Abstract
Plans to improve nonresident father engagement practices within child welfare systems often include enhancing worker attitudes. Yet, there is little research on the role of worker attitudes in father engagement practices. To address this, we conducted logistic regression analysis with a sample of 1,300 child welfare cases in four U.S. states in 2004–2005. We found that workers’ attitudes that involving nonresident fathers helps child well-being increased their odds of contacting a nonresident father (odds ratio [OR] = 1.43 [1.00, 2.03]), seeking health insurance (OR = 2.25 [1.26, 4.02]), but did not affect workers seeking financial support from fathers. We call for better measurement of attitudes, their contexts, and their impact.
Keywords
Introduction
Involving nonresident fathers (NRFs) in child welfare casework practice has significant implications for the well-being of children, families, and fathers (Casey Family Programs, 2024; Lauritzen et al., 2018; Lee et al., 2025; McDonald & Lindsay, 2025; Zanoni et al., 2013). Research, evaluations, and practice recommendations often point to child welfare caseworker attitudes as one important factor in the ecology that influences worker efforts and decisions to involve fathers (Allan et al., 2017; Casey Family Programs, 2024; Johnson & Briggs, 2025; Spielfogel et al., 2023; Vann & Perry, 2023). Since the federal government and states began collaborating to enhance father engagement over two decades ago, reports produced by researchers, advocates, and administrators have echoed the idea that worker attitudes and biases matter for father engagement (Agosti et al., 2021; Brooks, 2010; Campbell et al., 2015; Coady et al., 2013; Coakley et al., 2014; McDonald & Lindsey, 2025. Yet, only one dataset (analyzed in the current study)exists linking worker attitudes to outcomes at the case level, hampering our understanding of when and how workers’ positive attitudes can be leveraged to enhance father engagement. What little we know about how attitudes are associated with casework practice outcomes is based on original and secondary analysis of said dataset and a single, more recent study that relies on worker reports of their typical casework practice (Brewsaugh et al., 2018; de Peña, 2019; Malm et al., 2006). Evidence is scant and mixed, with some evidence that attitudes are not associated with efforts to identify and locate fathers and some evidence that workers with more positive attitudes are more likely to involve fathers in case planning by including them in routine practice alongside mothers and considering them as a foster placement option (Brewsaugh et al., 2018; de Peña 2019; Malm, Murray and Geen, 2006). This study extends the evidence base through secondary analysis of survey collected in 2004–2005 from child welfare workers in four U.S. states, examining whether workers are more likely to contact and subsequently involve a father when they have more positive attitudes about whether involving NRFs improves child well-being, whether NRFs want to be involved, and whether involving NRFs complicates a case.
Background
Child Welfare System Case Processing: Where Do Nonresident Fathers Fit?
Most children who are subject to a child protection services investigation and/or who enter foster care live with their biological mother at the time of the investigation (Bellamy, 2009). Although many children also live with a biological father, stepfather, mother’s partner, or another adult at the time of investigation and removal, most children have a biological father who does not reside with them, whom researchers often refer to as an NRF (Bellamy, 2009; Malm, 2002). Most NRFs maintain some contact and involvement in children’s lives up to the point of foster care entry (Bellamy, 2009). Involving NRFs in casework has been mandated by the federal government since 1980 (U.S. Department of Health & Human Services [USDHHS], Administration for Children and Families, Children’s Bureau, 2005). Doing so is associated with increased chances that a child will be returned to their mother or placed with other family during or after foster care and may enhance the well-being of children placed in care (Coakley, 2013; Dunkerley, 2021; D’Andrade, 2017; Hindt et al., 2020; Kim et al., 2020; Leon et al., 2016; Malm et al., 2008; Malm & Zielewski, 2009; Trahan et al., 2020). Thus, understanding how attitudes are associated with workers’ efforts and decisions regarding NRFs has the potential to improve child and family well-being in the context of child removal. Case work practice with NRFs occurs on a continuum, with initial steps including identifying and locating a father so that the agency can contact him. Once a father is contacted, a caseworker can begin making efforts to involve him in case planning and children’s lives. Involving a father in case planning looks like identifying his needs, working together to set goals, and facilitating access to services and supports. Involving a father in his child’s life looks like facilitating contact, closeness, support, and visitation (Baumgartner et al., 2023; Pearson & Wildfeuer, 2022).
How does a child enter foster care? In the United States, each state designs its own child welfare system. Yet, key decision points characterize the flow of children into and out of foster care in all states, and agency efforts to engage fathers intersect all stages of case processing (see Figure 1; Baumgartner et al., 2023). For a child to enter foster care, someone must first report suspected maltreatment, and the child welfare agency must decide whether the report merits investigation. When an agency concludes the investigation, the agency may or may not substantiate the report, depending on the level of evidence to support the claim (U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2024). In cases where the child would be in danger if the child remained in the home during an investigation, the agency removes the child while conducting the investigation. If, at the end of the investigation, the worker substantiates the report and assesses a high level of risk or danger, the child is removed from the home. Removal is synonymous with entry into foster care, and the agency may place the child with relatives, nonrelative foster parents, or in a group home or institutional setting (Child Welfare Information Gateway, 2020). The federal government mandates that the agency attempts to notify all known relatives, including noncustodial parents, when an agency places a child in foster care (Children’s Bureau, 2022). In addition, the federal government mandates that workers make efforts to learn whether a child’s relatives, including noncustodial parents, are interested in serving as foster parents (Children’s Bureau, 2022). In several states, noncustodial parents (typically fathers) are given priority before grandparents and other relatives are considered (Children’s Bureau, 2022). Thus, child welfare workers often begin efforts to identify, locate, and contact the child’s NRF either upon opening an investigation or by the time the child is placed in foster care.
Progression of a Child Welfare Case From Report to Reunification.
How does a child move through the child welfare system? After a child enters foster care, the child welfare agency establishes a permanency goal and the worker develops a case plan to facilitate the child’s exit from foster care. Federal law mandates that the permanency goal for all cases, except where not in the child’s best interests, be reunification, which means returning the child to their parent/caregiver (s) (Child Welfare Information Gateway, 2020). The case plan outlines activities or milestones parent/caregiver (s) must complete for the agency to deem it safe for the child to return home. For example, a mother whose child was removed due to risks associated with substance dependence or abuse may need to complete substance counseling, consistently test negative for using substances, and maintain stable housing (Child Welfare Information Gateway, 2020). Throughout the time that children are in foster care, the caseworker ideally keeps consistent contact with the parent(s) from whom the child was removed, as one of their overarching goals is supporting and monitoring parent(s) efforts to complete the case plan. Periodically, court dates are scheduled to document agency and parent efforts, assess progress with the case plan, and make major decisions around reunification. Apart from helping parent(s) work toward reunification, child welfare workers are also tasked with augmenting family resources available to children in foster care (Child Welfare Information Gateway, 2020). As such, workers should take steps to identify available financial and nonfinancial support in the family network. Perhaps most importantly, they should make efforts to maintain children’s relationships with caring adults, including parent(s), siblings, and kin (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a). While a child is in foster care, workers should also make efforts to involve the father in case planning by engaging in practices such as assessing his strengths and needs, referring him to relevant services and programs to improve his capacity to support child well-being, facilitating father-child visits, and supporting father-child relationships beyond visitation, regardless of whether he lived with the child at the time the child was removed (see Figure 2; Baumgartner et al., 2023; USDHHS, Administration for Children and Families, Children’s Bureau, 2020a).
