Maltreatment definitions and challenges

The challenge of defining and operationalizing child maltreatment has been noted for more than a century (see Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993 for a comprehensive review.) One of the first formal definitions, “battered child syndrome,” reflected diagnostic terminology grounded in the medical model of viewing maltreatment as a form of pathology (Kempe et al., Reference Kempe, Silverman, Steele, Droegemueller and Silver1962). Subsequent legislation in the Child Abuse Prevention and Treatment Act (CAPTA) of 1974 (P.L. 93-247), and early recognition of the negative consequences of abuse and neglect were followed by federal legislation to prevent maltreatment of children (Nelson, Reference Nelson1984). These efforts spurred increasing discussions over the subsequent decades calling for clearer operational definitions for individuals and systems involved in maltreatment-related research, social policy, legislation, surveillance, substantiation, and case management (Hutchison, Reference Hutchison1990; Korbin, Reference Korbin2022; Slep & Heyman, Reference Slep and Heyman2006). Currently, the conversation continues to emphasize how assessment, intervention, and prevention of maltreatment remain impeded by the lack of a universal consensus on how to define and operationalize it (Laajasalo et al., Reference Laajasalo, Cowley, Otterman, Lamela, Rodrigues, Jud and Nurmatov2023). Clearer definitions are needed on an individual level for identifying potential maltreatment and deterring its consequences and on a societal level for analyzing prevalence and incidence and comparing across jurisdictions and over time. While standardized technical definitions of child maltreatment are needed for research and legislation on its causes and consequences, clearer and non-technical definitions are also needed for community providers, child welfare workers, and survivors (Havlicek & Courtney, Reference Havlicek and Courtney2016; Slep & Heyman, Reference Slep and Heyman2006).

Current definitions continue to fall short in many ways. Despite long-standing calls for more culturally relevant definitions of maltreatment, the vast majority of research on child maltreatment, as well as policies and prevention efforts to reduce it, rely on Western-oriented definitions and assumptions of what constitutes harm to children (Havlicek & Courtney, Reference Havlicek and Courtney2016; Korbin, Reference Korbin2022; Raman & Hodes, Reference Raman and Hodes2012; Roberts, Reference Roberts2014). Furthermore, most definitions utilized by the legal system both historically and recently have relied on narrow definitions of serious existing or potential harm to the child resulting from alleged or observed caregiver behavior. Caregiver behavior is often judged without sufficient consideration of the context of the behavior or the circumstances outside of the caregivers’ control that may have influenced the behavior, such as poverty, social isolation, and other unmet basic needs, such as homelessness and food insecurity (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Hutchison, Reference Hutchison1990; Korbin, Reference Korbin2022; Raman & Hodes, Reference Raman and Hodes2012; Rebbe, Reference Rebbe2018). When there is ambiguity, decisions are often left up to the best judgement of professionals and systems who often, despite best intentions, impose outsider perspectives that lack adequate consideration of the values, circumstances, and lived experiences of the parents and youth most affected by the decisions (Havlicek & Courtney, Reference Havlicek and Courtney2016; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022).

Moreover, few definitions of maltreatment used in legal, policy, practice, or research settings have been created or tested in collaboration with community members affected by these definitions, such as adults with childhood maltreatment and members of non-Western cultural groups who may have different child-rearing practices and standards for what is considered a deviation in culturally appropriate child-rearing (APA, 2023a; Korbin, Reference Korbin2022; Laajasalo et al., Reference Laajasalo, Cowley, Otterman, Lamela, Rodrigues, Jud and Nurmatov2023; Raman & Hodes, Reference Raman and Hodes2012). Despite widespread attempts at “differential response” (the practice of serving families of low-to moderate-risk for child maltreatment through family engagement, diversion from formal CPS investigations, and provision of services) and other similar voluntary pathways, parents involved in the child welfare system still largely perceive it as punitive, traumatizing, and racially biased (Cénat et al., Reference Cénat, McIntee, Mukunzi and Noorishad2021; Kokaliari et al., Reference Kokaliari, Roy and Taylor2019; Merritt, Reference Merritt2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Legal definitions of maltreatment also rarely take the child’s perspective into account (Laajasalo et al., Reference Laajasalo, Cowley, Otterman, Lamela, Rodrigues, Jud and Nurmatov2023), yet court records are often considered the best available evidence that maltreatment took place (Danese & Widom, Reference Danese and Widom2020; Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023).

The World Health Organization (WHO) has called for more consistency in defining maltreatment. Their working definition of maltreatment is broad, including: “all types of physical and/or emotional ill-treatment, sexual abuse, neglect, negligence and commercial or other exploitation, which results in actual or potential harm to the child’s health, survival, development, or dignity in the context of a relationship of responsibility, trust, or power” (WHO, 2023). For the purposes of this paper, maltreatment most typically refers to an individual’s childhood experiences of abuse, such as emotional, physical, and sexual abuse; or neglect, including emotional and physical neglect [Centers for Disease Control and Prevent (CDC), 2023]. We acknowledge, however, that this definition of maltreatment is not exhaustive nor adequately culturally sensitive. It depends on how “harm” is defined and operationalized across cultures, and how individuals integrate their traumatic life experiences, which often consolidate across development (Poletti et al., Reference Poletti, Gebhardt and Raballo2022). We return to these issues in the concluding sections.

Theoretical overview of the developmental psychopathology (DP) perspective

Historical research on child maltreatment and the developmental processes involved in continuity versus discontinuity of maltreatment across the lifespan and over generations have been closely intertwined with the emergence of the DP perspective. Some of the earliest DP-oriented research questions addressed the extent to which maltreated children showed maladaptive versus resilient functioning, such as positive socioemotional, cognitive, and academic outcomes (Cicchetti, Reference Cicchetti1984; Reference Cicchetti, Cicchetti and Carlson1989; Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Egeland et al., Reference Egeland, Carlson and Sroufe1993). Indeed, central DP concepts of multifinality and equifinality were originally applied to pathways of childhood maltreatment (Cicchetti & Rogosch, Reference Cicchetti and Rogosch1996). These concepts illustrated that many individuals exposed to similar adversities, such as maltreatment, showed diverging developmental pathways (multifinality), including pathways of resilient functioning. Alternatively, many individuals exposed to different early adversities (e.g., maltreatment versus domestic violence) showed converging outcomes, leading to similar psychopathology symptoms or violent behavior (equifinality; Cicchetti, Reference Cicchetti2016; Egeland et al., Reference Egeland, Carlson and Sroufe1993; Masten, Reference Masten2006; Narayan et al., Reference Narayan, Englund and Egeland2013).

