FATHERS, NEGLECT, AND CPS INVOLVEMENT 1 Paternal and Household Characteristics Associated With Child Neglect and Child Protective Services Involvement Shawna J. Lee University of Michigan, Ann Arbor Recommended Citation: Lee, S.J. (2013). Paternal and household characteristics associated with child neglect and Child Protective Services involvement. Journal of Social Service Research, 39 (2), 171-187. DOI: 10.1080/01488376.2012.744618 Author Note: Shawna J. Lee, School of Social Work and Research Center for Group Dynamics, University of Michigan, Ann Arbor, Michigan. Please address correspondence concerning this article to: Shawna J. Lee, University of Michigan, School of Social Work, 1080 South University Avenue, Ann Arbor MI 48109. Phone: 734-763-6565. Fax: 734-763-3372. Email: shawnal@umich.edu. FATHERS, NEGLECT, AND CPS INVOLVEMENT 2 Abstract This study examined the association of paternal and household characteristics with householdlevel measures of child neglect and Child Protective Services (CPS) involvement, measured when the index child was 5 years of age. Secondary analyses of the Fragile Families and Child Wellbeing Study were conducted using a subsample of 1,089 residential, biological fathers. Logistic regression models indicated that paternal depression was associated with greater than doubled odds of child neglect and CPS involvement. Paternal alcohol use and parenting stress were associated with approximately 50% increased odds of child neglect, and a scale measuring 13 caregiving tasks to reflect positive father involvement with the child was also associated with less risk for child neglect. However, paternal alcohol use, parenting stress, and positive involvement with the child were not associated with CPS involvement. An implication of this study is that paternal psychosocial functioning are important to consider in conjunction with sociodemographic factors when examining maltreatment risk in two-parent families. KEY WORDS: fragile families, child maltreatment, abuse, child welfare, depression, father, parenting, Parent-Child Conflict Tactics Scales FATHERS, NEGLECT, AND CPS INVOLVEMENT 3 Paternal and Household Characteristics Associated With Child Neglect and Child Protective Services Involvement The developmental ecological framework, often used to examine the etiology and consequences of child maltreatment, proposes that risk for child maltreatment is influenced by individual-level parental characteristics, family-level factors, and the broader context (Belsky, 1993). Numerous studies, reviewed below, have examined maternal risk for Child Protective Services (CPS) involvement, linking factors such as poverty, father absence, and maternal psychosocial risks to incidence of neglect and CPS involvement. Nevertheless, little attention has been directed to individual-level parental characteristics of fathers that may be associated with the risk for child neglect or CPS involvement among families of young children. This gap in the research literature is problematic for many reasons. The incidence of child maltreatment perpetrated by fathers or father surrogates is higher than expected given that fathers spend far less time than mothers caring for young children (Yeung, Sandberg, Davis-Kean, & Hofferth, 2011). National child welfare data have indicated that fathers or father surrogates were implicated in more than half of all of child maltreatment fatalities in which a parent was the perpetrator (U.S. Department of Health and Human Services [USDHHS], 2011). However, other studies have suggested that state child-protection data have likely underestimated actual father involvement in child maltreatment fatalities. Data from the National Violent Injury Statistics Systems indicated that 83% of beating and shaking injuries that resulted in child fatality involved father or father surrogates as perpetrators (Fujiwara, Barber, Schaechter, & Hemenway, 2009). Similarly, Brewster et al.(1998) found that 84% of infanticides involved father or father surrogates as perpetrators. FATHERS, NEGLECT, AND CPS INVOLVEMENT 4 The overrepresentation of fathers as perpetrators of maltreatment underscores the importance of identifying markers of maltreatment risk in father-involved families of young children. Failure to examine fathers’ unique and complex roles in families might hinder identification of specific mechanisms at multiple levels that can be targeted to reduce or prevent neglect in particular and child maltreatment more broadly. The current study sought to address this gap in the literature by examining paternal and household characteristics associated with household-level measures of child neglect (i.e., whether either parent was unable to meet the child’s basic needs or whether the child was inadequately supervised; Coohey, 1998, 2003; Dubowitz et al., 2005; English, Thompson, Graham, & Briggs, 2005; Theodore, Runyan, & Chang, 2007) and CPS involvement (Berger, Paxson, & Waldfogel, 2009) in a community-based sample of father-involved families with young children. Maltreatment Risk at the Community, Family, and Individual Levels The developmental ecological theory emphasizes that child maltreatment is determined by multiple factors, involving an interaction of these factors at the community, family, and individual levels (Belsky, 1993). Belsky’s (1993) model offers a distinct advantage in that it acknowledges not only the potentiating factors that increase risk but also the compensatory factors that minimize risk. Numerous studies, most of which have focused on mothers, have examined relevant family and individual factors. Family-level factors that are related to maltreatment include household economic hardship and unemployment (Berger, 2004; Slack et al., 2004), with single-mother headed households facing high risk for maltreatment (Paxson & Waldfogel, 1999; Theodore et al., 2007; Wu et al., 2004). A wide range of demographic and psychosocial maternal factors for maltreatment, including young maternal age (Lee & Goerge, 1999; Wu et al., 2004), low educational achievement (Wu et al., 2004), personal history of FATHERS, NEGLECT, AND CPS INVOLVEMENT 5 neglect (Lounds, Borkowski, & Whitman, 2006), alcohol or drug problems (Coohey, 1998; Dube et al., 2001), parenting stress (Taylor, Guterman, Lee, & Rathouz, 2009), and depression (Coohey, 1998; Taylor et al., 2009) are also related to maltreatment. Yet, from the perspective of the developmental ecological model, to best understand child maltreatment in two-parent families it is also necessary to extend examination beyond factors of the maternal primary caregiver to consider (a) the aspects of household composition, (b) the interplay between mothers and fathers, and (c) the interaction of parent and child. It is important to note that due to the limitations of the study design, it is not possible to comprehensively assess the influence of community factors, which are an important aspect of the developmental ecological model. Most research has failed to examine the paternal role in neglect and child maltreatment even though fathers or father surrogates were acknowledged as the perpetrator in 50% of all maltreatment-related child fatalities (U.S. Department of Health and Human Services [USDHHS], 2011). To address this gap, the current study examined paternal individual-level parenting factors potentially associated with increased risk for child maltreatment. In addition, this investigation accounted for household variables (i.e., economic hardship and unemployment) and variables assessing mother-father interaction as well as child characteristics that might influence risk for maltreatment. Maltreatment Risk in Father-Involved Families Research has recognized the complexity of the paternal role in heightening or minimizing family-level risk for child maltreatment (Dubowitz, 2006). Consistent with the developmental ecological model, this research has pointed to the ways in which fathers directly and indirectly contribute to child maltreatment. For example, a father’s absence from the household can indirectly influence maltreatment because such absence is generally associated with fewer FATHERS, NEGLECT, AND CPS INVOLVEMENT 6 household economic resources and increased rates of poverty (Guterman & Lee, 2005). The economic contributions that fathers make to a household are likely to reduce overall household economic strain, which might have an indirect salutary effect on maternal strain and parenting stress. However, maltreatment risk might be heightened when fathers are present in the home, but unemployed (Paxson & Waldfogel, 2002; Wolfner & Gelles, 1993). Therefore, to better understand the paternal role in maltreatment risk, it is important to consider fathers’ unique human capital contributions in the context of overall household economic strain (Slack et al., 2004), as well as other paternal-related protective factors, such as fathers’ attendance at religious services (e.g., Zolotor & Runyan, 2006). Fathers also influence risk for maltreatment via the “immediate interactional context,” which is the father’s interactions with the mother (Belsky, 1993). In general, better quality marital relationships and positive interactions between parents “spillover” to promote enhanced parent-child relations (Erel & Burman, 1995). In contrast, interpersonal violence and aggression is associated with increased harsh punishment of children and child maltreatment (Taylor et al., 2009; Thackeray et al., 2010). The quality of the parental relationship can be especially important when examining fathers’ parenting behaviors and maltreatment of children in fatherinvolved families. Moreover, studies among diverse families have documented the ways in which a positive parenting relationship benefits children (Cabrera, Shannon, & Tamis-LeMonda, 2007; Carlson & McLanahan, 2006; Lunkenheimer, Kittler, Olson, & Kleinberg, 2006). The developmental ecological model also points to the importance of considering both potentiating and compensatory factors such as the nature and context of fathers’ involvement in caring for the child. Even in two-parent families, fathers’ involvement in daily caregiving for young children varies more than mothers’ involvement (Yeung et al., 2011). As such, theorizing FATHERS, NEGLECT, AND CPS INVOLVEMENT 7 around positive father-involvement has held such involvement is a protective factor in relation to child-maltreatment risk (Guterman & Lee, 2005). Findings from one study have shown increased duration of father or father surrogate involvement minimized risk of child neglect, whereas the father’s or father surrogate’s direct involvement in child care heightened the risk for child maltreatment (Dubowitz, Black, Kerr, Starr, & Harrington, 2000). This increased risk might be because fathers, and especially father surrogates, are frequently called upon to provide child care under stressful conditions. The positive effects of father involvement can also be moderated by paternal psychosocial characteristics, with high levels of father involvement associated with positive outcomes for children only if the father was not depressed (Kahn, Brandt, & Whitaker, 2004; Mezulis, Hyde, & Clark, 2004). Nevertheless, few studies have examined paternal variability in caregiving for children might relate to child-maltreatment risk in father-involved families. Overview of the Current Study The current study used the developmental ecological model as a framework to examine the immediate interactional context of child maltreatment (Belsky, 1993) by focusing on individual-level paternal and household variables potentially related to household-level measures of child neglect and CPS involvement among father-involved families of young children with residential biological fathers. Due to limitations of the data used for this study, it is not possible to examine macro or community level factors that, as informed by the ecological model, are also hypothesized to influence maltreatment risk. All variables were assessed when the target child was 5 years or younger. Based on numerous studies that have examined and identified important maternal risks for child neglect and maltreatment, this study’s primary research goal was to examine the contribution of paternal psychosocial risk factors to maltreatment risk, accounting FATHERS, NEGLECT, AND CPS INVOLVEMENT 8 for numerous household variables (e.g., economic conditions, unemployment), which many prior studies have shown to raise risk for maltreatment in father-involved families of young children. The central study hypothesis holds that poor paternal psychosocial functioning, as measured by paternal depression, parenting stress, and alcohol use, is associated with greater risk for maltreatment even after accounting for key family and household factors. To test this hypothesis, models included variables that assessed the immediate interactional context, such as parental relationship quality and child characteristics; paternal demographic characteristics; human capital characteristics; and household economic hardship (Harris, Hilton, Rice, & Eke, 2007; Hussey, Chang, & Kotch, 2006; Knutson, 1995; Lee & Goerge, 1999). In addition, the models incorporated key predictor variables that assessed paternal psychosocial characteristics. All fathers in the study sample were the biological father of the target child; this relationship is important to consider because patterns of maltreatment perpetration differ between men who are biologically related to the child versus men who are not biologically related to the target child (Berger et al., 2009; Harris et al., 2007; Radhakrishna, Bou-Saada, Hunter, Catellier, & Kotch, 2001). The study selected for families in which the mother and father were in an ongoing relationship, therefore the sample did not include single-mother headed households. This sampling approach was expected to bias the sample toward more advantaged individuals (Guzzo & Lee, 2008). Thus, it was anticipated that there would be lower than average levels of CPS involvement and neglect in this subsample of father-involved families. Method Procedure The study sample was obtained from the Fragile Families and Child Wellbeing Study (FFCWS), a community-based cohort study of new births obtained from hospitals in 20 large FATHERS, NEGLECT, AND CPS INVOLVEMENT 9 U.S. cities. The original FFCWS sample (N = 4,898) was collected from 1998 to 2000; detailed description of the FFCWS study design is available elsewhere (see Reichman, Teitler, Garfinkel, & McLanahan, 2001). The original intent of the FFCWS was to examine parenting relationship quality and wellbeing of children in unmarried families. Core interviews were conducted with mothers and fathers at four time points: the birth of the index child (baseline), and when the index child was 1-, 3-, and 5-years-old. Most baseline interviews with mothers took place at the hospital, whereas baseline interviews with fathers were conducted either at hospitals or over the phone (Reichman et al., 2001). Subsequent core interviews took place over the phone. When the index child was age 3 years and 5 years, the mothers who had completed core interviews participated in the In-Home Longitudinal Study of Pre-School Aged Children (hereafter, the InHome Study). The In-Home Study was an observational study conducted in the participants’ home with only the mother and child. In the current study, secondary analyses of the FFCWS examined residential biological fathers, and the characteristics of these fathers that were associated with child neglect and CPS involvement. Participants The FFCWS subsample used in the current study consisted of 3,299 biological fathers who participated in the 3-year core interview. Of this subsample, 161 fathers were excluded because they were in jail at the time of the 3-year core interview and did not complete measures of parenting stress and involvement with the child. Of the 3,138 remaining fathers, only 1,134 were residing in the home at the time of the 5-year In-Home Study interview when mothers completed the dependent measures of CPS involvement and child neglect. An additional 45 fathers were dropped from study inclusion because the mother had not completed the necessary items from the In-Home Study measures, yielding a sample of 1,089 residential, biological FATHERS, NEGLECT, AND CPS INVOLVEMENT 10 fathers in the sample used for the current study. It should be noted that the FFCWS sampling design intentionally oversampled nonmarital births, a factor which contributes to the large number of nonresidential fathers when the child was age 5 (Reichman et al., 2001). Measures Most variables were self-reported by the father. Items collected at the baseline core interview included time invariant demographic variables such as father’s age at time of child’s birth, education level, and race/ethnicity. Father-reported items collected at the 3-year core interview included frequency of religious attendance, current employment status, all measures of household economic hardship, paternal psychosocial characteristics, parental relationship quality, co-parenting support from mother, and child health. Because paternal reports were not available for all relevant study variables, some variables were taken from the mother’s interviews, including male child and child low birth weight (collected at baseline); and child neglect and CPS involvement (collected as part of the 5-year In-Home interview). Not all fathers