Involving Nonresident Fathers in Case Planning After Contacting Them: Potential Practices.
How does a child exit foster care? In most cases, parent(s) complete the case plan and children are reunited once the agency and court deem that it is safe enough for this to occur. If the court deems that the parent (s) has/have not made strong enough efforts to complete the case plan despite the agency’s support, federal law mandates that the court must terminate the parent’s legal rights (Child Welfare Information Gateway, 2020). At this point, the agency will work toward another permanent placement, which they most likely have already explored and pursued as an alternative or secondary placement option through a process called concurrent planning. Thus, reunification with an NRF or relative may be the secondary case goal, even if a mother/caregiver is engaging consistently with the case planning process and meeting milestones. If the caregiver from whom the child was removed loses their parental rights, the agency should determine if it would be in the child’s best interest to be placed with their NRF or other relatives. If so, this should be prioritized over other nonrelative placements (Baumgartner et al., 2023; Children’s Bureau, 2022).
Despite years of federal and state attention to father engagement, child welfare workers in the United States tend to direct more of their focus and efforts toward mothers (Franck, 2001; O’Neal-Luna, 2021; Murguia et al., 2020; USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). Analyses of the most recent Child & Family Service Review data gathered from 2015 to 2017 illustrate this (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). When all states’ data were aggregated, researchers found that agencies made adequate efforts to involve fathers in half of the cases, whereas they made adequate efforts to engage mothers in two thirds of the cases (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). When considering how often workers adequately assessed parents’ needs, the pattern was similar. For nearly two out of three cases, workers assessed mothers’ needs, whereas fathers’ needs were assessed in under half of the cases (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). For about 60% of cases, workers provided mothers appropriate services, compared with 44% of cases for fathers (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). Furthermore, workers were less likely to assess and address single fathers’ service needs relative to single mothers’ needs (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). Workers came closest to parity in mother and father engagement on measures of efforts to promote parent-child relationships, such as encouraging and facilitating parents’ participation in school activities, sports, and medical appointments. Workers did an adequate job of this in 65% of cases for mothers, and 56% of cases for fathers (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a). When looking at individual states, some measures of father engagement dipped as low as 2% to 12% (Pearson & Wildfeuer, 2022).
When describing the barriers to involving fathers in child welfare practice, researchers have argued that gendered expectations for family roles place fathers in the periphery of practice (Baum, 2017; Brown et al., 2009; Critchley, 2022; Philip et al., 2019). From the outset of case processing, practice patterns suggest that workers view mothers as more responsible for, and relevant to, children’s well-being (Brewsaugh et al., 2018; Middel et al., 2020). For instance, case files are typically organized by mother’s name. Studies also have found that case files often do not include information on family structure or that details about NRFs are often missing (Murguia et al., 2020; O’Donnell, 1999; Sonenstein et al., 2002; Wells et al., 2015). Although barriers exist to identifying and engaging fathers, studies have found that workers tend to see mothers as the default client, even when fathers reside with their children (Perez-Vaisvidovsky et al., 2023; Skramstad & Skivenes, 2017). In this vein, workers’ perceptions of and attitudes toward fathers can themselves be a barrier to engagement. This starts at the outset of case processing. For instance, workers are more likely to open a case for services and to place children in foster care when a mother, rather than a father, is the subject of a maltreatment report (Crawford & Bradley, 2016; Kobulsky et al., 2021; Middel et al., 2020). Studies have also found that workers tend to focus services on mothers and avoid fathers, even when the main concern is fathers’ use of violence against mothers and/or children (Bedston et al., 2019; Brown et al., 2009; Nygren et al., 2021).
Far From Ideal in Practice: Involving Fathers in Child Welfare Casework
Despite years of federal and state attention to father engagement, child welfare workers in the United States tend to direct more of their focus and efforts toward mothers (Franck, 2001; O’Neal-Luna, 2021; Murguia et al., 2020; USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). Federally commissioned reports based on the most recent Child & Family Service Review data illustrate this (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). When all states’ data collected in the 2015 to 2017 round of service reviews were aggregated, researchers found that agencies made adequate efforts to involve fathers in half of the cases, whereas they made adequate efforts to engage mothers in two thirds of the cases (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). When considering how often workers adequately assessed parents’ needs, the pattern was similar. For nearly two out of three cases, workers assessed mothers’ needs, whereas fathers’ needs were assessed in under half of the cases (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). For about 60% of cases, workers provided mothers appropriate services, compared with 44% of cases for fathers (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). Furthermore, workers were less likely to assess and address single fathers’ service needs relative to single mothers’ needs (USDHHS, Administration for Children and Families, Children’s Bureau, 2020b). Workers came closest to parity in mother and father engagement on measures of efforts to promote parent-child relationships, such as encouraging and facilitating parents’ participation in school activities, sports, and medical appointments. Workers did an adequate job of this in 65% of cases for mothers, and 56% of cases for fathers (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a). When data were from individual states were compared, measures of father engagement dipped as low as 2% to 12% in some states (Pearson & Wildfeuer, 2022).
When describing the barriers to involving fathers in child welfare practice, researchers have argued that gendered expectations for family roles place fathers in the periphery of practice (Baum, 2017; Brown et al., 2009; Critchley, 2022; Philip et al., 2019; Scourfield, 2001; Scourfield, 2006). Several reviews and theoretical essays frame barriers to engaging fathers ecologically (see, for example, Gordon et al., 2012; Harty & Banman, 2023; Walsh & Graves, 2025). Among the most commonly cited barriers is a lack of system resources and supports to address fathers’ typical challenges (e.g., unemployment, criminal justice involvement, alcohol, or substance use). Workers and fathers interviewed in previous studies also describe how unconscious bias against fathers leads workers to form ideas about what fathers (do not) have to offer (Walsh & Graves, 2025). From the outset of case processing, practice patterns suggest that workers view mothers as more responsible for, and relevant to, children’s well-being (Brewsaugh et al., 2018; Middel et al., 2020). For instance, case files are typically organized by mother’s name. Multiple studies also find that case files often do not include information on family structure or that details about NRFs are missing (Murguia et al., 2020; O’Donnell, 1999; Sonenstein et al., 2002; Wells et al., 2015). Although barriers exist to identifying and engaging fathers, studies have found that workers tend to see mothers as the default client, even when fathers reside with their children (Perez-Vaisvidovsky et al., 2023; Skramstad & Skivenes, 2017). In this vein, workers’ perceptions of and attitudes toward fathers can themselves be a barrier to engagement. For instance, workers are more likely to open a case for services and to place children in foster care when a mother, rather than a father, is the subject of a maltreatment report (Crawford & Bradley, 2016; Kobulsky et al., 2021; Middel et al., 2020). Studies have also found that workers tend to focus services on mothers and avoid fathers, even when the main concern is fathers’ use of violence against mothers and/or children (Bedston et al., 2019; Brandon et al., 2009; Brandon et al., 2019; Brown et al., 2009; Nygren et al., 2021; Strega et al., 2008).