One of the most important and enduring contributions of the DP perspective to the field of child maltreatment research has been the infusion of this theoretical perspective into a better understanding of why maltreatment may occur, for whom the risks are the highest (e.g., for parents as potential perpetrators and children as potential victims), and how to operationally define maltreatment. Some of the earliest scholarly progress on initial definitions of child maltreatment (which, as mentioned, continue to remain elusive today) stemmed from incorporation of the ecological-transactional model of maltreatment (Belsky, Reference Belsky1980; Cicchetti & Rizley, Reference Cicchetti, Rizley, Rizley and Cicchetti1981), which in turn was influential for delineating the DP perspective itself (Cicchetti, Reference Cicchetti, Cicchetti and Carlson1989; Masten, Reference Masten2006). According to the ecological-transactional model of childhood maltreatment and the subsequent systems principle of DP, the risks for and outcomes following maltreatment, as well as the range of experiences that may reflect maltreatment, depend on several factors. These factors may include the specific needs and capacities of the developing child, issues influencing the parent and the parent-child dyad, and the nature of the context surrounding the family, including the resources available to them and the vulnerabilities and stressors they face (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Belsky, Reference Belsky1980; Cicchetti & Toth, Reference Cicchetti and Toth2009; Masten, Reference Masten2006). Indeed, the DP perspective has always emphasized that development is the result of transactional interactions between individuals, their relationships, and their broader ecological systems (Cicchetti, Reference Cicchetti, Cicchetti and Carlson1989, Reference Cicchetti2016; Masten, Reference Masten2006; Narayan et al., Reference Narayan, Lieberman and Masten2021). In this paper, we argue that just as the DP perspective incorporates many transactional influences, our approach to assessing and understanding maltreatment and its consequences must also involve many influences. These influences have included and should continue to include rigorous multi-method, multi-informant empirical strategies; legal definitions and standards; environmental risk factors affecting families, and communities, and organizations; and children’s developmental characteristics (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993). We must also consider societal norms and expectations, such as DEI-informed perspectives that assess maltreatment in ways that are as free of bias as possible.

The legacy of prospective longitudinal studies on childhood maltreatment

Child maltreatment research over the past 40 years has become increasingly rigorous and nuanced (Cicchetti, Reference Cicchetti1984; Reference Cicchetti2016). Some of the first investigators of the enduring, predictive significance of child maltreatment in the early years realized the critical importance of documenting maltreatment prospectively, at the time it occurred in development, to observe its effects over the life course (Sroufe et al., Reference Sroufe, Egeland, Carlson and Collins2005). Indeed, one of the longest-running prospective studies has shown that the offspring of parents with childhood maltreatment histories were significantly more likely to report experiencing sexual abuse and neglect compared to offspring from a matched sample of parents who did not have childhood maltreatment histories (Widom et al., Reference Widom, Czaja and DuMont2015).

Several other prospective studies have illuminated the extremely harmful long-term effects of childhood maltreatment, including but not limited to the Dunedin Multidisciplinary Health and Development Study (Poulton et al., Reference Poulton, Moffitt and Silva2015), the Environmental Risk (E-Risk) Study (Moffitt et al., Reference Moffitt2002), the Minnesota Longitudinal Study of Risk and Adaptation (MLSRA; Sroufe et al., Reference Sroufe, Egeland, Carlson and Collins2005) and the Widom sample (Widom, Reference Widom1989). Foremost, these studies have immeasurably shaped the field by underscoring that maltreatment is not only harmful but also prevalent (issues that were themselves under debate until the last few decades of the 20th century; Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Billingsley & Giovannoni, Reference Billingsley and Giovannoni1972; Kempe et al., Reference Kempe, Silverman, Steele, Droegemueller and Silver1962). These longitudinal studies and others documented the pernicious effects of child maltreatment with compelling data from independent sources that were not contingent on individuals’ self-reports and recollections. In other words, many of these studies showed that maltreatment, assessed from court and child welfare records and other prospective sources (e.g., home visits, interviews with parents, reviews of medical records) and documented at the time maltreatment occurred in individuals’ childhoods, predicted many negative life outcomes. These outcomes included higher levels of mental and physical health problems, lower educational and vocational attainment, and more negative parenting in the next generation, after accounting for other developmental influences (e.g., Moffitt et al., Reference Moffitt2002; Raby et al., Reference Raby, Labella, Martin, Carlson and Roisman2017; Widom, Reference Widom1989). These studies emphasized the deleterious effects of childhood maltreatment during a time when society did not widely accept that children could experience trauma, much less that child abuse and neglect were common (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993). Because of these studies’ contributions, we are now well positioned to revisit and redefine how we assess and interpret maltreatment data across formats and sources.

The debate about racism in the child welfare system

On one side of the debate (which we call Side 1 only for clarity and simplicity here), scholars articulate the disproportionality and overrepresentation of minoritized children, and particularly Black children, at every decision-making stage in the child welfare system from being reported, having reports investigated and substantiated, moving to out-of-home placements, entering foster care, having lower rates of reunification with biological families, and exiting more slowly from foster care to permanent adoptions. This side argues that racial biases have been documented in school administrators, medical providers, caseworkers, and judges who have differentially higher rates of making CPS reports, advocating for removal, and ruling for premature termination of parental rights in Black families compared to White families (Cénat et al., Reference Cénat, McIntee, Mukunzi and Noorishad2021; Detlaff et al., Reference Detlaff, Boyd, Merritt, Plummer and Simon2021; Doyle, Reference Doyle2007; Lane et al., Reference Lane, Rubin, Monteith and Christian2002; Rivaux et al., Reference Rivaux, James, Wittenstrom, Baumann, Sheets, Henry and Jeffries2008). This side also points to the overrepresentation of American Indian/Alaskan Native (AIAN) children in the child welfare system and to a lesser extent, Latinx children, compared to White children (Hanna, Reference Hanna and Detlaff2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). They argue that the disproportionalities of minoritized children in the child welfare system exceed the proportions that would be expected based on the representation of children from these groups in the general populations of the United States, as well as Canada and England (Berkman et al., Reference Berkman, Brown, Scott and Adiele2022; Cénat et al., Reference Cénat, McIntee, Mukunzi and Noorishad2021; Roberts, Reference Roberts2014).