were interviewed at baseline; therefore, the parents’ marital status at time of child’s birth was taken from the maternal report to avoid the problem of missing data. Dependent variables: Physical child neglect and household CPS involvement. As part of the 5-year In-Home Study, interviewers administered the revised Parent-Child Conflict Tactics Scales (CTS-PC: Straus, Hamby, Finkelhor, Moore, & Runyan, 1998) to measure any instance of child neglect by either the father or mother within the past year (i.e., when the child was between 4 and 5 years old). The CTS-PC asked the mother to report the number of times in the past year she had engaged in the following situations indicating possible neglect: she had to leave the child home alone, even though some adult should have been with the child; she was so caught up with her own problems that she was not able to show or tell child that she loved her FATHERS, NEGLECT, AND CPS INVOLVEMENT 11 child; she was not able to make sure that the child got the food child needed; she was not able to make sure child got to a doctor or hospital when child needed medical attention; and, she was so drunk or high she had a problem taking care of her child. Mothers were then asked the same set of questions but pertaining to the care provided by the child’s father. Because child neglect was infrequent, this study used a dichotomous variable to indicate whether either parent had been involved in any instance of child neglect within in the past year (0 = none; 1 = any instance of child neglect from mother or father in the past year). The variable for CPS involvement was based on the question, “Since the child was born, has Child Protective Services contacted you about any child or children in this household?” Families that had not been contacted by CPS since the index child’s birth were coded “0” and families with an affirmative response to this item were categorized as CPS involved and coded “1”. Following precedent established by other researchers (Berger et al., 2009) who used the same measure of CPS involvement as in the current study, these families were considered to be CPS involved because CPS is unlikely to contact a family regarding a ‘‘screened-out’’ child maltreatment report. Further, as discussed earlier in this article, the CPS involvement variable might be subject to self-report biases similar to those encountered with other proxy measures such as the CTS-PC. A second shortcoming of the measure for CPS involvement is that the measure does not identify which child in the household was the focus of the CPS report, which adults (mother, father, or someone else) were the alleged perpetrators, or which types of maltreatment were alleged (Berger et al., 2009). Therefore, the CPS-involvement measure should be viewed as a general indicator of household risk for CPS involvement rather than a direct indicator of father-perpetrated maltreatment. FATHERS, NEGLECT, AND CPS INVOLVEMENT 12 Paternal demographic and human capital characteristics. The study variables representing demographic and human capital characteristics included age, education (1 = less than high school, 2 = high school diploma or equivalent, 3 = some college or technical school, 4 = college or higher), race/ethnicity (1= White, 2 = African American, 3 = Hispanic, 4 = other), and whether the index child’s birth was a marital birth (0 = not married, 1 = married) to account for parental relationship status. During the 3-year core interview, fathers indicated their frequency of religious attendance in the past year (0 = never or less than once a year, 1 = a few times a year, 2 = a few times a month, 3 = once a week, 4 = a few times a week or every day) and employment status (0 = no, 1 = yes), measured by whether the father had worked for pay in the past week. Household economic hardship. Except as noted, all measures of household economic hardship asked about events during the 12 months before the father’s 3-year core interview. The variable for annual household income summed household income from all sources before taxes and deductions, which included the income of all persons living in the household. The variable for nongovernmental financial assistance indicated receipt of money from any source, including relatives and friends of either the father or mother, but excluded money received from the father’s mother or financial support from any government or private agency (0 = no, 1 = yes). Financial support from all government sources, including assistance through programs such as Temporary Assistance for Needy Families (i.e., TANF), Food Stamps, unemployment insurance, or workers’ compensation , was represented by the dichotomous governmental financial assistance variable (0 = no, 1 = yes). The electricity turned off variable indicated the utility company had interrupted the household’s electrical service because of nonpayment (0= no, 1 = yes). The household stability variable indicated the number of times the family had FATHERS, NEGLECT, AND CPS INVOLVEMENT 13 moved in the 2 years before the 3-year core interview (0 = none, 1 = moved once, 2 = moved two or more times). Another variable measured the number of children younger than 18 years old living in the household; fathers who reported more than five children younger than 18 years residing in the home were top-coded at “5.” Paternal psychosocial characteristics. Assessments of paternal psychosocial characteristics used the Parenting Stress Index-Short Form (PSI-SF; Abidin, 1995); this instrument used a 4-point scale (1 = strongly agree to 4 = strongly disagree) to measure the father’s agreement with four statements such as “I feel trapped by my responsibilities as a parent” (α = .62). A variable for involvement with the child indicated the average number of days per week (0 = never to 7 = every day) the father provided each of 13 common types of care to the child (e.g., sing songs or nursery rhymes with child, read stories to child, assist child with eating; α = .89). The Composite International Diagnostic Interview-Short Form: Section A (Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998) was used to measure incidence of paternal depression in the past year that lasted for 2 weeks or more, and whether the symptoms lasted for most of the day and occurred every day of the 2-week period. Respondents were classified as depressed if they endorsed the screening items and three or more symptoms (0 = no, 1 = yes). The variable alcohol use was based on paternal self-report, and indicated the largest number of alcoholic drinks the father consumed in a single day during the past 12 months (0 = no drinks consumed in the past 12 months, 1 = one to three drinks consumed in any single day during the past 12 months, 2 = four or more drinks consumed in any single day during the past 12 months). According to the National Institute on Alcohol and Alcoholism (2005), heavy drinking for men is defined as five or more drinks in a single day. FATHERS, NEGLECT, AND CPS INVOLVEMENT 14 Relationship with the child’s mother and child characteristics. Fathers indicated parental relationship quality (1 = poor to 5 = excellent) as well as co-parenting support from the child’s mother (1 = never true to 4 = always true), indicated by six items, including “You can trust the mother to take good care of the child” and “You and the mother talk about problems that come up raising the child” (α = .61). Variables representing child characteristics included male child (0 = female, 1 = male) and low birth weight (0 = no, 1 = yes, child weighed < 2,500 grams at birth). The variable for child health indicated the father’s evaluation of the child’s health and development at age 3 years (1 = poor to 5 = excellent). Statistical Analysis Plan Analyses were conducted using Stata 10.1. Sample characteristics and bivariate results with Bonferroni corrected post-hoc tests (i.e., chi-square and one-way analysis of variance) are presented in Tables 1 and 2. Table 1 examines between-group differences by comparing families with no neglect to those in which neglect was present in the past year, measured when the child was age 5; Table 2 between-group differences by comparing families with no CPS involvement since the child’s birth to those with CPS involvement since the child’s birth. -----------------------------------------------------Please place Table 1 and Table 2 about here -----------------------------------------------------Table 3 and Table 4 present odds ratios (OR) and 95% confidence intervals (CI) for logistic regression analyses examining child neglect and CPS involvement, respectively. Associations were tested in a stepwise fashion to examine the relative contribution of sets of variables. In each