Conceptual Framework: Decision-Making Ecology
Ecological decision-making perspectives propose that child welfare workers’ characteristics and attitudes, among other factors at multiple levels, influence their decision-making (Allan et al., 2017; Baumann et al., 2014; Benbenishty et al., 2015; Mallette et al., 2018; Keddell, 2021). At multiple stages in the case planning process, despite objective case characteristics, workers’ decisions vary as a function of factors at the case-level, worker-level, organizational-level, and broader- or external-level factors (Maxwell et al., 2012; Pfitzner et al., 2017; Platt, 2012). For instance, one worker may substantiate maltreatment while another may not. Evidence from multiple countries shows that workers’ professional attitudes partly explain this variation. Particularly, workers’ attitudes may be more oriented toward risk or more risk- averse, impacting their interpretation of case characteristics during risk assessment (Davidson-Arad & Benbenishty, 2010) and their willingness and intent to intervene once risk is assessed (Keddell, 2017). Furthermore, workers may tend more toward family preservation or child removal, and positive attitudes toward removal increase the chance that a worker will decide to remove children from the home (Benbenishty et al., 2015). Applying the ecological decision-making perspective to workers’ efforts and decisions to involve NRFs, workers’ attitudes are assumed to affect their judgments, such as of the risks and benefits of involving a given father. In turn, their judgments affect their efforts toward achieving a given outcome, such as contacting and involving said father (Fluke et al., 2016).
Apart from attitudes, other caseworker characteristics influence risk assessment and subsequent decisions. For instance, workers with less experience tend to support substantiation and removal more frequently than do more tenured workers, in part due to their attitudes toward risk and removal (Fluke et al., 2016; Middel et al., 2022) In addition, workers with better training, more experience, and greater skill levels may be more confident in achieving a positive return on their efforts to engage a father, leading them to invest more in engaging fathers (Cryer-Coupet et al., 2021; Gladstone et al., 2014; Maxwell et al., 2012; Scourfield et al., 2012; Toros & Tiko, 2018).
Contextual factors matter as well—workers in different countries and states tend to have different average attitudes and they tend toward different decisions when presented with the same case scenarios (Benbenishty et al., 2015). Factors at the organizational level that may influence efforts and decisions include agency resources and infrastructure, such as well-functioning accessible databases for locating fathers, or staff dedicated to father engagement, or an organizational culture where supervisors or judges support workers and hold them accountable to identifying and engaging fathers (Malm et al., 2006; McDonald & Lindsey, 2025; Pearson & Wildfeuer, 2022). Furthermore, if a father feels valued by an agency and its staff due to a father-positive culture, a worker will likely be more successful at moving him into deeper levels of involvement appropriate to the particular case, child, and father (Bellamy et al., 2022; Cryer-Coupet et al., 2021; Leonard, 2009). Contextual factors at the external level include federal and state policies that incentivize workers’ diligent efforts to identify and engage fathers and the extent to which there are local services and programs tailored to fathers’ needs in a given community context, which is shaped in part by funding and legislation at the state and higher levels (Gordon et al., 2012; Malm et al., 2006; McDonald & Lindsey, 2025; Pearson and Wildfeuer, 2022).
Factors at the case level that may influence decisions and efforts to involve fathers include case circumstances (i.e., maltreatment type and perpetrator), family dynamics, mothers’ and fathers’ parenting strengths and barriers to involvement, as well as characteristics of children, mothers, and fathers (Gordon et al., 2012; Malm et al., 2006; Maxwell et al., 2012). For instance, some mothers may not wish to help with contacting and involving an NRF, including due to unresolved conflict or fear of losing custody to the father or his family network (Coakley et al., 2014; Malm et al., 2006; O’Donnell et al., 2005; Sicouri et al., 2018). In addition, mothers who have repartnered after dissolving their relationship with children’s fathers may consider their current partner to be the children’s father (Ferguson & Hogan, 2004; Huebner et al., 2008). Father’s paternity status (Malm et al., 2006; McDonald & Lindsey, 2025), pressures on fathers’ time due to employment (Seymour et al., 2021; Tully et al., 2018), incarceration (Pearson & Wildfeuer, 2022), and fathers’ own stereotypical beliefs about gender roles and parenting can also influence engagement outcomes from identification to contact to involvement in case planning (Philip et al., 2019; Sicouri et al., 2018). Furthermore, in a large minority of cases, workers may be concerned about exposing a mother, child, or themselves to a father’s use of violence especially in organizational contexts that lack adequate services, supports, or worker guidance (Malm et al., 2006; McDonald & Lindsey, 2025; Walsh & Graves, 2025).
Theory of planned behavior and related frameworks (i.e., reasoned action, reasonable goal pursuit) complement ecological decision-making perspectives and have informed efforts to measure worker attitudes toward father involvement since the creation of the Dakota Father Friendly Assessment (Azjen et al., 2018; Kammeyer-Meuller et al., 2024; White et al., 2011). According to theory of reasoned action, workers form behavioral intentions which then increases chances of a given decision or behavior. Within this theoretical framework, attitudes are conceptualized as positive or negative dispositions toward a given behavior that make the behavior more or less likely, respectively (White et al., 2011). Worker decisions and efforts are partly a function of attitudes toward that behavior (Azjen et al., 2018; Hale et al., 2002). Also consistent with ecological decision-making perspectives, theory of planned behavior posits that workers’ perceptions of organizational norms influence their behaviors (Azjen et al., 2018; Hale et al., 2002). In other words, workers with more positive attitudes toward a practice or policy and/or those who perceive stronger norms of enacting a practice will intend to implement that policy or practice. In turn, those who intend to act are more likely to act. Workers’ self-efficacy also shapes their intentions and actions, with workers who believe they have control over an outcome being more likely to pursue it (Kammeyer-Mueller et al., 2024). Applied to casework practice with NRFs, those who believe that involving fathers in caregiving and casework practice benefits child or family well-being, those who perceive their peers, supervisors, or organization as favoring father involvement, and those who feel more capable of successfully involving fathers will make more frequent and/or persistent efforts to engage fathers, regardless of case and father characteristics.
Literature Review
Worker Attitudes and Casework Practice With NRFs in Child Welfare
How Have Attitudes Been Measured?
Although improving worker attitudes and addressing bias are typically included among the recommendations for improving child welfare agency efforts to improve casework practice, there have only been two studies that directly measure worker attitudes toward fathers and their associations with case outcomes (Malm et al., 2006; Brewsaugh et al., 2018). Of the two studies, only Malm and colleagues (2006) focuses exclusively on NRFs. Malm and colleagues (2006) surveyed over 1,200 child welfare workers in four U.S. states as part of a broader study aimed at understanding barriers and facilitators of casework practice with NRFs. Based on an extensive literature review and interviews with child welfare administrators, Malm and colleagues (2006) developed five items to measure workers’ perceptions of involving NRFs in casework practice. For each item, workers were presented with a statement and asked the extent to which they agreed or disagreed on a five-point scale. Items measured workers’ attitudes about whether involving NRFs improves child well-being, NRFs want to be involved, NRFs need help with parenting, involving NRFs complicates a case, and working with NRFs is more trouble than it is worth. Of these five attitudes, Malm and colleagues (2006) only analyzed the first two, in part due to the skewed distribution of the remaining measures. Rather than analyzing them separately, they combined them to form a dichotomous measure of father-positive orientation, comparing workers who agreed or strongly agreed both that NRF want to be involved and that involving NRF improves child well-being to all other workers. We discuss their findings and those of a more recent secondary analysis of their data in the next section (Malm et al., 2006; de Peña, 2019).