The other side of the debate (Side 2) acknowledges that Black children do indeed have disproportionately higher rates of contact with child welfare and higher rates of case substantiation and foster care placement. However, this side argues that Black children experience higher rates of poverty and sociodemographic and health-related risk factors (e.g., lower maternal age and educational attainment, higher rates of paternal absence, less usage of prenatal care, and higher birth abnormalities). Proponents of this side argue that when these factors are controlled in large national datasets, Black children are statistically no more likely (and may even be less likely) to be reported to CPS than White children (Barth et al., Reference Barth, Berrick, Garcia, Drake, Jonson-Reid, Gyourko and Greeson2021; Drake et al., Reference Drake, Jones, Chen, Font, Putnam-Hornstein, Barth and Jonson-Reid2023; Putnam-Hornstein et al., Reference Putnam-Hornstein, Needell, King and Johnson-Motoyama2013). Side 2 further asserts that because poverty rather than race is the main correlate of CPS surveillance and allegations of maltreatment, new social policies should focus on reducing socioeconomic disparities that render Black children more vulnerable to entering child welfare in the first place (Barth et al., Reference Barth, Berrick, Garcia, Drake, Jonson-Reid, Gyourko and Greeson2021; Drake et al., Reference Drake, Jones, Chen, Font, Putnam-Hornstein, Barth and Jonson-Reid2023).

In response, Side 1 argues that racism continues to exist even when race and SES factors are experimentally or statistically controlled, and such statistical models fail to acknowledge that race and SES often interact with each other in multidimensional ways to predict higher rates of child welfare surveillance and involvement (Detlaff et al., Reference Detlaff, Weber, Pendleton, Boyd, Bettencourt and Burton2020). Side 1 also notes that many minoritized groups, such as Native American families, have not been represented in large-enough datasets to quantitatively examine disproportionalities with adequate statistical power that compares them to other groups (Hanna, Reference Hanna and Detlaff2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Studies that conclude that the role of race in the disproportionality of minoritized children in child welfare is statistically non-significant ignore that poverty itself is racialized and stems from centuries-old systemic oppression based on skin color (Detlaff et al., Reference Detlaff, Boyd, Merritt, Plummer and Simon2021; Merritt, Reference Merritt2021). This systemic racism dates back to the enslavement of African people and to the forced separation of Native American children into Indian boarding schools as a result of the Indian Relocation Act, both of which continue to influence the present disproportionalities in the child welfare system (Briggs et al., Reference Briggs, Hanson, Klika, LeBlanc, Maddux, Merritt, Palusci, Panlilio, Roygardner, Schelbe, Stormer, Valentino, Vaughan-Eden and Barboza2023; Hanna, Reference Hanna and Detlaff2021).

Evidence of racial bias in child welfare is evident in Lau and colleagues’ (Reference Lau, McCabe, Yeh, Garland, Hough and Landsverk2003) quantitative study conducted over two decades ago that involved interviews with over 1,000 youth from various racial/ethnic groups (White, African American, Hispanic American, Asian/Pacific Islanders) and their primary caregivers sampled from five different service sectors, including child welfare. The authors prioritized youth interviews because child welfare investigations typically do not focus on youth reports of their own experiences despite that these reports are a better predictor of psychopathology than maltreatment ratings from social workers, CPS, and parents (Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023). Lau and colleagues’ (Reference Lau, McCabe, Yeh, Garland, Hough and Landsverk2003) study found no racial/ethnic differences in youth-reported maltreatment for any of the five maltreatment subtypes (i.e., physical abuse, sexual abuse, emotional abuse, emotional neglect, physical neglect) when all sectors were examined together. For youth in the child welfare sector specifically, however, they found that African American youth self-reported maltreatment less frequently than other youth. However, African American youth were 12 times more likely to have a history of foster care placement after controlling for youth-reported maltreatment, income, age, and gender. Additionally, youth who did self-report maltreatment were three times more likely to have a history of foster care placement, and this association held for youth of all races (European American, Hispanic American, and Asian Pacific Islander) except African American youth, who were equally as likely to have a history of foster care placement, irrespective of their self-reported maltreatment history (Lau et al., Reference Lau, McCabe, Yeh, Garland, Hough and Landsverk2003). These findings suggest that African American youth are disproportionately taken into custody, even when reporting that they have experienced similar levels of maltreatment compared to other racial/ethnic groups. They also suggest that African American youth’s reported experiences do not correspond to child welfare-related outcomes in a similar way as other youth’s reported experiences correspond to outcomes, pointing to the presence of outside biases in determining outcomes for African American youth.

Side 1 also argues that qualitative data illustrating minoritized children and families’ lived experiences of racism in this system is often devalued in favor of multivariate, quantitative data (Detlaff et al., Reference Detlaff, Weber, Pendleton, Boyd, Bettencourt and Burton2020; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). A qualitative study with Black and Latinx parents revealed that most parents in the child welfare system felt judged by caseworkers’ racial stereotypes. In turn, they felt overwhelmed, blamed, and intimidated; perceived loss of control over outcomes; and feared their children would be removed (Merritt, Reference Merritt2021). Other qualitative data has found lack of justification in caseworker files for substantiation of maltreatment (Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Devaluing of qualitative data in research is common despite that caseworkers’ qualitative decision-making and record-keeping are often among the strongest pieces of evidence used in decisions about out-of-home placement and termination of parental rights (Havlicek & Courtney, Reference Havlicek and Courtney2016; Henry et al., Reference Henry, Victor, Ryan and Perron2020).