table, Model 1 examined the associations of paternal demographic and human capital characteristics to the dependent variable; Model 2 assessed the additional contribution of FATHERS, NEGLECT, AND CPS INVOLVEMENT 15 household economic hardship variables; Model 3 added psychosocial characteristics; and Model 4 included family-level characteristics, including parental-relationship quality and child characteristics. All regression models were adjusted for two key aspects of the FFCWS sampling design: interview city and parental marital status at birth of the index child. -----------------------------------------------------Please place Table 3 and Table 4 about here -----------------------------------------------------Results Among the data collected from the 1,089 families meeting the inclusion criteria for the current study, variables had less than 1% missing data except: low birth weight (2.75%), coparenting support from mother (1.65%), and parental relationship quality (1.19%). Therefore, the final analytic sample reported in Table 3 contained data from 1,016 fathers. Table 1 includes the maternal reports of child neglect by either the mother or the father, and shows that 11.85% of mothers reported at least one instance of neglect. Of these instances of neglect, 9% were cases of maternal-related neglect (n = 98) and 8.26% were cases of father-related neglect (n = 90). In the majority of cases, neglect occurred at least once from both parents (5.42%) whereas 3.58% were mother-only and 2.85% were father-only cases of neglect. In the regression models for CPS involvement (Table 4), none of the fathers who identified as other race/ethnicity had been CPS involved. Given the perfect correlation between other race/ethnicity and CPS involvement, these fathers were dropped from the analyses, yielding a sample of 1,000 families. Of these families, 6.4% (n = 64) had been contacted by CPS about any child in the household since the index child was born (see Table 2). Household Child Neglect FATHERS, NEGLECT, AND CPS INVOLVEMENT 16 As seen in Table 3, African American, Hispanic, and other race/ethnicity status fathers (as compared with White) was associated with greater odds of parental neglect in Model 1; however, these association were no longer significant for African American fathers after accounting for household economic hardship in Model 2 or for other race/ ethnicity fathers after accounting for paternal psychosocial characteristics in Model 3. Model 2 suggests a connection between current employment and annual household income. Current employment was associated with increased risk of parental child neglect. When adding psychosocial variables in Model 3, paternal depression was associated with more than double odds of child neglect. In addition, higher levels of paternal parenting stress and having one to three alcoholic drinks in any one day within the past year were also associated with greater risk for neglect. In contrast, higher levels of paternal positive involvement with the child were protective against risk for neglect. Further, as compared with Model 2, the odds ratio for current employment increased risk of child neglect. Model 4 did not yield greatly altered patterns of results and none of the variables added in Model 4 (parental relationship quality, co-parenting support, male child, child health, low birth weight) were significantly associated with risk for parental neglect. However, Hispanic race/ethnicity status (compared with White) was associated with a 58% increase in the odds of child neglect. Other risk factors for neglect identified in Model 4 included current employment and lower household income. In this final model, presence of paternal depression nearly doubled the odds of neglect, and the parenting stress and alcohol use variables were each associated with an approximate 50% increased odds of neglect. However, paternal involvement remained a significant protective factor against risk of child neglect. CPS Involvement FATHERS, NEGLECT, AND CPS INVOLVEMENT 17 In Model 1, a higher level of paternal education was associated with a lower level of CPS involvement (see Table 4). The patterns of ORs for education were similar between Model 1 and Model 2, but achieved statistical significance only when comparing educational level of a college degree or higher with an educational level of less than a GED. In Model 3, receipt of nongovernmental and governmental assistance were associated with increased odds of CPS involvement. The odds associated with CPS involvement increased in relation to the number of young children in the home. Adding variables in Model 3 and Model 4 did not change the patterns observed in Model 2. Consistent with the results obtained for neglect in Table 3, Models 3 and 4 in Table 4 demonstrated that paternal depression was associated with more than double the odds of CPS involvement. Discussion Child welfare statistics indicate that neglect is the most pervasive form of child maltreatment in the United States. In 2008, child neglect accounted for 71.1% of all substantiated maltreatment cases (U.S. Department of Health and Human Services [USDHHS], 2010). From the early 1990s to 2003, there was a nearly 50% decrease in substantiated cases of child sexual abuse and 36% fewer substantiated cases of child physical abuse. In comparison, child neglect has proven to be a far more intractable form of maltreatment, with only a 7% decrease in neglect during that same period (Jones, Finkelhor, & Halter, 2006). Moreover, recurrence of maltreatment is most common when neglect is the primary complaint (Hindley, Ramchandani, & Jones, 2006). Specifically, the intractable nature of child neglect might be due, at least in part, to the fact that risk for maltreatment is multiply determined and no single, common factor has been identified that can be easily targeted in interventions to prevent maltreatment. For example, FATHERS, NEGLECT, AND CPS INVOLVEMENT 18 developmental ecological theory (Belsky, 1994) points to numerous community-, household-, family-, and individual-level factors that are related to risk for child maltreatment. However, few prior studies have examined child neglect and CPS involvement in father-involved families of young children; this gap is problematic for two reasons: fathers are present in many households, and research shows biological fathers are more likely to be involved in maltreatment of young children than older children (U.S. Department of Health and Human Services [USDHHS], 2005). As anticipated, the rate of maternal-reported CPS involvement in this sample of fatherinvolved families was low, with 4% less CPS involvement than what was observed in a study sample (Berger et al., 2009) that included dual- and single-parent households sample and that used the same measure of CPS involvement as was used in the current study. CPS involvement was 10.4% in a sample of 2,927 families in which the mother was in an ongoing relationship (i.e., living with either the child’s father or another partner/ spouse, or was in a dating relationship) or was not romantically involved; this rate is comparable to those obtained from administrative data for similar populations (Berger et al., 2009). However, because nearly 12% of mothers indicated that she or the father had engaged in acts of neglect (e.g., leaving the child unsupervised, not making sure the child had enough food, or being too drunk or high to care for the child) in the past year, the self-reported measure of CPS involvement used in this study likely underreports actual maltreatment. Consistent with the developmental ecological model, the results of this study (Tables 3 and 4) suggest that risk for maltreatment was determined by multiple factors in this sample of father-involved families. In this study, individual-level parenting factors as well as demographic, human capital, and household economic hardship variables were related to risk for neglect and FATHERS, NEGLECT, AND CPS INVOLVEMENT 19 CPS involvement. The patterns among these variables differed depending on the outcome being examined; these results are discussed in detail later in this article. Paternal Psychosocial Characteristics Study results confirmed a central hypothesis of the study: namely, that poor paternal psychosocial functioning (measured by paternal depression, parenting stress, and alcohol use) is associated with increased risk for maltreatment, even after accounting for key family and household factors. Specifically, paternal depression was a strong risk factor for both CPS involvement and neglect. Rates of paternal depression were more than double among the fathers involved in child neglect and the CPS-involved fathers. Paternal depression not only