More recently, Brewsaugh et al. (2018) surveyed 490 child welfare workers in several U.S. states about their attitudes and practices as part of her study of how benevolent and hostile sexism correlates to professional attitudes and practices with fathers. Although Brewsaugh’s study relies on self-reports from a voluntary, nonrandom sample of child welfare workers, it includes well-validated measures of attitudes toward involving fathers. To measure workers’ attitudes, Brewsaugh et al. (2018) tailored three subscales from the Dakota Father Friendly Assessment, traditionally used in early childhood settings: worker attitudes, worker bias toward practice with mothers, and worker behaviors (White et al., 2011). To adapt the scales to child welfare settings, Brewsaugh et al. (2018) relied on her own extensive experience in child welfare practice and consulted an expert panel of child welfare workers and administrators.
For each scale, Brewsaugh et al. (2018) presented workers with several statements and asked them to respond to several items on a 7-point scale. These included statements such as “my organization’s mission statement applies to fathers as well as mothers,” “fathers should be involved in the intake and assessment process,” “it is important that fathers attend case functions,” “it is important to have services for the whole family,” and “fathers bring unique strengths to parenting that meet a child’s growth and development needs.” Furthermore, Brewsaugh and colleagues (2018) also measured bias against fathers (conceptualized as preference toward working with mothers) using multiple items such as, “mothers are more committed to the care and well-being of their children than most fathers,” “mothers put more effort into completing their case/treatment plan,” “I find it hard to let fathers be in charge of providing child care,” and “I usually don’t interact with fathers when mothers are present.” The father-friendly attitudes scale was negatively associated with a preference toward working with mothers (Brewsaugh et al., 2018). Below we discuss the results of Brewsaugh et al.’s (2018) analysis.
At What Points in Case Processing Are Attitudes Associated With Efforts to Involve Fathers?
A small body of research suggests that child welfare caseworkers’ positive attitudes toward engaging fathers may enhance worker efforts and case outcomes, but perhaps only in the latter steps of the father involvement continuum (Brewsaugh et al., 2018; Coakley et al., 2018; de Peña, 2019; Malm et al., 2006). For instance, Malm and colleagues’ (2006) found that child welfare workers’ positive attitudes toward casework practice with NRFs did not affect the chances a worker would successfully identify an NRF. More specifically, the vast majority (88%) of fathers were identified and workers with father-positive attitudes were no more likely to report successfully identifying a father than those who endorsed one or neither statement (Malm et al., 2006). Workers’ use of specific strategies for locating fathers was not associated with this measure of attitudes, with one exception. Workers with positive attitudes were more likely to consult public aid records among the resources they used to locate a father, but they were not more or less likely to consult other specific resources, such as family members or child support systems (Malm et al., 2006). Recently, de Peña (2019) conducted a secondary analysis of Malm and colleagues’ (2006) data, combining all five items into a scale after reverse coding apparently negative statements. de Peña (2019) found no association between attitudes and the number of strategies used to identify and locate NRFs.
In contrast, research focused on outcomes later in the engagement process tends to find that positive attitudes are associated with efforts to involve fathers. Malm and colleagues (2006), who measured positive attitudes as believing fathers want to be involved and that involving them helps child well-being, found that attitudes did not influence whether workers told a father that their child was placed in foster care or whether they shared the case plan with the father (Malm et al., 2006). However, Malm and colleagues (2005) limited their sample for this analysis to fathers who had been contacted, and virtually all fathers (95+%) who were contacted were told about the foster placement and the vast majority were given a copy of the case plan (Malm et al., 2006). In the same study, workers with more positive attitudes were more likely to consider placing a child with their NRF (Malm et al., 2006). In addition, workers’ attitudes showed a weak, positive correlation with the number of services workers offered to an NRF, the number of services a father accepted, and the overall level of father-child involvement documented in the case file (de Peña, 2019). In the only other existing survey of child welfare workers that covers multiple U.S. states, workers who had father friendly attitudes were less likely to have a preference for working with mothers (Brewsaugh et al., 2018). Workers with more positive attitudes and those with lower preference for working with mothers also self-reported that they made more frequent efforts to involve fathers with greater consistency than did those with more negative attitudes or a stronger preference toward working with mothers. Practices that occurred more frequently among workers with more positive attitudes included making fathers feel critical to children’s lives, scheduling appointments and meetings when fathers can be available, educating mothers on the value of father involvement and developing a case plan that reflects fathers’ input, among other practices (Brewsaugh et al., 2018).
Evidence Is Sparse. But Why Is It Inconsistent?
It is possible that worker attitudes are not associated with their efforts until a father is contacted for a number of reasons, both practical and theoretical. For instance, the vast majority of fathers are identified at case opening. Some of these fathers may identify themselves to the system, requiring no effort on the worker’s behalf. Furthermore, compared with involving fathers, identifying and locating them is a more straightforward and concrete task. These early tasks may be more routinely monitored, less susceptible to worker discretion, and/or less effortful compared with the assessment and engagement entailed in considering a father as a placement option.
Low validity of measures of worker attitudes and worker efforts is another strong possibility that may explain inconsistent evidence. For instance, de Peña’s (2019) work ignores previous evidence that Malm and colleagues’ (2006) five items did not form an underlying construct (Arroyo & Peek, 2015) and de Peña’s (2019) own evidence of low scale reliability. Furthermore, Malm and colleagues (2006) asked workers to endorse all strategies and resources they used to identify and locate fathers from a list. From these responses, de Peña (2019) developed a summative scale of worker effort to identify and locate fathers. As such, workers who had success after using only one strategy or resource (i.e., asking mother for information) would be operationalized as making lesser effort than workers who used additional strategies after their first attempt was unsuccessful. Another threat to validity is social desirability bias. In studies based on Malm and colleagues’ (2006) data, social desirability may dampen the association of attitudes with outcomes, given that outcomes were measured based on caseworker’s review of case files. Yet, in Brewsaugh’s study, social desirability is likely to inflate associations given that both attitudes and casework practices were self-reported leaving the potential for workers to overreport both positive attitudes and frequency of casework practice with fathers.
The Current Study
The evidence suggests that workers’ positive attitudes may enhance casework practice with NRFs, although there are important limits to existing knowledge. This study aims to add to our understanding of how child welfare workers’ positive attitudes may enhance casework practice with NRFs. This study advances the literature by examining previously unanalyzed outcomes and re-examining whether a father was considered as a placement option. Whereas previous analyses were bivariate, the present study uses multivariate analysis to better isolate the association of attitudes with outcomes. Using unique survey data representative of children placed in foster care in 2004–2005 in four U.S. states, we examine whether child welfare workers’ positive attitudes toward NRFs were associated with five outcomes: contacting an NRF, seeking financial support as part of the case plan, and seeking health insurance as part of the case plan, arranging father-child visits, and considering a father as a placement option. We expect to find that each of the attitudes we examine is associated with each of the casework practices we examine, given strong theoretical grounding that workers’ attitudes tend to be associated with their decisions and actions.