Finally, Side 1 argues that the claim that disproportionalities are a result of poverty and not racism is itself disproportionately advanced by White researchers and individuals without lived experiences of child welfare. Side 1 points out that researchers’ “social identities and positionalities shape their worldviews and their biases, affect how they approach and conceptualize their work, and influence their research and scholarship” (Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022, p. 514). Research on these issues by scholars of color is underrepresented, as are datasets that are sufficiently equipped with minoritized and multiply intersectional individuals to make adequate comparisons across many minoritized groups (Kokaliari et al., Reference Kokaliari, Roy and Taylor2019; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). In turn, Side 1 argues against a quo that tends towards normalizing and rationalizing the disproportionate representation of minoritized families in the child welfare system as acceptable, excusable, and justifiable (Briggs et al., Reference Briggs, Hanson, Klika, LeBlanc, Maddux, Merritt, Palusci, Panlilio, Roygardner, Schelbe, Stormer, Valentino, Vaughan-Eden and Barboza2023; Detlaff et al., Reference Detlaff, Boyd, Merritt, Plummer and Simon2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022).

Both sides of the debate converge in agreement that more research is needed and that the literature specifically lacks a sufficient corpus of studies on Black or Latinx children, much less AIAN children or those with multiracial and intersectional identities (Drake et al., Reference Drake, Jones, Chen, Font, Putnam-Hornstein, Barth and Jonson-Reid2023; Hanna, Reference Hanna and Detlaff2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Both sides also agree that unmeasured aspects of racialized poverty and economic stratification render many minoritized children at disproportionate risk for maltreatment and other childhood adversities (Barth et al., Reference Barth, Berrick, Garcia, Drake, Jonson-Reid, Gyourko and Greeson2021; Drake et al., Reference Drake, Jones, Chen, Font, Putnam-Hornstein, Barth and Jonson-Reid2023; Hanna, Reference Hanna and Detlaff2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Racial, ethnic, and intersectional identities are not variables that can statistically controlled (APA, 2023b). They are complex and dynamic aspects of individuals’ lived experiences that are frequently subjected to immeasurable systemic injustices.

Given the reality that racism has existed in the child welfare system for as long as the system itself has existed, we recommend that official CPS reports and child welfare records not be viewed as the gold standard or objective source for maltreatment data. This practice reinforces a deficit view of Black and other minoritized families by privileging evidence from court records over other types of information such as individuals’ self-reports, giving too little attention to enduring structural and systemic racism and racialized poverty that contribute to disproportionalities, and potentially leading to problematic interpretations (APA 2023a; Briggs et al., Reference Briggs, Hanson, Klika, LeBlanc, Maddux, Merritt, Palusci, Panlilio, Roygardner, Schelbe, Stormer, Valentino, Vaughan-Eden and Barboza2023; Hanna, Reference Hanna and Detlaff2021).

Discordance between prospective and retrospective assessment of childhood maltreatment

Research has increasingly highlighted the lack of concordance or agreement between prospective and retrospective assessment of childhood maltreatment data (Baldwin et al., Reference Baldwin, Reuben, Newbury and Danese2019; Danese, Reference Danese2020; Danese & Widom, Reference Danese and Widom2020). Meta-analytic evidence from 16 studies has shown that the agreement on overall maltreatment (including child abuse and neglect) between prospective and retrospective assessment formats was poor (k = .19, p < .001). Less than half of individuals (48%) with prospective documentation of child maltreatment later retrospectively reported childhood maltreatment as adults, and similarly, less than half of adult individuals (44%) who retrospectively reported childhood maltreatment had prospective data that documented it (Baldwin et al., Reference Baldwin, Reuben, Newbury and Danese2019). Baldwin and colleagues’ primary conclusions were that each format – prospective and retrospective – identified different groups of people, that prospective information from court records may be underestimates, and that retrospective information from self-reports may contain biases. The presence of racism as a potential source of bias in prospective assessment that relies primarily on court records was not acknowledged.

As another example, recent research drew conclusions from a large, prospective longitudinal study of childhood maltreatment documented in the legal system between 1967 and 1971, in which neglect was by far the most common subtype (rates of child neglect were 45.4%, whereas rates of physical and sexual abuse, respectively, were 9.2% and 8.0% of children; Danese & Widom, Reference Danese and Widom2020). In a subsequent paper using data drawn from this same sample, the authors indicated, “Court-substantiated records provide the legal standard on which child protection actions are based and thus provide the strongest possible evidence for the objective experience of child maltreatment” (Danese & Widom, Reference Danese and Widom2023, pp. 1011). However, arguments that court records are the best evidence that maltreatment occurred need to be to considered in light of all available data. The presence of racism in the child welfare system calls for greater care in reaching conclusions about maltreatment based on data that is primarily drawn from court records. It also calls into question the extent to which court records can be considered objective, especially for data gathered in the late 1960s and early 1970s, when concerns about racism in the child welfare system were just beginning and racial bias was likely even more pronounced than today (Billingsley & Giovannoni, Reference Billingsley and Giovannoni1972; Detlaff et al., Reference Detlaff, Boyd, Merritt, Plummer and Simon2021).

Given evidence of racism in the child welfare system, we cannot assume that data drawn primarily from court records and CPS reports always accurately identify individuals who have been maltreated. In the study above by Danese and Widom (Reference Danese and Widom2020), maltreatment prevalence rates were broken down according to “objective” and “subjective” measures and by race and ethnicity. Although “objective” data (prospective court and CPS records) and “subjective” data (retrospective self-reports) on prevalence rates among Black participants were comparable for overall maltreatment (54.2% and 56.8%, respectively), there was a large discrepancy in prevalence of neglect. The prevalence of neglect among Black participants was 45.8% for “objective” records, but 26.4% by “subjective” measures. This discrepancy was not as pronounced among White participants, who had a 44.8% prevalence rate of neglect according to “objective” measures and a 36.6% prevalence rate of neglect per “subjective” measures. (Discrepancies also existed for both Black and White participants in “objective” versus “subjective” prevalence of physical and sexual abuse, with a much higher percentage of both Black and White individuals reporting these two subtypes by “subjective” measurement relative to “objective” measures.) Taken all together, these data suggest that overestimates of maltreatment could plausibly vary across racial lines depending on which subtype was assessed.