doubled the odds of child neglect but also raised the odds of CPS involvement. In addition, paternal parenting stress and higher levels of alcohol use were associated with elevated risk for parental child neglect (see Table 3, Model 4) but not for CPS involvement (see Table 4, Model 4). Results also confirmed the important role of compensatory factors in child maltreatment cases. In this study of biological, residential fathers, higher levels of father involvement in common child-care activities was a protective factor for child neglect only (see Table 3, Model 4). This finding stands in contrast to Dubowitz et al.’s (2000) finding that more father involvement in child care was associated with heightened risk for neglect among a sample that included biological fathers, social fathers, stepfathers, and other male caregivers of young children. This apparent paradox may be explained by the nature and context of father involvement among social versus biological fathers. Social father-child relationships might be more fraught, particularly in a high-risk sample, and especially when compared with father-child relationships when fathers are biological and residential, as in the current study, with relatively high and presumably more stable levels of involvement. Although participants in both the current FATHERS, NEGLECT, AND CPS INVOLVEMENT 20 study and the Dubowitz et al. (2000) study were recruited from large urban areas, the rates of neglect reported by Dubowitz et al. (2000) ranged from 11% to 30%. In comparison, on the whole, the characteristics of the current sample of married or co-residential parents (see Table 1) would not necessarily be considered high-risk and indeed the sampling procedure for the FFCWS was intended to reflect urban communities rather than a high-risk sample. Moreover, the incidence of neglect in the current study was lower than that reported by Dubowitz et al. (2000). That father involvement was not associated with CPS involvement is consistent with other studies that have shown father involvement was not associated with paternal physical aggression, psychological aggression, or spanking (Lee, Guterman, & Lee, 2008; Lee, Perron, Taylor, & Guterman, 2011). When viewed together, these studies suggest that positive paternal involvement might be uniquely influential as it relates to the potential for child neglect, and as underscored by the results from the current study and Dubowitz et al. (2000), the protective benefits of father involvement may vary as a function of the nature and context of the fathering role. Demographic, Human Capital, and Household Economic Hardship Demographic and household characteristics, such as more children younger than 18 years in the home, receipt of financial assistance (from governmental and non-governmental sources) and low paternal education were linked to CPS involvement but were not linked to neglect (see Model 4 in Table 3 and Table 4). When comparing the final models for neglect and CPS involvement, paternal psychosocial variables, such as parenting stress and alcohol use, stand out as significantly associated with household parental neglect, whereas economic factors seem to play a larger role in predicting CPS involvement. These results may suggest that having more children in the household and relying on governmental assistance are influential in bringing FATHERS, NEGLECT, AND CPS INVOLVEMENT 21 families to the attention of CPS, whereas paternal psychosocial functioning is most likely to interfere with a father’s ability to provide basic emotional support and caregiving to his child. Although prior research has shown that paternal unemployment could heighten maltreatment risk, the type of maltreatment examined in some studies was either physical aggression or unspecified (for example, see Paxson & Waldfogel, 2002; Wolfner & Gelles, 1993). The results of the current study pointed to an interesting juxtaposition of factors and associations pertaining to fathers’ current employment, household income, and perhaps race/ ethnicity, with fathers’ current employment heightened their risk for child neglect (see Table 3). The measure of child neglect used in this study may have been sensitive to the issue of parental employment, since one of the five items asked about leaving the child home alone, even though some adult should have been with the child, a phenomenon that may be more common when parents are employed. An important direction for future research would be to further examine how the interaction of fathers’ employment status, mothers’ employment status, and household income are related to risk for child maltreatment and examine whether young children in families with two working parents may be more likely to be left unsupervised or unattended. Study Limitations This study examined a subset of father-involved families and, therefore, the results cannot be generalized to other types of male caregivers such as nonresidential biological fathers, father surrogates (e.g., stepfathers) or male caregivers who are not biologically related to the children in their care. All families were living in urban areas; therefore, the findings of this study do not generalize to families living in nonurban areas. It is also important to note that this study utilizes secondary data, and the original intention of the FFCWS was not to study child FATHERS, NEGLECT, AND CPS INVOLVEMENT 22 maltreatment but rather to examine family structure, specifically, parenting relationship quality and child wellbeing in unmarried families. The selection of father-involved families biases the sample toward more advantaged individuals (Guzzo & Lee, 2008). Prior research has extensively examined differences among married and unmarried fathers in the FFCWS sample. Such studies have shown that as compared with married fathers, unmarried fathers were younger at the time of their child’s birth, had lower incomes, were less well-educated, and were less involved with their children (Carlson & McLanahan, 2010). In turn, these characteristics of unmarried fathers contributed to relationship instability and dissolution of cohabiting unions over the first years of a child’s life (Carlson, McLanahan, & England, 2004). Although the CTS-PC subscale measure of neglect used in this study is a valid and reliable proxy measure for child maltreatment, a limitation of the CTS-PC is that the instrument assesses only a limited set of neglectful behaviors, including physical and supervisory neglect. Future research should examine a wider range of potential neglectful barriers (e.g., educational neglect) to more fully inform prevention efforts (e.g., Dubowitz et al., 2005; English et al., 2005). Moreover, this study relied on maternal reports to the CTS-PC and paternal report of maltreatment was unavailable. However, this study included only residential fathers, and research using the CTS-PC has indicated that in two-parent families, mothers are reliable and accurate reporters of the fathers’ behavior toward their children (Lee, Lansford, Pettit, Dodge, & Bates, 2012). Lee et al. (2012) compared maternal reports of fathers’ behaviors with fathers’ self-reports and found that mothers tended to underreport fathers’ negative behaviors toward the child. Therefore, it is reasonable to assume that our use of maternal report most likely resulted in an underreporting of child neglect, which would result in conservative estimate of the influence FATHERS, NEGLECT, AND CPS INVOLVEMENT 23 of predictor variables and bias the study toward null findings. As with all studies that use selfreported measures, maternal responses must be viewed as a proxy for actual behavior, particularly when the behaviors being measured are socially undesirable acts such as a parent failing to properly care for his or her child. As discussed elsewhere (Berger et al., 2009), the measure of CPS involvement has several limitations. The measure does not isolate the timing when CPS intervention occurred, nor does it clearly identify the perpetrator(s) of maltreatment that prompted the CPS report or whether the report was filed in reference to the study index child. Results should be interpreted to highlight that associations were with household CPS involvement rather than fatherperpetrated maltreatment per se. Given limitations of the available data available, it is impossible to precisely test associations of father-perpetrated CPS involvement. Although preferable to compare the self-report CPS measure with administrative records, such data is not available. Prior research has shown that CPS involvement is higher in neighborhoods with high rates of poverty (Coulton, Crampton, Irwin, Spilsbury, & Korbin, 2007; Lee & Goerge, 1999). Another important limitation