Method
Participants and Procedures
We conduct a secondary analysis of the Study of Fathers’ Involvement in Permanency Placement Planning and Child Welfare Casework (2004–2005) survey data, collected from child welfare workers in four U.S. states selected to represent a range of practice contexts. Workers responded by telephone to questions about the case, father, mother, child, and about their own characteristics and attitudes. To increase data accuracy, the original investigators requested that the workers review the case file while responding to questions about the case. Using a two-stage sampling design, the original investigators gathered information regarding 1,928 cases from 1,222 workers in Arizona, Massachusetts, Tennessee, and Minnesota. At stage one, the investigators included all caseworkers working on at least two cases that met the study criteria: the children were in foster care for the first time, the case had been open between 3 and 36 months, and the children had a living, NRF at the time of foster care entry. At stage two, the investigators randomly selected two eligible cases per caseworker. The present study analyzes only those cases for whom there was valid data and for whom the father had been identified and located, which is a prerequisite for worker-father contact. After listwise deletion of cases with missing data on dependent variables (n = 52) and/or independent variables (n = 38), the analytic sample includes 1,300 cases for the dependent variable “contacted father.” The sample includes only 911 cases for postcontact outcomes because these could not occur for fathers who were not contacted.
The data used in this article were obtained with permission from the National Data Archive on Child Abuse and Neglect (NDACAN), Cornell University, Ithaca, NY. Data from the Study of Fathers’ Involvement in Permanency Placement Planning and Child Welfare Casework (2004–2005) were originally collected by Karin Malm, Robert T. Geen, and Timothy Triplett. Funding for the project was provided by the U.S. Department of Health and Human Services (Award Number: 100-01-0014). The collector of the original data, the funder, NDACAN, Cornell University, and their agents or employees bear no responsibility for the analyses or interpretations presented herein.
Measures
Dependent Variables
We analyzed five casework practice outcomes, all of which were dichotomous. Worker contacted father is a dichotomous variable equal to one if the worker reported one or more instances of contact with the NRF by phone, mail, or in-person since the investigation was opened. The remaining four variables could only occur upon contact. These include whether the worker sought financial support from father as part of the case plan, sought health insurance from father as part of the case plan, planned father-child visits, or considered father as a foster placement option.
Independent Variables
Caseworker Attitudes and Characteristics
Caseworker attitudes were measured in the original study by asking each worker to respond to five statements by selecting from five potential responses: strongly disagree, disagree, neither agree nor disagree, agree, and strongly agree. The statements include the following:
NRFs want to be part of the decision-making process with regard to their children
Involvement of NRFs enhances a child’s well-being
NRFs need help with their parenting skills
Dealing with NRFs makes a case more complicated
Working with NRFs is more trouble than its worth
Given previous evidence that the five statements do not represent a broader, underlying construct (Arroyo & Peek, 2015; de Peña, 2019), we considered each attitude separately. We dichotomized the attitude measures, comparing positive responses with all other responses. We assumed that agreeing or strongly agreeing that NRF help child well-being or NRFs want to be involved represent positive attitudes. We assumed that disagreeing or strongly disagreeing that NRFs need help, NRFs complicate the case, or NRFs are too troublesome represent positive attitudes. We did not include the attitudes NRFs need help with parenting and NRFs are too troublesome in multivariate analysis, given limited variation in workers’ responses.
We analyzed several caseworker characteristics, including field of degree, level of education, years of experience, work position, training experience, sex, and ethnoracial identity. Field of degree is a dichotomous variable, with social work as the reference category to which degrees in all other fields are compared. Level of education is a dichotomous indicator, with Bachelor’s degree or lower as the reference category to which higher levels are compared. Work position is a dichotomous indicator equal to one if the worker was a supervisor and zero if not. Years of experience is a dichotomous variable with 2 years of less as the reference category, compared with those with 3 or more years. Training experience includes two separate dichotomous indicators. The first is for having received training to identify, locate, or engage fathers. The second is having received training on how to refer cases to the child support system for help with locating fathers. Sex is measured with two categories, male and female. Ethnoracial identity is a categorical variable that compares non-Latinx white workers with those who identified as Latinx of any race(s), non-Latinx Black, and non-Latinx of another race.
Organizational- and Broader-Level Characteristics
We included a dichotomous indicator comparing workers who reported that their agency encourages workers to use the state parent locator services to identify and locate fathers to those who did not. State is a categorical variable. Minnesota is the reference category, compared with Arizona, Tennessee, and Massachusetts.
Case and Child Characteristics
We included a dichotomous indicator for cases in which neglect was the only alleged maltreatment compared with cases with any other form of alleged maltreatment. We included a dichotomous indicator for whether NRF was named as a perpetrator. Child’s age is a dichotomous variable with children five and younger as the reference category to which children of all other ages are compared. Sex is a dichotomous variable comparing males and females. Child’s ethnoracial identity compares non-Latinx white children with those identified as Latinx of any race, non-Latinx Black, non-Latinx Multiracial, or non-Latinx of any other race.
Data Analytic Strategy
We used logistic regression analysis to examine whether child welfare caseworkers’ positive attitudes were associated with father engagement outcomes, net of other factors. Logistic regression analysis is an extension of linear regression used to estimate the probability that a dichotomous outcome occurs given a set of independent variables. With logistic regression, a logistic or sigmoid function is used to transform the outcome so that there are only two possible values for the dependent variable: 1 (outcome occurred) and 0 (it did not). Logistic regression coefficients, then, represent the increase or decrease in the log-odds that an outcome (i.e., contacting a father) occurs given a unit increase or decrease in an independent variable. Thus, we exponentiated coefficients and presented odds ratios (ORs) using 95% confidence intervals (CIs). In all analyses, we applied survey weights to adjust for the complex sampling design. We conducted all analyses using Stata SE 17. We report associations at a marginal level of significance (p < .10) given the theoretical assumption that attitudes will be associated with outcomes and given that the sample gets increasingly small as we examine later stages of involvement.
Logistic regression rests on several assumptions which if violated could lead to inaccurate coefficient estimates and/or standard errors. Logistic regression assumes that the relationship between independent variables and the log-odds of the probability of the dependent outcome is linear (Long & Freese, 2006; UCLA Statistical Consulting Group, 2024). We tested this using STATA’s linktest command and found no violations. Logistic regression also assumes that independent variables are not strongly correlated, a phenomenon called multicollinearity (Long & Freese, 2006; UCLA Statistical Consulting Group, 2024). To test for multicollinearity, we examined variance inflation factors, which capture the extent to which correlation among independent variables is responsible for variance of an estimate, for each independent variable. All variables fell below the typically used thresholds of 5 and 10. Another assumption is that observations are independent of one another (i.e., their errors are not correlated) (Long & Freese, 2006; UCLA Statistical Consulting Group, 2024). Given that workers in the sample were asked about more than one case, we used cluster robust standard errors to account for dependence among cases assigned to the same caseworker. Finally, logistic regression assumes that no extraneous variables are included and that all relevant variables are included. This is rarely possible in practice; thus, ORs in the present study should be interpreted as evidence of an association and not evidence of a causal effect (Mood, 2010).
Results
Univariate Results
Table 1 presents descriptive statistics for the study sample. Most workers were non-Latinx white women. Nearly 80% of workers had received some form of training to identify, locate, and engage NRFs. Yet, less than a quarter had received training to refer cases to federal or state parent locator services. Still, about 60% of workers reported that their agency encouraged them to make such referrals. Less than a quarter of workers had an advanced degree or a degree in social work. Turning to case characteristics, about half of children were non-Latinx white (50%), with a sizable proportion of Latinx (22%), Black (17%), and Multiracial children (8%) and a small proportion of children of Another race (3%). Nearly two thirds (62%) of children were five and older. About half (49%) of cases involved neglect allegations with no other alleged form of maltreatment, and NRFs were named as perpetrators in around 15% of cases.
Weighted Descriptive Statistics (N = 1,327).