We suggest that an incomplete set of interpretations has been offered in psychology research to explain weak agreement between prospective and retrospective data on childhood maltreatment. Some existing interpretations have centered on problems with individuals’ memory abilities and motivations, such as memory recall and retrieval issues because of the young age at which maltreatment occurred (Baldwin & Degli Esposti, Reference Baldwin and Degli Esposti2021; Danese & Widom, Reference Danese and Widom2020; Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023). Additional interpretations posit that individuals may withhold, fabricate, or confabulate experiences of maltreatment, due to “misinterpretation of internal images or dreams as actual experiences” (Danese, Reference Danese2020, p. 239). Recall and retrieval issues, as well as motivations to withhold disclosures, are well documented (Herman, Reference Herman2003), but they are more likely to occur due to elevated traumatic stress rather than a fabricated or confabulated trauma history.

According to scholars of trauma and memory, the term “false memories” of abuse gained popularity in the 1990s in relation to the belief that child welfare workers could elicit false accusations of sexual abuse from suggestible young children (DePrince et al., Reference DePrince, Allard, Oh and Freyd2004). However, in the child and adult trauma research literature, false recall errors more commonly reflect errors of omission (not disclosing trauma that did happen), rather than errors of commission (saying trauma happened when it did not). In terms of omission errors, children who experience maltreatment from a trusted caregiver may be vulnerable to long-term suppression of the memories of it, or “motivated forgetting or misremembering” to reconcile survival needs with the extreme betrayal characterizing the abuse (DePrince et al., Reference DePrince, Brown, Cheit, Freyd, Gold, Pezdek, Quina and Belli2012, p. 193). Furthermore, research has shown that maltreated children are no more likely than non-maltreated children to make recall and recognition errors, nor to be more or less suggestible or susceptible to believe incorrect information (Howe et al., Reference Howe, Cicchetti, Toth and Cerrito2004). Moreover, traumatized adults with or without PTSD may indeed make more false recall errors than individuals without trauma histories (Zoellner et al., Reference Zoellner, Foa, Brigidi and Przeworski2000). However, this pattern suggests that more trauma exposure, rather than less, is associated with recall problems. Regarding commission errors, rates of false recall of abuse in adults are extremely low (Lisak et al., Reference Lisak, Gardinier, Nicksa and Cote2010; Weiser, Reference Weiser2017). Evidence generally does not support that individuals confabulate or exaggerate maltreatment experiences during research participation.

Underestimation and overestimation of maltreatment in child welfare data

Foremost, we do not mean to suggest that any specific data documenting child maltreatment through prospective or retrospective formats in any study is inherently inaccurate or false. Like others have acknowledged (Danese & Widom, Reference Danese and Widom2020; Widom et al., Reference Widom, Czaja and DuMont2015), it is more likely that CPS reports and child welfare cases fail to detect (i.e., underestimate) the extent of maltreatment that occurs (i.e., leading to false negatives), rather than overestimate it (leading to false positives). Indeed, when comparing the self-reports of maltreatment in a large sample of foster care youth, half of whom were Black, with their records from CPS files, the majority of the disagreement between self-reports and case files was due to case files underestimating maltreatment (Cooley et al., Reference Cooley, Jackson and Stoolmiller2022).

However, underestimates of maltreatment or future risk for it may fall along racial lines. In a study of medical records of pediatric fractures, Lane et al. (Reference Lane, Rubin, Monteith and Christian2002) found evidence of underreporting to CPS for White families and overreporting for minoritized families. In another study, Rivaux et al. (Reference Rivaux, James, Wittenstrom, Baumann, Sheets, Henry and Jeffries2008) found that when controlling for income and other factors, caseworkers assessed Black families to be at lower risk for future maltreatment than White families. However, Black families were more likely to have their cases opened and when actions were taken, to experience removal of the child. Alternatively, White families were more likely to have their cases closed, but if actions were taken, to receive supportive family services. We want to emphasize that it should be of utmost priority for no system to fail to detect maltreatment or future risk of it.

We also want to point out, however, that overestimation of maltreatment by objective records compared to self-reports may also occur. Overestimation is present when maltreatment allegations are made, investigated, and substantiated even when maltreatment may not have occurred, and when the risk thresholds for taking action to prevent future maltreatment (such as removing children from families) are differentially applied depending on race (Eastman et al., Reference Eastman, Herz, Palmer and McCroskey2023; Rivaux et al., Reference Rivaux, James, Wittenstrom, Baumann, Sheets, Henry and Jeffries2008). Indeed, among youth in foster care, a subset with substantiated maltreatment in their case files did not self-report experiencing maltreatment, suggesting overestimation likely occurs to some extent (Cooley et al., Reference Cooley, Jackson and Stoolmiller2022).

Overestimation may be particularly true of cases characterized by neglect, which are among the most difficult and controversial to substantiate and the most likely to confound issues of poverty that disproportionately affect minoritized families (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Berkman et al., Reference Berkman, Brown, Scott and Adiele2022; Havlicek & Courtney, Reference Havlicek and Courtney2016; Roberts, Reference Roberts2014). Neglect is the most frequently substantiated subtype of maltreatment, particularly for minoritized (and specifically, Black) families (Rebbe, Reference Rebbe2018; Roberts, Reference Roberts2014), and the most frequently occurring subtype in many prospective studies (e.g., Danese & Widom, Reference Danese and Widom2020; Raby et al., Reference Raby, Labella, Martin, Carlson and Roisman2017). However, definitions of neglect are not standardized across child welfare systems, which vary by state, and despite its prevalence, neglect is the least-researched subtype (Laajasalo et al., Reference Laajasalo, Cowley, Otterman, Lamela, Rodrigues, Jud and Nurmatov2023; Rebbe, Reference Rebbe2018).