of the current study is related to the structure of the FFCW study that makes it impossible to comprehensively assess the role of community factors that are an important aspect of the developmental ecological model. This limitation should be considered in light of the comprehensive measurement of father, child, and household characteristics allowing us to examine factors that may be most proximal in terms of influence maltreatment risk, particularly of young children such as those examined in the current study. A future direction for research is to examine how community factors, such as collective efficacy and social cohesion, may influence paternal parenting strain and risk for maltreatment (e.g., Guterman, Lee, Taylor, & Rathouz, 2009). FATHERS, NEGLECT, AND CPS INVOLVEMENT 24 Despite these limitations, the current study represents an advance in research on paternal parenting for several reasons. Most measures used in the study were self-reported by the fathers. Most existing studies have focused on the highest-risk families, that is, single-parent households. In such studies, father involvement is often crudely measured based solely on the presence or absence of the father in the home. For a detailed discussion of issues in measuring father involvement, c.f. Guterman & Lee, 2005 and Lee, Bellamy, & Guterman, 2009. Further, such studies have not clarified whether the household father figure is a biological father, father surrogate, or in some other caregiving relationship with the child. Therefore, the self-report data and measures of father involvement used in this study (e.g., father involvement in child caregiving activities, marital status, and presence in the home when index child was 5 years) are an advance over prior research. In addition, the results of this study are enhanced by the nature of the study design that allowed for prospective analyses of multiple individual-level and household-level characteristics as well as maternal- and child-related characteristics. Conclusions The study findings have several implications for research and practice. Many of the father-related risk factors for maltreatment identified in this study were similar to those observed in studies of mothers. For example, both parenting stress and alcohol use have been identified here and elsewhere as risk factors for child neglect. This study provided evidence that when examining the paternal role in child maltreatment, it is imperative to consider the nature and context of father involvement rather than simply the presence or absence of the father from the home. Results of this study indicate that even in this sample of father-involved families, which was presumed to have relatively high levels of father involvement given that all fathers were FATHERS, NEGLECT, AND CPS INVOLVEMENT 25 married or co-residing with the child’s mother, father involvement had a salutary effect; specifically, father involvement was related to lowered levels of child neglect. Perhaps the strongest conclusion based on study findings is the further evidence that paternal depression compromises the father-child relationship (Bronte-Tinkew, Moore, Matthews, & Carrano, 2007). Much of the existing research indicates that maternal depression is associated with numerous negative child outcomes; similarly, this study suggests that paternal depression likewise serves as an important risk factor for maltreatment in father-involved families. Although demographic and human capital characteristics are difficult to change in the short-term, psychosocial risk factors are malleable. Therefore, an important implication of this study is the need to assess and treat depression to promote optimal fathering. Thus, recommendations include the need to screen fathers, as well as mothers, at prenatal and perinatal clinic visits (Flaherty & Stirling, 2010), as well as to include both mothers and fathers in the provision of evidence-based services to prevent child maltreatment in pediatric health care settings (e.g., Safe Environment for Every Kid; Dubowitz, Feigelman, Lane, & Kim, 2009). FATHERS, NEGLECT, AND CPS INVOLVEMENT 26 References Abidin, R. (1995). Parent stress inventory (3rd ed.). Odessa, FL: Psychological Assessments Resources. Belsky, J. (1993). Etiology of child maltreatment: A developmental-ecological analysis. Psychological Bulletin, 114(3), 413-434. doi:10.1037/0033-2909.114.3.413 Berger, L. M. (2004). Income, family structure, and child maltreatment risk. Children and Youth Services Review, 26, 725-748. doi:10.1016/j.childyouth.2004.02.017 Berger, L. M., Paxson, C., & Waldfogel, J. (2009). Mothers, men, and child protective services involvement. Child Maltreatment, 14(3), 263-276. doi:10.1177/1077559509337255 Brewster, A. L., Nelson, J. P., Hymel, K. P., Colby, D., Lucas, D. R., McCanne, T. R., & Milner, J. S. (1998). Victim, perpetrator, family, and incident characteristics of 32 infant maltreatment deaths in the United States Air Force. Child Abuse & Neglect, 22, 91-101. doi:10.1016/S0145-2134(97)00132-4 Bronte-Tinkew, J., Moore, K. A., Matthews, F., & Carrano, J. (2007). Symptoms of major depression in a sample of fathers of infants: Sociodemographic correlates and links to father involvement. Journal of Family Issues, 28, 61-99. doi:10.1177/0192513X06293609 Cabrera, N. J., Shannon, J. D., & Tamis-LeMonda, C. (2007). Fathers' influence on their children's cognitive and emotional development: From toddlers to pre-K. Applied Developmental Science, 11, 208-213. doi:10.1080/10888690701762100 Carlson, M. J., & McLanahan, S. S. (2006). Strengthening unmarried families: Could enhancing couple relationships also improve parenting? Social Service Review, 80, 297-321. doi:10.1086/503123 FATHERS, NEGLECT, AND CPS INVOLVEMENT 27 Carlson, M. J., & McLanahan, S. S. (2010). Fathers in fragile families. In M. E. Lamb (Ed.), The role of the father in child development (5th ed.). Hoboken, NJ: John Wiley & Sons. Carlson, M. J., McLanahan, S. S., & England, P. (2004). Union formation in fragile families. Demography, 41, 237-261. doi:10.1353/dem.2004.0012 Coohey, C. (1998). Home alone and other inadequately supervised children. Child Welfare, 77(3), 291-310. Coohey, C. (2003). Making judgments about risk in substantiated cases of supervisory neglect. Child Abuse & Neglect, 27, 821-840. doi:10.1016/S0145-2134(03)00115-7 Coulton, C. J., Crampton, D. S., Irwin, M., Spilsbury, J. C., & Korbin, J. E. (2007). How neighborhoods influence child maltreatment: A review of the literature and alternative pathways. Child Abuse & Neglect, 31, 1117-1114. doi:10.1016/j.chiabu.2007.03.023 Dube, S. R., Anda, R. F., Felitti, V. J., Croft, J. B., Edwards, V. J., & Giles, W. H. (2001). Growing up with parental alcohol abuse: Exposure to childhood abuse, neglect, and household dysfunction. Child Abuse & Neglect, 25, 1627-1640. doi:10.1016/S01452134(01)00293-9 Dubowitz, H. (2006). Where's dad? A need to understand father's role in child maltreatment. Child Abuse & Neglect, 30, 461-465. doi:10.1016/j.chiabu.2006.04.002 Dubowitz, H., Black, M. M., Kerr, M. A., Starr, R. H., Jr., & Harrington, D. (2000). Fathers and child neglect. Archives of Pediatrics & Adolescent Medicine, 154, 135-141. Dubowitz, H., Newton, R. R., Litrownik, A. J., Lewis, T., Briggs, E. C., Thompson, R., … Feerick, M. M. (2005). Examination of a conceptual model of child neglect. Child Maltreatment, 10(2), 173-189. doi:10.1177/1077559505275014 FATHERS, NEGLECT, AND CPS INVOLVEMENT 28 English, D. J., Thompson, R., Graham, J. C., & Briggs, E. C. (2005). Toward a definition of neglect in young children. Child Maltreatment, 10, 190-206. doi:10.1177/1077559505275178 Erel, O., & Burman, B. (1995). Interrelatedness of marital relations and parent-child relations: A meta-analytic review. Psychological Bulletin, 118, 108-132. doi:10.1037/00332909.118.1.108 Flaherty, E. G., & Stirling, J., Jr. (2010). The pediatrician’s role in child maltreatment prevention. Pediatrics, 126, 833–841. Fujiwara, T., Barber, C., Schaechter, J., & Hemenway, D. (2009). Characteristics of infant homicides: Findings from a U.S. multisite reporting system. Pediatrics, 124, e210-e217. doi:10.1542/peds.2008-3675 Guterman, N. B., & Lee, Y. (2005). The role of fathers in risk for physical child abuse and neglect: Possible pathways and unanswered questions. Child Maltreatment, 10, 136-149. doi:10.1177/1077559505274623 Guterman, N. B., Lee, S. J., Taylor, C. A., & Rathouz, P. J. (2009). Parental perceptions of neighborhood processes, stress, personal control, and risk for physical child abuse and neglect. Child Abuse & Neglect, 22, 897-906. doi: 10.1016/j.chiabu.2009.09.008 Guzzo, K. B., & Lee, H. (2008). Couple relationship status and patterns in early parenting