Table 2 presents univariate statistics for workers’ attitudes and casework practice outcomes. Most workers (70%) agreed that involving NRFs in case-planning benefits child well-being, and an even greater proportion of workers rejected the idea that involving fathers is more trouble than its worth (80%). Still, other attitudes suggest some ambivalence or trepidation among workers. The majority (83%) felt that fathers need help with their parenting, nearly two thirds (64%) felt that involving fathers makes a case complicated, and only a little over half (55%) agreed that NRFs want to be involved in case-planning. Although most workers contacted fathers and involved them in some way upon contact, their chances of reporting each individual form of postcontact involvement were lower. Among NRFs who had been located by the child welfare agency, most (74%) were contacted. In nearly three quarters of cases in which the agency contacted a father, at least one post-contact outcome occured. For just under half (45%) of cases in which a father was contacted, workers reported planning father-child visits. For nearly half (47%), workers reported that they considered the NRF as a placement option. Less than half (38%) had a case plan that included seeking financial support from the NRF. The case plan included seeking health insurance from the NRF in less than a sixth (15%) of father-contacted cases.
Weighted Univariate Statistics (N = 1,327).
Bivariate Results
Table 3 presents associations between worker attitudes and involving NRF. Each of the attitudes examined was associated with one or more outcomes, and workers who expressed positive attitudes tended to involve NRF at rates that were between 5 and 10 percentage points higher than those who did not (see Table 3). Of the five attitudes, agreeing that involving NRFs helps child well-being had the most pervasive positive association with father involvement outcomes. Workers were more likely to contact NRFs (76% vs. 68%) and upon contacting them, more likely to seek financial support (40% vs. 33%), seek health insurance (17% vs. 11%), arrange father-child visits (48% vs. 38%), and consider an NRF as a placement option (50% vs. 37%). Workers who agreed that NRFs want to be involved were more likely to contact fathers (76% vs. 70%), and upon contacting them, more likely to report planning father-child visits (48% vs. 38%). This attitude was not associated with seeking financial support or health insurance or considering a father as a placement option. None of the remaining attitudes was associated with contact, but each had a modest or weak association with one or two postcontact involvement outcomes. This tended to be at a marginal level of significance. Workers who agreed that involving NRFs complicates the case were less likely to seek financial support from fathers (36% vs. 43%). Workers who agreed that involving NRFs is more trouble than it is worth were more likely to seek health insurance (20% vs. 14%) but less likely to consider an NRF as a foster placement (39% vs. 49%). Workers who agreed that fathers need help with their parenting skills were more likely to seek health insurance (22% vs. 14%), but less likely to arrange father-child visits (37% vs. 47%).
Casework Practice Outcomes by Child Welfare Caseworkers’ Attitudes.
Multivariate Results
Table 4 presents the results of the logistic regression analysis of contact with an NRF who had been successfully identified and located. Workers who agreed that involving NRFs in case-planning improves child well-being were more likely to report contacting an NRF than those with neutral or negative attitudes (OR = 1.43, CI = [1.00, 2.03]). Attitudes that NRFs want to be involved in case-planning or that involving NRFs complicates the case were not associated with contacting a father. Workers who reported that agencies encourage them to refer cases to parent locator databases were less likely to report contacting fathers, suggesting that agencies that emphasize use of locator services may be those who have experienced greater difficulties identifying and contacting fathers. Workers with advanced degrees were less likely to report contacting a father (OR = 0.60, CI = [0.41, 0.86]). Black workers were less likely than white workers to report contacting a father (OR = 0.64, CI = [0.42, 0.99]). Other worker characteristics, including training, social work background, years of experience, and gender, were unassociated with contact. Workers were more likely to contact fathers on cases with a father named among perpetrators of maltreatment (OR = 3.15, CI = [1.72, 5.78]), aligned with the goal of assessing risk and ensuring child safety during child maltreatment investigations. Workers were less likely to contact fathers of Black and Latinx children and girls, which has been found in previous research examining the same data set as the present study (Arroyo et al., 2019; Malm et al., 2006). Maltreatment type and low-income status were unassociated with contact.
Logistic Regression Model for Agency Contacting NRF.
Table 5 presents the results of logistic regression analysis of seeking financial support from NRFs who had been contacted. Workers’ attitudes were not associated with this outcome. Workers who were trained to refer cases to parent locator databases were more likely to include financial support from the father in the case plan (OR = 2.61, CI = [1.66, 4.09]) as were those who worked in agencies that encouraged use of parent locator databases (OR = 1.53, CI = [0.98, 2.38]). Workers were less likely to seek financial support from fathers in Massachusetts (OR = 0.26, CI = [0.14, 0.50]) and more likely to do so in Tennessee (OR = 1.97, CI = [1.10, 3.53]) relative to Minnesota. Case characteristics were unassociated with seeking financial support, except that workers were less likely to include financial support in case plans for Black children (OR = 0.45, CI = [0.27, 0.76]) and Latinx children (OR = 0.64, CI = [0.38, 1.08]) relative to white children. Other worker characteristics were unassociated with this outcome, perhaps suggesting that contextual factors, rather than worker discretion, predominantly shape whether financial support is included in the case plan.
Logistic Regression Model for Agency Seeking Financial Support From NRF.
Table 6 presents the results of the logistic regression of seeking health insurance from NRFs who had been contacted. Workers who believed that involving NRF helps child well-being were more likely to report seeking health insurance from a child’s father as part of the case plan (OR = 2.25, CI = [1.26, 4.02]). Workers who were trained to refer cases to parent locator databases were more likely to include health insurance in the case plan (OR = 2.63, CI = [1.59, 4.37]), which makes sense given that successful referrals could lead to better information about active child support cases and father’s employment and insurance status. Black workers (OR = 1.92, CI = [1.02, 3.60]) had greater odds than did white workers of including health insurance. Potentially, Black workers may be more likely to work in agencies which emphasized this particular practice, as worker race did not pattern other outcomes. Other worker attitudes and characteristics were unassociated with including health insurance in the case plan. Children’s race and gender were unassociated with seeking health insurance, but workers were more likely to seek health insurance from fathers of older children (OR = 1.75, CI = [1.10, 2.77]). Perhaps this is because children who enter foster care for the first time when they are older tend to come from more advantaged families relative to those who enter before age five. Two case characteristics, both of which proxy a family’s low-income status, were associated with seeking health insurance from a father. Workers were less likely to seek health insurance from fathers when the only maltreatment allegation was neglect (OR = 0.58, CI = [0.36, 0.94]) and more likely to do so when a case was ineligible for Title IV-E reimbursement (OR = 1.61, CI = [0.92, 2.82]). This likely reflects the realistic assessment that less advantaged fathers do not have means to provide health insurance through their work, let alone if they are unstably employed or incarcerated.
Logistic Regression Model for Agency Seeking Health Insurance From NRF.
Table 7 presents results of logistic regression analysis of arranging father-child visits for fathers who had been contacted. No attitude was associated with arranging visits, and only worker training, state in which case was located, and children’s race were associated with workers arranging visits, suggesting this outcome may be more responsive to fathers’ circumstances (and workers judgments thereof) and contextual factors than to worker attitudes. Workers who were training to identify and engage fathers were more likely to arrange visits (OR = 1.43, CI = [0.98, 2.08]). Workers in Massachusetts were more likely to arrange visits than those in Minnesota (OR = 1.75, CI = [1.04, 0.93]). Workers were less likely to arrange visits when working with Multiracial children (OR = 0.49, CI = [0.34, 1.08]) and those of Another Race (OR = 0.60, CI = [0.34, 1.08]).