Overestimation could occur especially if racism is present (Rivaux et al., Reference Rivaux, James, Wittenstrom, Baumann, Sheets, Henry and Jeffries2008; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022), and it could lead to irreparable harm through forcible separation of children from families. To return to a main point of this paper, overestimation could account for error in prospective and “objective” data on child maltreatment by inaccurately classifying individuals, particularly those who are minoritized, as maltreated when they were not. In turn, this misclassification could influence discordance with retrospective assessment of maltreatment (i.e., court records indicate maltreatment occurred, whereas self-reports do not), and discrepancies in associations between prospective maltreatment data with adult outcomes and retrospective self-reported maltreatment data with adult outcomes. However, the possibility of overestimation of maltreatment in court records due to racism, particularly for minoritized families, has not been offered in many longitudinal studies as an explanation for the weak concordance between prospective and retrospective assessment formats (Danese & Widom, Reference Danese and Widom2020, Reference Danese and Widom2023; Reuben et al., Reference Reuben, Moffitt, Caspi, Belsky, Harrington, Schroeder, Hogan, Ramrakha, Poulton and Danese2016), nor in recent meta-analyses (Baldwin et al., Reference Baldwin, Reuben, Newbury and Danese2019; Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023), nor in studies by child welfare researchers who have observed the stark discrepancies between official investigated reports versus youth self-reports (Havlicek & Courtney, Reference Havlicek and Courtney2016).

In light of evidence that racism exists in the child welfare and legal systems (e.g., Lane et al., Reference Lane, Rubin, Monteith and Christian2002; Rivaux et al., Reference Rivaux, James, Wittenstrom, Baumann, Sheets, Henry and Jeffries2008), we encourage future researchers from all fields and disciplines to acknowledge racism as a possible interpretation for the lack of agreement in prospective versus retrospective assessments of child maltreatment. Furthermore, we also argue that official and substantiated maltreatment from court records and CPS reports may not be as accurate as previously assumed. Though it is likely that many cases of substantiated maltreatment identify children who have experienced victimization and are in great need of safety-promoting services, it is also possible that some children with CPS cases have been subjected to racism that inaccurately classified them as maltreated.

Discrepant associations between prospective and retrospective maltreatment with outcomes

The strength of the associations between prospectively documented maltreatment and retrospectively reported maltreatment with outcomes is often discrepant. For instance, a recent meta-analysis of 17 studies showed a significant association between “subjective” maltreatment and self-reported psychopathology (r = .16, p < .0001) but a null association between “objective” maltreatment and self-reported psychopathology (r = .06, p = .14; Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023). As another example, the prospective longitudinal study that assessed maltreatment from court records between 1967 and 1971 (Danese & Widom, Reference Danese and Widom2020, Reference Danese and Widom2023) described that individuals who only had “objective” data on child maltreatment from court records did not show significantly poorer adult mental health outcomes than individuals with no maltreatment, whereas individuals with “subjective” (retrospectively-reported) maltreatment had significantly poorer mental health outcomes, whether or not they also had “objective” data on maltreatment.

The authors gave several explanations for the lack of significant effects of objective child maltreatment on adult psychopathology. One plausible interpretation was that underestimation of maltreatment occurred in the legal system in the form of false negatives. Another interpretation was that recall biases associated with contemporaneous adult psychological disorders and distress could have led to higher subjective appraisal that individuals had been maltreated, but the authors ruled this out. They noted, however, that residual memory biases following earlier psychological disorders or previous distress could have become stable vulnerability factors and rendered adults more susceptible to subjective appraisal of maltreatment in the absence of current disorders or distress. The authors’ conclusions were that there was minimal impact of objective maltreatment on long-term psychopathology, “The risk of psychopathology linked to objective experiences of childhood maltreatment, even for severe cases of maltreatment identified through official court records, is minimal in the absence of a subjective appraisal” (Danese & Widom, Reference Danese and Widom2020, p. 811). In other words, they concluded that the effects of maltreatment, documented at the time it occurred in childhood, were not harmful unless individuals also appraised it as harmful when they were adults. A recent meta-analysis also concluded that adults’ mental health problems were more likely to be associated with their subjective experiences of maltreatment, rather than their objective experiences of it according to court records or child welfare data (Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023). Neither of these papers discussed the possibility that data on maltreatment from court records might not be entirely accurate (perhaps in part due to the presence of racism), which might explain why weaker-than-expected associations have been continuously observed across studies between prospective maltreatment and adult health outcomes. We offer the interpretation that adults’ retrospective reports of maltreatment may be more reliably linked to adulthood outcomes because adults’ self-reports, rather than court-records (in some but not all cases), may more accurately reflect lived experiences, correcting for both underestimates and overestimates of child welfare data confounded by race, ethnicity and income.

More broadly, we suggest that more nuanced and flexible interpretations are needed when prospective and retrospective assessments of maltreatment show differential associations with adulthood outcomes. For example, Newbury et al. (Reference Newbury, Arseneault, Moffitt, Caspi, Danese, Baldwin and Fisher2018) examined prospectively documented maltreatment via aggregated information from several sources, including multiple interviews with primary caregivers at multiple time points, home observations, and clinical case conferences for maltreatment coded from child welfare reports. Findings showed that both prospective and retrospective formats were associated with a range of adults’ psychological disorders. However, when associations between both formats were modeled together to predict outcomes, the effects of prospective reports became weaker and the effects of retrospective reports became stronger. The investigators articulated that because prospective data may underestimate individuals’ lived experiences, the most comprehensive approach is to include both retrospective and prospective information for a composite of childhood maltreatment.

Similarly, Shaffer et al. (Reference Shaffer, Huston and Egeland2008) found that when prospective child maltreatment data from multiple sources (e.g., home observations, caregiver interviews, reviews of child protection and medical records) were combined with retrospective self-reports of maltreatment, individuals with maltreatment from both formats reported the highest levels of socioemotional problems. Similar to Newbury et al. (Reference Newbury, Arseneault, Moffitt, Caspi, Danese, Baldwin and Fisher2018), they also concluded that both prospective and retrospective formats should be aggregated for comprehensive identification of maltreatment, and “it is erroneous to assume that retrospective reports of child maltreatment are inherently problematic as compared to prospectively identified reports. Rather, both methods provide valuable information regarding maltreatment experiences” (Shaffer et al., Reference Shaffer, Huston and Egeland2008, p. 691). The added value of both methods, rather than the superiority of one, was also documented in a study on the associations between maltreatment and antisocial behavior (Smith et al., Reference Smith, Ireland, Thornberry and Elwyn2008).