practices. Journal of Marriage and Family, 70, 44-61. doi:10.1111/j.17413737.2007.00460.x Harris, G. T., Hilton, N. Z., Rice, M. E., & Eke, A. W. (2007). Children killed by genetic parents versus stepparents. Evolution and Human Behavior, 28, 85-95. doi:10.1016/j.evolhumbehav.2006.08.001 FATHERS, NEGLECT, AND CPS INVOLVEMENT 29 Hindley, N., Ramchandani, P. G., & Jones, D. P. H. (2006). Risk factors for recurrence of maltreatment: A systematic review. Archives of Diseases in Childhood, 91, 744-752. doi:10.1136/adc.2005.085639 Hussey, J. M., Chang, J. J., & Kotch, J. B. (2006). Child maltreatment in the United States: Prevalence, risk factors, and adolescent health consequences. Pediatrics, 118, 933-942. doi:10.1542/peds.2005-2452 Jones, L. M., Finkelhor, D., & Halter, S. (2006). Child maltreatment trends in the 1990s: Why does neglect differ from sexual and physical abuse? Child Maltreatment, 11(2), 107-120. doi:10.1177/1077559505284375 Kahn, R. S., Brandt, D., & Whitaker, R. C. (2004). Combined effect of mothers' and fathers' mental health symptoms on children's behavioral and emotional well-being. Archives of Pediatric and Adolescent Medicine, 158, 721-729. doi:10.1001/archpedi.158.8.721 Kessler, R. C., Andrews, G., Mroczek, D., Ustun, B., & Wittchen, H. U. (1998). The World Health Organization Composite International Diagnostic Interview Short-Form (CIDISF). International Journal of Methods in Psychiatric Research, 7(4), 171-185. doi:10.1002/mpr.47 Knutson, J. F. (1995). Psychological characteristics of maltreated children: Putative risk factors and consequences. Annual Review of Psychology, 46, 401-431. doi:10.1146/annurev.ps.46.020195.002153 Lee, B. J., & Goerge, R. M. (1999). Poverty, early childbearing, and child maltreatment: A multinomial analysis. Children and Youth Services Review, 21, 755-780. doi:10.1016/S0190-7409(99)00053-5 FATHERS, NEGLECT, AND CPS INVOLVEMENT 30 Lee, S. J., Bellamy, J. B., & Guterman, N. B. (2009). The role of fathers in risk for physical child abuse and neglect: Possible pathways and unanswered questions. Child Maltreatment, 14, 227-231. doi:10.1177/1077559509339388 Lee, S. J., Guterman, N. B., & Lee, Y. (2008). Risk factors for paternal physical child abuse. Child Abuse & Neglect, 32, 846-858. doi:10.1016/j.chiabu.2007.11.006 Lee, S. J., Lansford, J. E., Pettit, G. S., Bates, J. E., & Dodge, K. A. (2012). Parental agreement of reporting parent to child aggression using the Conflict Tactics Scales. Child Abuse & Neglect, 36, 510-518. doi: http://dx.doi.org/10.1016/j.chiabu.2012.04.005 Lee, S. J., Perron, B. E., Taylor, C. A., & Guterman, N. B. (2011). Paternal psychosocial characteristics and corporal punishment of their 3-year-old children. Journal of Interpersonal Violence, 26, 71-87. doi:10.1177/0886260510362888 Lounds, J. J., Borkowski, J. G., & Whitman, T. L. (2006). The potential for child neglect: The case of adolescent mothers and their children. Child Maltreatment, 11, 281-294. doi:10.1177/1077559506289864 Lunkenheimer, E. S., Kittler, J. E., Olson, S. L., & Kleinberg, F. (2006). The intergenerational transmission of physical punishment: Differing mechanisms in mothers' and fathers' endorsement? Journal of Family Violence, 21, 509-519. doi:10.1007/s10896-006-9050-2 Mezulis, A. H., Hyde, J. S., & Clark, R. (2004). Father involvement moderates the effect of maternal depression during a child's infancy on child behavior problems in kindergarten. Journal of Family Psychology, 18, 575-588. doi:10.1037/0893-3200.18.4.575 Paxson, C., & Waldfogel, J. (1999). Parental resources and child abuse and neglect. American Economic Review, 89, 239-244. doi:10.1257/aer.89.2.239 FATHERS, NEGLECT, AND CPS INVOLVEMENT 31 Paxson, C., & Waldfogel, J. (2002). Work, welfare, and child maltreatment. Journal of Labor Economics, 20(3), 435-475. doi:10.1086/339609 Radhakrishna, A., Bou-Saada, I. E., Hunter, W. M., Catellier, D. J., & Kotch, J. B. (2001). Are father surrogates a risk factor for child maltreatment? Child Maltreatment, 6, 281-289. doi:10.1177/1077559501006004001 Reichman, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 32, 303-326. doi.org/10.1016/S0190-7409(01)00141-4 Slack, K. S., Holl, J. L., Lee, B. J., McDaniel, M., Yoo, J., & Bolger, K. (2004). Understanding the risks of child neglect: An exploration of poverty and parenting characteristics. Child Maltreatment, 9, 395-408. doi:10.1177/1077559504269193 Straus, M. A., Hamby, S. L., Finkelhor, D. W., Moore, D. W., & Runyan, D. (1998). Identification of child maltreatment with the Parent-Child Conflict Tactics Scales: Development and psychometric data for a national sample of American parents. Child Abuse & Neglect, 22, 249-270. doi:10.1016/S0145-2134(97)00174-9 Taylor, C. A., Guterman, N. B., Lee, S. J., & Rathouz, P. (2009). Intimate partner violence, maternal stress, nativity, and risk for maternal maltreatment of young children. American Journal of Public Health, 99, 175-183. doi:10.2105/AJPH.2007.126722 Thackeray, J. D., Hibbard, R., Dowd, M. D., Committee on Child Abuse and Neglect, & Committee on Injury, Violence and Poison Prevention. (2010). Intimate partner violence: The role of the pediatrician. Pediatrics, 125, 1094-1100. doi:10.1542/peds.2010-0451 FATHERS, NEGLECT, AND CPS INVOLVEMENT 32 Theodore, A., Runyan, D., & Chang, J. J. (2007). Measuring the risk of physical neglect in a population-based sample. Child Maltreatment, 12, 96-105. doi:10.1177/1077559506296904 U.S. Department of Health and Human Services [USDHHS], Office of the Assistant Secretary for Planning and Evaluation. (2005). Male perpetrators of child maltreatment: Findings from NCANDS. Available from http://www.ncjrs.gov/App/Publications/abstract.aspx?ID=210576 U.S. Department of Health and Human Services [USDHHS], Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. (2010). Child Maltreatment 2008. Available from http://www.acf.hhs.gov/programs/cb/stats_research/index.htm#can U.S. Department of Health and Human Services [USDHHS], Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. (2011). Child Maltreatment 2010. Available from http://www.acf.hhs.gov/programs/cb/stats_research/index.htm#can Wolfner, G. D., & Gelles, R. J. (1993). A profile of violence toward children: A national study. Child Abuse and Neglect, 17, 197-212. doi:10.1016/0145-2134(93)90040-C Wu, S. S., Ma, C. X., Carter, R. L., Ariet, M., Feaver, E. A., Resnick, M. B., & Roth, J. (2004). Risk factors for infant maltreatment: A population-based study. Child Abuse & Neglect, 28, 1253-1264. doi:10.1016/j.chiabu.2004.07.005 Yeung, J.W., Sandberg, J.F., Davis-Kean, P.E., & Hofferth, S.L. (2011). Children's time with fathers in intact families. Journal of Marriage and Family, 63, 136-154, FATHERS, NEGLECT, AND CPS INVOLVEMENT 33 Zolotor, A. J., & Runyan, D. K. (2006). Social capital, family violence, and neglect. Pediatrics, 117, 1124-1131. doi:10.1542/peds.2005-1913 FATHERS, NEGLECT, AND CPS INVOLVEMENT 34 Table 1 Descriptive Statistics of Study Variables by Child Neglect at Age 5 years Full Sample Neglect – No Neglect – Yes N = 1,089 (100%) n = 960 (88.15%) n = 129 (11.85%) % or M (SD) % or M (SD) % or M (SD) 29.42 (7.04) 29.53 (7.11) 28.64 (6.47) Less than high school degree 25.85 25.31 29.92 High school degree or equivalent 27.52 27.30 29.13 Some college 26.69 27.20 22.83 College degree or higher 19.94 20.19 18.11 White (non-Hispanic) 32.32 33.33 24.81 African-American 35.35 35.21 36.43 Hispanic 28.56 27.81 34.11 Other 3.76 3.65 4.65 46.78 47.65 40.31 1.82 (1.37) 1.83 (1.38) 1.73 (1.36) 87.96 87.80 89.15 216.36 (91.64) 219.36 (92.66) a 194.06 (80.58) a Nongovernmental financial assistance, y 24.77 24.35 27.91 Governmental financial assistance, y 3.49 3.23 5.47 Electricity turned off in past year, y 17.57 17.54 17.83 No moves 61.55 61.17 64.34 Moved 1 time 29.07 30.06 21.71 Moved 2+ times 9.38 8.77 13.95 2.02 (1.22) 2.02 (1.21) 2.02 (1.31) Parenting stress (1-4) 2.04 (0.66) 2.02 (0.65) a 2.24 (0.68) a Involvement with the child (0-6.77) 4.47 (1.09) 4.49 (1.09) 4.32 (1.11) Variable (Range) Father age (15-61 years) Father education Race/ ethnicity Marital birth, y Freq. of religious attendance (0-4) Current employment status, y Annual household income (1-1000) Household stability * # of children in household < 18 yrs (05) Depression, y Alcohol use 10.28 9.06 b 19.38 b FATHERS, NEGLECT, AND CPS INVOLVEMENT 35 0 drinks in one day in past yr 30.36 31.32 23.26 1-3 drinks in one day in past yr 39.56 38.41 48.06 4+ drinks in one day in past yr 30.08 30.27 28.68 Parental relationship quality (1-5) 4.20 (0.92) 4.21 (0.86) a 4.06 (0.92) a Co-parenting support from mother 3.86 (0.23) 3.87 (0.22) a 3.82 (0.26) a 52.07 52.50 48.84 4.61 (0.64) 4.61 (0.65) 4.62 (0.64) 7.46 6.42 1.04 (1.17-4) Male child, y