anecdotal support as case workers in their verbal communications with the evaluator have commented on how these men's disclosures are more frequent as the case management relationship develops. Therefore, although we did not detect an association between substance use and the outcomes, it is recommended that agencies such as fatherhood programs screen for substance use and abuse and evaluate their impact on parenting capacities (Ammerman et al., 1999; Terling, 1999; Wolfner & Gelles, 1993).
Parental involvement and communication was not found to be a protective factor or predictor of CPS involvement and court-restricted access to their child. This lack of relationship could be due to socially desirable responding. Generally, the men reported that they engaged in a relatively high number of prosocial parenting activities. More research is needed to investigate the role of paternal parental involvement, its definition for this population, accurate assessment of involvement by men, and how increasing their abilities decreases the probability of maltreatment.
Lastly, the current study explored the differences in needs between men with and without criminal histories. Fathers with histories of criminal offending had significantly more treatment (i.e., substance use, psychological and health services) and educational needs. These findings are consistent with the prisoner reentry literature, which shows that men with a history of incarceration have more concerns and needs (Council of State Governments, 2003). Fatherhood programs should consider developing strategies that on entry identify men who have an extensive history of criminal offending and emphasize how building additional supports and skills helps them to achieve their goal of being available and healthy fathers to their children. They should also develop clear strategies and language that moves men from a place where criminal histories are not negatively associated with risk but an avenue for capacity building and strength. The results highlight potential interventions for individuals entering fatherhood programs with criminal justice histories (Council of State Governments, 2003).
Although the findings are promising, there are several limitations that are worth noting. Because of the cross-sectional and retrospective design, no causal relationships can be drawn from these findings. Given that this survey was self-report, accuracy and social desirability may have affected the responses given. The relatively small and convenient sample size limits the generalizability of our findings. There were also a number of alternative hypotheses that would have enhanced our understanding of the phenomenon of interest if measured. Future studies should note these challenges and expand the body of knowledge available. The outcome measures used in this investigation did not directly observe the phenomenon of child abuse and neglect. More robust measures of this phenomenon are needed as this discussion and line-of-inquiry continues. Finally, there are many other factors that predict child maltreatment that were not accounted for in the current model. Developing strategies to integrate these factors into a more definitive study is indicated.
Despite the limitations, the findings highlight how fatherhood programs can play a significant role in reducing the risks to fathers and their children. Understanding how a father's factors at all ecological levels operate for men often described as being on the “fringe” will greatly enhance our ability to intervene and prevent problems for them and their children. Further, they can help to support healthy family and community involvement (Gordon et al., in press; Scott et al., 2006). Our findings lend additional support for Alaggia and Millington's (2008) recommendation that sexual victimization be a part of a regular assessment used by clinicians and caseworkers who serve these men. Although some men might not be forthcoming in disclosing this information, for those who do, appropriately targeted interventions may help to reduce negative outcomes for them and their children. It may be in the best interest of fatherhood programs and/or clinicians to help men process their traumatic experiences, and increase their active coping strategies. These results also elucidate the need for proven strategies to treat male victims of sexual abuse. Finally, the findings suggest that programs need to acknowledge that the histories of these fathers relate to their abilities to engage fully with their children in healthy ways. By addressing these concerns and raising a father's awareness of his potential risks to his children, programs and fathers could model and develop skills sets that increase the positive parenting behavior they use.
There is evidence that men want to be assets to their families and children. Documenting how to help meet this need while protecting fathers and those they care for, would help to keep all involved safe (Gordon et al., in press). Given that multiple contexts where men and boys receive messages about what constitutes the masculine ideal, future studies should also explore how specific gender-identity mechanisms such as conformity to masculine norms (Liu & Iwamoto, 2007; Mahalik et al., 2003) and gender-role conflict (O'Neil et al., 1986) may contribute to understanding how to reduce child maltreatment.
Funding for the Promoting Responsible Fatherhood program was provided by the U.S. Department of Health and Human Services, Administration of Children and Families Grant #90FR0031 administered by the Connecticut Department of Social Services. Support for Dr. Iwamoto's time on this study was supported by the National Institute on Drug Abuse Grants # T32-DA-019426-0, and 3R01-DA018730-04A1S1.