Meta-analytic evidence for the intergenerational transmission of maltreatment

Finally, although intergenerational transmission of maltreatment is not the focus of this paper, recent meta-analytic findings also support the conclusion that prospective maltreatment assessment formats are not necessarily superior to retrospective formats because both formats yield comparable information on rates of intergenerational transmission. Across 142 studies, one meta-analysis found that the association of maltreatment in parents’ childhoods to their children’s rearing environments was significant but modest (d = .45; Madigan et al., Reference Madigan, Cyr, Eirich, Fearon, Ly, Rash, Poole and Alink2019). Another meta-analysis of 84 studies also reported a medium effect size of maltreatment (r = .29) across generations (Assink et al., Reference Assink, Spruit, Schuts, Lindauer, van der Put and Stams2018). To address whether aspects of maltreatment measurement moderated transmission, both studies used a 10- or 13-item dimensional score of methodological quality (Thornberry et al., Reference Thornberry, Knight and Lovegrove2012) applied to all studies that included whether the sample was representative of the population, maltreatment was assessed from validated measures, maltreatment was drawn from official records, multiple sources were assessed, measurement was prospective or retrospective, and attrition was <40% (Assink et al., Reference Assink, Spruit, Schuts, Lindauer, van der Put and Stams2018; Madigan et al., Reference Madigan, Cyr, Eirich, Fearon, Ly, Rash, Poole and Alink2019).

Madigan et al.s’ (Reference Madigan, Cyr, Eirich, Fearon, Ly, Rash, Poole and Alink2019) study showed that the intergenerational transmission of maltreatment was not moderated by a dimensional or categorical score of methodological quality. Furthermore, effect sizes of intergenerational maltreatment did not depend on how maltreatment was assessed (i.e., by official records versus self-report) or by prospective versus retrospective format. In other words, the effect sizes were similar if official records assessed parents’ childhood maltreatment and their maltreatment of children, if self-reports were used to assess both, or if official reports assessed parents’ childhood maltreatment but self-reports assessed their maltreatment of children (or vice versa). Although Assink et al.s’ (Reference Assink, Spruit, Schuts, Lindauer, van der Put and Stams2018) study found that overall methodological quality moderated observed effects of transmission (rates of transmission were lower when study quality was higher), the only individual quality moderator that strengthened effects of transmission was official records in the present generation, as opposed to official records from parents’ childhoods. No aspect of assessment method or format for parents’ childhood maltreatment history (e.g., official records, prospective versus retrospective assessment) moderated the effect of transmission. Effect sizes for intergenerational maltreatment from parents’ childhoods to their children’s rearing environments seem to be similar in magnitude regardless of whether parents’ childhood maltreatment was assessed prospectively (from official records or other sources) or retrospectively.

Recommendations against invalidating terminology

We recommend against labeling court records as “objective” or “true” evidence for maltreatment because there is evidence suggesting these records are not always accurate, particularly for minoritized groups. We also recommend a shift away from devaluing adults’ retrospective reports of child maltreatment when self-reports are not previously documented prospectively, such as in “individuals who construe their childhood experiences as maltreatment despite the lack of documented history” (Danese & Widom, Reference Danese and Widom2020, p. 815). Moreover, we recommend that researchers avoid terms such as “subjective”, “appraisal”, or “perceived” to refer to individuals’ reports of lived experiences. Instead of using the term “subjective” experience, such as in “the effects of childhood adversity on psychopathology are primarily driven by a person’s subjective experience” (Francis et al., Reference Francis, Tsaligopoulou, Stock, Pingault and Baldwin2023, p. 1185), we advise that scholars use the term “lived” experience. Labeling experiences of potentially egregious harm and victimization as “subjective” because there is no record of it happening from “objective” sources may be an invalidating and potentially traumatizing – and retraumatizing – experience. Such terminology could hinder survivors’ willingness to disclose traumatic experiences, and it could reinforce a culture of silencing victims and enabling perpetrators (Herman, Reference Herman2003). More neutral language is illustrated in a study by Smith et al. (Reference Smith, Ireland, Thornberry and Elwyn2008) that used the terms “official substantiated maltreatment” versus “self-reported maltreatment” to differentiate the sources of their maltreatment data. The authors did not presume that one source of data was more valid than the other, which was consistent with their findings that both sources comparably predicted antisocial behavior. We recommend using neutral but precise terms that indicate the source of the data, such as “court-substantiated maltreatment” or “child welfare records” rather than “objective data” that is imprecise and potentially inaccurate and whose opposite, “subjective data,” devalues individuals’ perspectives. In addition to advocating for validating terminology in regards to individuals’ self-reports of maltreatment that did occur, we also caution against using terms such as “false negatives” in regards to individuals’ self-reports that maltreatment did not occur (Havlicek & Courtney, Reference Havlicek and Courtney2016).

Furthermore, even if disproportionalities cannot be measured as precisely as desired, investigators should hold under consideration the reality that the child welfare system may function differently for minoritized youth and families, and for Black individuals specifically. Black youth have been found to self-report experiencing maltreatment less than youth from other racial/ethnic groups yet to be placed out of home at much higher rates. Moreover, their reports of experiencing maltreatment do not correspond to their odds of placement in the same ways that they correspond for White youth (Lau et al., Reference Lau, McCabe, Yeh, Garland, Hough and Landsverk2003).

On a different note, it is also possible that prospective data on childhood maltreatment is accurate, but the now-grown individual has valid reasons for choosing not to disclose it. These reasons may include hesitancy to disclose painful experiences to unfamiliar sources without adequate therapeutic rapport, or preference to avoid thinking about potentially traumatic stimuli (Herman, Reference Herman2003; Narayan et al., Reference Narayan, Rivera, Bernstein, Castro, Gantt, Thomas, Nau, W. and Lieberman2017). A study with college students indicated that of those reporting a history of childhood victimization such as maltreatment, only 53.3% had ever told someone about their experiences and if they did, they were more likely to first disclose to a parent or friend. Only 20.8% had first disclosed to a professional, suggesting that if disclosure does happen, it often happens to informal sources (Desir & Karatekin, Reference Desir and Karatekin2019). Preference to avoid thinking about maltreatment could also certainly be the case for individuals who have avoidance symptoms of PTSD, feel unsafe or invalidated in the disclosure context, or perceive themselves as vulnerable in other ways because of having minoritized identities, being pregnant, or having low social support (Narayan et al., Reference Narayan, Rivera, Bernstein, Castro, Gantt, Thomas, Nau, W. and Lieberman2017; Herman, Reference Herman2003). Individuals’ odds of disclosing maltreatment are also dynamic. They may vary across time and change from decade to decade depending on one’s developmental stage of life, parenting status, and other life events. Odds of disclosure also vary within cultural groups and depend on the sociopolitical context.