Child health (2-5) Child low birth weight , y Note: Cell percents may not add to 100% due to rounding. Annual total household income was square-root transformed for all analyses; actual mean = $53,555 (SD = $43,239), median = $43,000. a Paired superscripts indicate Bonferroni corrected one-way analysis of variance (ANOVA), p ≤ .05. b Paired superscripts indicate Pearson χ2 test significantly different at p ≤ .05. * Pearson χ2 test significant p ≤ .05 for the omnibus test. FATHERS, NEGLECT, AND CPS INVOLVEMENT 36 Table 2 Descriptive Statistics of Study Variables by CPS Involvement (n = 1,000) CPS Involvement – No CPS Involvement – Yes n = 936 (93.6%) n = 64 (6.4%) % or M (SD) % or M (SD) 29.37 (7.03) 27.77 (6.10) Less than high school degree 25.53 39.06 High school degree or equivalent 27.99 34.38 Some college 26.92 25.00 College degree or higher 19.55 1.56 White (non-Hispanic) 33.87 26.56 African American 35.58 54.69 Hispanic 30.56 18.75 Marital birth, y 47.22 a 32.81 a Freq. of religious attendance 1.86 (1.38) a 1.46 (1.43) a Current employment status, y 88.14 Variable (Range) Father age Father education* Race/ ethnicity* Annual household income 85.94 217.73 (89.18) a 185.99 (67.79) a Nongovernmental financial assistance, y 24.04 b 37.50 b Governmental financial assistance, y 16.03 b 32.81 b Electricity turned off in past year, y 3.10 b 7.81 b No moves 62.50 67.19 Moved 1 time 28.85 23.44 Moved 2+ times 8.64 Household stability 9.38 a 2.59 (1.40) a # of children in household < 18 yrs 2.02 (1.20) Parenting stress 2.03 (0.66) 2.05 (0.56) Involvement with the child 4.44 (1.09) 4.64 (0.94) 8.65 b 21.88 b 0 drinks in one day in past yr 30.02 31.25 1-3 drinks in one day in past yr 39.42 32.81 4+ drinks in one day in past yr 30.56 35.94 Depression, y Alcohol use FATHERS, NEGLECT, AND CPS INVOLVEMENT 37 Parental relationship quality 4.21 (0.84) a 3.95 (1.03) a Co-parenting support from mother 3.86 (0.23) 3.84 (0.20) 52.14 42.19 Male child, y Child health (2-5) Child low birth weight , y 4.61 (0.64) a 7.05 4.41 (0.75) a 10.94 Note: Cell percents may not add to 100% due to rounding. Annual total household income was square-root transformed for all analyses. See Table 1 for the household income mean, SD, and median. a Paired superscripts indicate Bonferroni corrected one-way analysis of variance (ANOVA), p ≤ .05. b Paired superscripts indicate Pearson χ2 test significantly different at p ≤ .05. * Pearson χ2 test significant p ≤ .05 for the omnibus test. FATHERS, NEGLECT, AND CPS INVOLVEMENT 38 Table 3 Logistic Regression Models of Paternal and Household Characteristics Measured at Child’s Age 3 or Earlier That Predict Child Neglect at Age 5 Years (n = 1,016) Model 1 Model 2 Model 3 Model 4 OR (CI) OR (CI) OR (CI) OR (CI) Demographic & Human Capital Characteristics Father age (in years) Father education 0.99 (0.97-1.02) a --- 1.00 (0.97-1.03) --- 1.00 (0.97-1.02) --- 1.00 (0.97-1.03) --- H.S. or equivalent 0.96 (0.49-1.90) 1.00 (0.50-1.98) 1.05 (0.52-2.13) 1.06 (0.51-2.17) Some college 0.82 (0.47-1.43) 0.91 (0.53-1.56) 1.00 (0.59-1.70) 0.97 (0.58-1.72) College degree or higher 1.03 (0.64-1.66) 1.38 (0.84-2.29) 1.45 (0.92-2.31) 1.50 (0.96-2.34) Race/ ethnicity b --- --- --- --- African-American 1.55 (1.02-2.36)* 1.32 (0.86 -2.02) 1.33 (0.82-2.16) 1.31 (0.82-2.12) Hispanic 1.72 (1.09-2.70)* 1.51 (0.97 -2.34) 1.55 (1.00 -2.38)* 1.58 (1.03-2.42)* 1.88 (1.07-3.31)* 1.80 (1.01-3.18)* 1.61 (0.85-3.06) 1.71 (0.87-3.34) 0.92 (0.60-1.42) 0.99 (0.62-1.56) 0.93 (0.56-1.54) 0.98 (0.59-1.63) 0.94 (0.81-1.10) 0.94 (0.80-1.10) 0.99 (0.84-1.17) 0.99 (0.84-1.17) 1.67 (1.01-2.74)* 1.92 (1.16-3.19)* 2.26 (1.38-3.71)** 2.28 (1.36-3.84)** 1.00 (0.99-1.00)* 1.00 (0.99-1.00)* 1.00 (0.99-1.00)* 1.14 (0.80-1.64) 1.08 (0.72-1.62) 1.09 (0.73-1.64) 0.86 (0.58-1.26) 0.80 (0.53-1.22) 0.79 (0.52-1.20) 1.33 (0.60-2.96) 1.30 (0.56-3.00) 1.27 (0.51-3.15) Other Marital birth c Freq. of religious attendance ‡ Current employment status c Household Economic Hardship Annual household income ‡ Nongovernmental financial assistance Governmental financial assistance Electricity turned off in past year Household stability d c c c --- --- --- FATHERS, NEGLECT, AND CPS INVOLVEMENT 39 Moved 1 time 0.68 (0.41-1.13) 0.70 (0.41-1.18) 0.69 (0.41-1.17) Moved 2+ times 1.25 (0.70-2.26) 1.12 (0.63-1.97) 1.31 (0.65-1.96) 1.00 (0.88-1.15) 0.98 (0.85-1.11) 0.97 (0.84-1.13) 1.49 (1.07-2.08)* 1.49 (1.07-2.07)* 0.86 (0.76-0.98)* 0.87 (0.77-0.98)* 2.18 (1.36-3.47)** 2.04 (1.22-3.42)** # of children in household < 18 years old Paternal Psychosocial Characteristics Parenting stress ‡ Involvement with the child ‡ Depression c Alcohol use e --- --- 1-3 drinks in one day in past yr 1.56 (1.05-2.28)* 1.57 (1.07-2.31)* 4+ drinks in one day in past yr 1.26 (0.62-2.53) 1.24 (0.62-2.49) Relationship With Child’s Mother And Characteristics Of The Child Parental relationship quality ‡ Co-parenting support from mother Male child 0.93 (0.75-1.16) ‡ c 0.80 (0.40-1.63) 0.70 (0.46-1.08) Child health ‡ 1.13 (0.89-1.44) Child low birth weight c 0.99 (0.49-2.00) Note: OR = odds ratio; CI = confidence interval. All models were adjusted for the father’s city at the time of the 3-year interview because this was a key variable used in the sampling design. Household income was square root transformed for analyses. ‡ Higher scores indicate higher levels of the construct. a Reference group is less than high school degree. b Reference group is White. c Dichotomous variable coded 0 = no, 1 = yes. d Reference group is no moves since the child’s first birthday. e Reference group is no alcohol consumption in the past year. *p ≤ .05, **p ≤ .01, ***p ≤ .001. FATHERS, NEGLECT, AND CPS INVOLVEMENT 40 Table 4 Logistic Regression Models of Paternal and Household Characteristics At 3 Years or Earlier Predicting Household Child Protective Services Involvement (N = 1,000) Model 1 Model 2 Model 3 Model 4 OR (CI) OR (CI) OR (CI) OR (CI) Demographic & Human Capital Characteristics Father age (in years) Father education 1.00 (0.95-1.05) a --- 0.99 (0.94-1.05) --- 1.00 (0.93-1.05) --- 0.99 (0.93-1.05) --- H.S. or equivalent 0.55 (0.30-1.01) 0.65 (0.32-1.34) 0.60 (0.29-1.28) 0.65 (0.31-1.38) Some college 0.41 (0.20-0.84)* 0.53 (0.24-1.18) 0.55 (0.25-1.22) 0.61 (0.27-1.41) College degree or higher 0.04 (0.01-0.23)*** 0.06 (0.01-0.46)** 0.06 (0.01-0.47)** 0.07 (0.01-0.53)* Race/ ethnicity b --- --- --- --- African-American 1.33 (0.63-2.82) 1.24 (0.51 -2.97) 1.49 (0.59-3.74) 1.44 (0.60-3.47) Hispanic 0.41 (0.17-0.96)* 0.42 (0.16 -1.11) 0.52 (0.18-1.48) 0.51 (0.17-1.52) Other Marital birth --c Freq. of religious attendance ‡ Current employment status c --- --- --- 1.02 (0.52-2.01) 0.97 (0.49-1.92) 1.01 (0.49-2.11) 0.98 (0.50-1.94) 0.84 (0.68-1.04) 0.82 (0.66-1.03) 0.82 (0.66-1.03) 0.84 (0.68-1.04) 1.37 (0.52-3.59) 2.07 (0.66-6.49) 2.32 (0.73-7.45) 2.20 (0.64-7.52) 1.00 (0.99-1.00) 1.00 (0.99-1.00) 1.00 (0.99-1.00) 1.49 (1.05-2.12)* 1.39 (0.98-1.98) 1.45 (0.99-2.12) 2.22 (1.09-4.48)* 2.25 (1.17-4.35)* 2.18 (1.09-4.37)* 1.01 (0.39-2.59) 0.91 (0.33-2.49) 0.88 (0.30-2.65) Household Economic Hardship Annual household income ‡ Nongovernmental financial assistance c Governmental financial assistance Electricity turned off in past year c c FATHERS, NEGLECT, AND CPS INVOLVEMENT 41 Household stability d --- --- --- Moved 1 time 0.71 (0.40-1.28) 0.74 (0.40-1.34) 0.69 (0.37-1.32) Moved 2+ times 0.65 (0.26-1.64) 0.54 (0.23-1.26) 0.51 (0.20-1.33) 1.35 (1.10-1.67)** 1.35 (1.10-1.65)** 1.38 (1.12-1.69)** Parenting stress ‡ 0.96 (0.68-1.33) 0.90 (0.60-1.36) Involvement with the child ‡ 1.27 (0.94-1.72) 1.30 (0.95-1.78) Depression c 2.21 (1.03-4.74)* 2.42 (1.05-5.58)* # of children in household < 18 years old Paternal Psychosocial Characteristics Alcohol use e --- --- 1-3 drinks in one day in past yr 0.76 (0.31-1.86) 0.75 (0.30-1.87) 4+ drinks in one day in past yr 1.38 (0.62-3.06) 1.31 (0.58-2.97) Relationship With Child’s Mother and Characteristics of the Child Parental relationship quality ‡ Co-parenting support from mother 0.84 (0.64-1.09) ‡ Male child c 2.19 (0.80-6.00) 0.67 (0.48-0.94)* Child health ‡ 0.67 (0.44-1.04) Child low birth weight c 1.19 (0.50-2.86) Note: OR = odds ratio; CI = confidence interval. All models were adjusted for the father’s city at the time of the 3-year interview because this was a key variable used in the sampling design. Household income was square root transformed for analyses. ‡ Higher scores indicate higher levels of the construct. a Reference group is less than high school degree. b Reference group is White. c Dichotomous variable coded 0 = no, 1 = yes. d Reference group is no moves since the child’s first birthday. e Reference group is no alcohol consumption in the past year. *p ≤ .05, **p ≤ .01, ***p ≤ .001.