Logistic Regression Model for Agency Planned Visits.
Table 8 presents results of logistic regression analysis of considering an NRF as a placement option upon contacting him. Workers who believed that involving NRFs helps child well-being had greater odds of considering a father as a placement option (OR = 1.66, CI = [1.14, 2.41]). Attitudes about how much NRFs want to be involved or how involving an NRF complicates a case were not associated with considering a father as a placement option. Workers who had been trained to use parent locator services were more likely to consider a father (1.55, CI = [1.02, 2.37]). Workers with more years of experience were less likely to consider a father, raising the question as to how tenure in the child welfare workforce shapes father-related attitudes and skills over time. Other worker characteristics were not associated with this outcome.
Logistic Regression Model for Considering Father as Placement.
Discussion
This study provides new evidence that child welfare caseworkers’ attitudes are associated with the chances that an NRF will be contacted and involved beyond contact. When workers agreed that involving fathers helps child well-being, they were more likely to contact fathers. Focusing only on fathers who were successfully contacted, attitudes were associated with some outcomes (i.e., seeking health insurance and considering father as a placement option), but not others (i.e., seeking financial support and arranging father-child visits). This study also found that other attitudes, around fathers’ desire for involvement and the extent to which involving fathers complicates casework, were not associated with workers contacting and involving fathers. These results suggest that attitudes may impact worker-father contact, an early-stage caseworker practice that opens the door to deeper father engagement. Apart from workers’ attitudes, training had the most consistent impact on the outcomes we examined. Other characteristics of children, cases, workers, and organizations were not consistently associated with outcomes, suggesting that a specific set of factors, and workers’ judgments thereof, impacts each of the various potential father engagement outcomes that exist, including the limited set we examined.
We propose a few potential explanations for our mixed findings. Workers’ efforts to contact fathers may be one of the steps on the father engagement continuum that is most susceptible to worker attitudes. Latter steps may be more susceptible to other factors in the decision-making ecology, including fathers’ characteristics which are often learned or confirmed upon contact (Malm et al., 2006). This also includes organizational culture, which may influence steps that depend more on a workers’ interpretation of whether that step is suited to the particulars of a father, family, and case. Furthermore, once a father is contacted, worker factors such as their demeanor and skills may become more important to whether actions are taken than attitudes about fathers are. Existing research documents that fathers’ perceptions of worker attitudes and skills shape fathers’ motivation and receptiveness (Cryer-Coupet et al., 2021). In turn, workers’ perceptions of the utility and payoff of efforts may shape their efforts to involve that father in the case. Even more likely, a fathers circumstances and barriers as well as workers’ interpretation of his demeanor may become more important to whether a worker judges a step, such as seeking insurance or financial support, worthwhile for the case and even, positive and safe for the child and mother.
We turn now to a particularly puzzling finding. Workers who believed involving NRFs is good for child well-being were more likely to report seeking health insurance, but not financial support, from the NRF as part of the case plan. At the time the data were collected, the federal government had only recently begun to provide resources and technical support to state agencies to enhance collaboration between child protection and child support systems (Malm et al., 2006). In fact, data analyzed in the current study are sourced from a larger project whose aims included understanding how federal parent locator services could be better leveraged by states to increase rates of establishing paternity and child support orders for children in foster care, especially those receiving TANF and Medicaid (Malm et al., 2006; Sciamanna, 2011). In the decades leading up to the study, the federal government had also pursued recuperating Medicaid expenses, including among foster children, through insisting that states seek financial and medical support from noncustodial parents of children receiving Medicaid, but the strongest requirement was not implemented until 2005 after data were collected (Pirog & Ziol-Guest, 2006). Thus, it is possible that, at the time of the study, agency-level factors such as infrastructure, emphasis, and accountability varied more around seeking financial support than around health insurance. If so, this would have allowed workers less room for discretion based on attitudes when taking actions to seek financial support relative to those to seek health insurance. Importantly, it is difficult to disentangle the influence of individual attitudes from that of organizational climate and likely that workers’ positive attitudes grew in contexts where father engagement was encouraged and supported and/or where sufficient time to take additional steps above and beyond routine practice existed and was harnessed by those with positive attitudes (Walsh & Graves, 2025). If contacting fathers and seeking financial support were more routine or bare minimum, these outcomes would be less affected by this variation. Alternatively, workers with positive attitudes about the role of NRFs in child well-being may have viewed seeking financial support differently than they viewed contacting fathers and seeking health insurance. For example, some may have viewed collaborating with the child support system negatively or ambivalently, given their knowledge of hardships of low-income fathers and barriers to father-child involvement that often result from child support enforcement (Battle, 2019; Coakley et al., 2015; Miller & Mincy, 2012; National Child Support Enforcement Association, 2020; Robbins et al., 2022; Turner & Waller, 2017). If so, positive attitudes would not lead them to take this step. These possibilities lead us to suggest multiple questions for future research. Which steps along the continuum of father engagement vary more as a function of state and organizational context and which steps vary more as a function of workers’ attitudes and efforts? How do attitudes develop over time and how do contextual factors influence workers attitudes not only toward fathers but toward the risk and value of completing specific engagement outcomes, such as seeking financial support, arranging visitation, and pursuing reunification with fathers and their relatives?
Limitations
Our results document a positive association between attitudes and father engagement outcomes, yet the reader should not interpret results as causal for several reasons. First, the data are cross-sectional in nature. Ecological decision-making theories and theory of reasoned goal pursuit emphasize that workers attitudes shift over time and are responsive to their previous successes and failures. Thus, it is likely that we have only captured a cross-section of an unfolding feedback loop whereby attitudes shape outcomes, which in turn shape attitudes over time (Campbell et al., 2015; Kammeyer-Mueller et al., 2024). Longitudinal studies would be better suited to capture the full complexity of this dynamic process. Second, we were unable to include all case, father, agency, and broader characteristics that influence outcomes. It is possible that one or more of these factors may moderate or explain the relationships we found. Third, it is possible that workers may have reported more positive attitudes toward fathers due to social desirability bias, especially in contexts where they perceived others to have more positive attitudes. If this is the case, positive associations found here may partly capture workers’ perceptions of organizational norms (as in theory of planned behavior) and null relationships may be the result of a portion of workers’ true attitudes being masked by positive ones.
The associations found here may not generalize to all children who come into contact with child welfare systems today as the sample only included cases in which children entered foster care for the first-time and stayed there longer than 3 months. Among children who are re-entering foster care, case files may include substantial information on fathers’ past circumstances, struggles, and/or strengths. It is not clear how workers’ attitudes affect their interpretation of information about fathers gathered in previous cases and their subsequent involvement efforts. This study cannot speak to how attitudes impact casework practice in the initial stages of investigating maltreatment reports or among families whose cases are diverted onto alternative service pathways (an increasingly common practice) or served in home. This is important given recent evidence that responses during investigation may depend in part on resident parents’ gender and race and workers’ endorsement of stereotypes (Middel et al., 2020, 2022, 2024). It is also important given that attitudes toward fathers may impact workers’ recommendations to place children with NRFs and their families formally or informally in cases where removing a child from a mother’s custody is deemed necessary as a result of an investigation.
The study was also fielded over two decades ago in only four states. Although the states sampled represented a range of father engagement practice and policy contexts at the time the study was fielded, increasing state and federal attention to father engagement in child welfare has likely resulted in a greater proportion of U.S. child welfare workers experiencing substantial training, resources, support, and accountability around father engagement (Children’s Bureau, 2008; Children’s Bureau 2012). Yet, a recent report shows that only a minority of states have a fatherhood initiative and father engagement still lags behind mother engagement across states (Pearson & Wildfeuer, 2022; USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). In contexts where expectations are high, routines are clear, or father-related resources are plentiful, this may enhance worker attitudes and/or constrain the effect of attitudes (by reducing worker discretion). Greater geographic variation in any of these factors may increase the impacts of organizational and state variation on case outcomes and on the relationship between attitudes and outcomes. We must also note that our measures of casework practice are limited. For instance, we measure whether contact occurred at all. Better measures, potentially drawn from secondary analysis of Child & Family Service Reviews, would gauge the extent and quality of caseworker-father contact as frequent, quality contact can facilitate deeper engagement (USDHHS, Administration for Children and Families, Children’s Bureau, 2020a, 2020b). In addition, we were unable to include a more holistic measure of casework practice and did not directly examine father engagement, in the sense of father’s full participation in decision-making around the case. Despite these limitations, this study is among the few multivariate, case-level analyses to document the relationship between positive worker attitudes and case outcomes at multiple stages of the father engagement continuum.
Directions for Future Research
This study provides new evidence that workers’ positive attitudes are associated with multiple father engagement outcomes and suggests that workers’ attitudes shape their judgments of, and actions toward, a child’s NRF before and upon interacting with him. Yet, important questions remain. Among these questions is how to best measure worker attitudes and design research capable of examining their impacts. Qualitative and mixed methods research could support the development of reliable instruments that potentially capture multiple dimensions of attitudes. Furthermore, workers’ attitudes about complex and diverse family structures may impact their approach to practice with NRFs (Mallette et al., 2018; Saleh, 2013). Although the present study focused on NRFs, many children involved in child welfare systems also have a stepfather or adult male in their home, and a substantial minority of children reside with their biological father (Bellamy, 2009). The diversity of family structure continues to evolve in the United States since these data were collected. For example, among parents with two or more children, about 20% experience multiple partner fertility, and among children of low-income backgrounds, as many as 60% can expect to have a half sibling (Cooper & Pugh, 2020; Monte, 2019). Although it is likely that attitudes toward working with biological and stepfathers who reside with their children shape casework practice, no research has examined this.
In addition, research on attitudes toward father involvement among child welfare has relied exclusively on direct, quantitative, self-report measures. Future research could incorporate measures to gauge social desirability, as has been done with research on caseworkers’ attitudes about kinship care (Brisebois et al., 2013). In addition, mixed methods designs that present case vignettes can be used to capture if and how workers’ attitudes differentially affect cases depending on father or case characteristics. Studies of how workers’ perceptions of parents’ race and gender affect decision-making have become more numerous in recent years, but tend to focus on front-end case processing, such as decisions to screen in maltreatment reports, removing a child upon substantiating a report, and level of contact with parents during maltreatment investigations (Kobulsky et al., 2021; Middel et al., 2020, 2022, 2024). Future research should build on these designs to help move us beyond associations and understand how attitudes play out in later-stage practice.
A related direction for the field is to improve our understanding of how state and agency contexts shape workers’ attitudes and/or moderate their effects on father engagement outcomes. For example, in states with a broader fatherhood initiative, workers may have better resources and supports. In these contexts, workers may be more likely to have positive attitudes and better skills (i.e., building trust, creating a supportive environment, consistently and clearly communicating), and in turn, higher levels of engagement (Cryer-Coupet et al., 2021; Lehtme & Toros, 2020; Pearson & Fagan, 2019; Saleh, 2013). A similar process may occur at the agency level. In agencies which have staff dedicated to finding fathers or where there is more consistent, thorough training on navigating the complexities of engaging NRFs, workers may develop more positive attitudes toward working with fathers (Ewart-Boyle et al., 2015; Huebner et al., 2008; O’Hagan, 1997; Zanoni et al., 2013). In addition, in “father friendly” contexts where workers are held accountable by their supervisors, routinely asked about fathers in court, individual worker attitudes may have a weaker effect on engagement outcomes as there is less room for worker discretion (Edwards, 2009). It is important for future studies to differentiate between attitudes at the agency-level (i.e., organizational culture) and worker-level (i.e., individual attitudes), so that undue focus is not placed on changing workers’ attitudes rather than changing workers’ contexts.
Implications and Recommendations for Practice and Social Work Education
Given that positive worker attitudes are associated with NRF engagement outcomes, states and agencies should take steps to sustain positive attitudes and improve negative and ambivalent attitudes. As part of this, agencies should ensure that policies and practices consistently communicate to workers and families that fathers’ positive contributions to child and family well-being are valued (Spielfogel et al., 2023). This should occur as part of a broader change effort that ensures that workers have sufficient skills and resources for working effectively with the diverse needs of fathers. The Fathers and Continuous Learning in Child Welfare project is one recent and well-documented example of this process, with several resources freely available for child welfare agency use (Agosti et al., 2021; Bellamy et al., 2025; Fung 2023). As part of this, agencies should hold space for periodic reflection and dialogue around workers’ implicit biases and larger societal misconceptions about fathers (e.g., see “Self-Awareness and Work with Men Exercise” in Georgia State University School of Social Work, DCFS Professional Excellence Project, 2009).
It is also important to consider the ways preservice social workers are trained (or not trained) to engage fathers and how fathers are framed in teaching materials when fathers are in the curriculum (Brewsaugh & Strozier, 2016). Research has found that child welfare textbooks tend to leave fathers out of case vignettes and/or portray them negatively, making it likely that program graduates lack attitudes, knowledge and skills needed to engage them confidently (Brewsaugh & Strozier, 2016; Jiang et al., 2018). Efficacy is a cognitive precursor to implementation/practice of specific skills (Turner et al., 2011), and father engagement training has been shown to increase worker self-efficacy among child welfare workers in some contexts (Scourfield et al., 2012). Developing practitioners’ sense of efficacy, through the provision of father-focused education and training, may be imperative for effectively engaging fathers in practice (Cryer-Coupet et al., 2021; Saleh, 2013; Turner et al., 2011). Given that fathers of color may be more likely to be excluded from case planning due to worker bias and systemic racial inequalities, training and education should be attentive to how fathers’ race and other social identities shape their experiences of fathering and child welfare (Arroyo et al., 2019; Casey Family Programs, 2024).
Conclusion
Although not all NRFs want to be or can be safely involved with children, many desire to be involved or are already involved at high levels (Cameron et al., 2014; Malm et al., 2006). Engaging NRFs is aligned with social work ethics and the child welfare systems’ philosophy of strengthening families (Shadik & Shafer, 2022; Shafer & Bellamy, 2016). This article provides new evidence that workers who believe that involving NRFs enhances child well-being are more likely to contact and engage fathers in case planning and outlines new questions for the field. Children, fathers, and families stand to gain from efforts to examine the relationships among attitudes and father engagement outcomes using methods better suited to understanding how attitudes unfold in context and whether they have a causal impact on engagement outcomes, independent of contextual factors and fathers’ circumstances and characteristics.
Footnotes
Disposition editor: Cristina Mogro-Wilson
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