Recommendations for more inclusive and culturally informed methods

Studies that use each format of maltreatment assessment – whether prospective or retrospective – and draw on different sources, including parents, health providers, school personnel, child welfare workers, and individuals themselves, may contribute more nuanced perspectives. All sources might also contain elements of bias or measurement error. Thus we do not recommend that any one source be considered the gold standard. Ideally, two or more sources or formats would be aggregated to assess for the presence of maltreatment, as done in several studies (e.g., Newbury et al., Reference Newbury, Arseneault, Moffitt, Caspi, Danese, Baldwin and Fisher2018; Shaffer et al., Reference Shaffer, Huston and Egeland2008; Raby et al., Reference Raby, Labella, Martin, Carlson and Roisman2017).

We recommend that adults’ retrospective reports of maltreatment be considered in addition to prospective data as evidence of maltreatment, rather than disbelieved unless prospective evidence confirms it. We agree with the following recommendation from a recent bioethics report to “encourage medical providers to be thoughtful about their reporting, to consider social context and alternative explanations for things they think are worrisome, to base their concerns on evidence as much as possible, to avoid making assumptions and instead seek clarification from families about areas of concern, and to ensure that they are being internally consistent. Consideration of harm should be comprehensive and informed by historical contexts” (Berkman et al., Reference Berkman, Brown, Scott and Adiele2022, p. 34). We also concur with the recommendation of APA (2023a) that all parties from research, healthcare, policy, and legal systems be aware of biases and use multi-informant, multi-method, DEI-informed perspectives when assessing childhood maltreatment.

We also recommend that methods to assess maltreatment be dimensional (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993), moving beyond a binary distinction of maltreatment versus no maltreatment to examine measures of harmful parenting practices along a continuum. Many parenting practices that are not universal must be assessed in the cultural context in which they occur. Some deviations in culturally appropriate practices that might signal abuse or neglect in some cultures but also contain ambiguity should be assessed according to the perspectives of members from the identified cultural group (APA, 2023a, b; Raman & Hodes, Reference Raman and Hodes2012). There are cultural differences between White versus Black individuals on what types of parental discipline constitute maltreatment (Kesner et al., Reference Kesner, Kwon and Lim2016) and the subsequent impact on children’s development. For instance, mild physical discipline in early and middle childhood predicted higher parent- and youth-reported externalizing behaviors for European-American adolescents, but lower externalizing behaviors for African American adolescents. The cultural context influences the meaning and effects of physical parenting strategies on youth behavior. Milder physical discipline may be more normative in African American than European American families (Lansford et al., Reference Lansford, Deater-Deckard, Dodge, Bates and Pettit2004).

Cultural differences in perception and identification of maltreatment need to be continually examined across service sectors and from various types of mandated reporters. One study found that attitudes towards corporal punishment and ratings of its abusiveness differed across racial groups in U.S. teachers-in-training, with African American respondents exhibiting more favorable attitudes towards corporal punishment than Asian and White respondents (Kesner et al., Reference Kesner, Kwon and Lim2016). What behaviors are considered to be maltreatment also vary across time as social norms evolve (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993). Investigators from the MLSRA prospective longitudinal study re-reviewed all multi-informant maltreatment data originally gathered many decades ago and recoded all of it based on updated definitions (Raby et al., Reference Raby, Labella, Martin, Carlson and Roisman2017). However, this commendable practice is extremely time-intensive, often not possible, and as a result, is rarely undertaken. At the very least, research conducted in samples where maltreatment definitions might be outdated should state this as a limitation in relation to possibly affecting the interpretation of findings.

Cultural and generational differences in parenting practices and what constitutes maltreatment by Western definitions could contribute to discrepancies in prospectively documented versus retrospectively reported maltreatment, particularly in U.S. legal systems. To reduce the measurement error and more importantly, eliminate the practice of devaluing individuals’ lived experiences by assuming official court records of maltreatment are superior to self-reports, maltreatment could be assessed according to culturally specific definitions that providers adopt for all families in their care, ideally before allegations are made to increase standardized and equitable treatment across groups. It is important to weigh the perspectives of individuals with lived experiences of maltreatment fairly and equitably with other sources.

Furthermore, cross-talk is needed between researchers across social science and medical disciplines. Many scholars who have identified racial biases in the child welfare system come from social work, economics, sociology, and pediatrics, whereas many of the investigators who use official court records as a gold standard measure for maltreatment research come from psychology. Collaboration across disciplines and cultures on parenting and child development in non-Western societies could lead to diversity-informed discoveries about the causes and consequences of maltreatment. We are currently in an era where it is more feasible than ever before to conduct interdisciplinary work because of the easy access to information across disciplines.

Finally, we do not argue for trading methodological rigor for inclusivity. Both can be achieved. Regarding retrospective data on childhood maltreatment, we hope that researchers continue to consider the possibility of inflated associations between data from the same informants, such as retrospectively reported life experiences and contemporaneously reported health outcomes (Reuben et al., Reference Reuben, Moffitt, Caspi, Belsky, Harrington, Schroeder, Hogan, Ramrakha, Poulton and Danese2016). We agree with the conclusion of Reuben et al. (Reference Reuben, Moffitt, Caspi, Belsky, Harrington, Schroeder, Hogan, Ramrakha, Poulton and Danese2016) that “‘sunny’ and ‘gray’ dispositions may bias predictions from retrospective [adversity] measures toward underestimating [the impacts of adversity] on objectively measured outcomes and overestimating impacts on subjectively measured outcomes” (p. 1111). Aspects of individuals’ current mental states, internalized distress, perceived stress levels, or unmeasured personality characteristics may influence associations between independent and dependent variables if all are reported by the same informant. When possible, we recommend covarying for factors affecting reporting biases (Narayan et al., Reference Narayan, Merrick, Lane and Larson2023).

Recommendations for more inclusive research questions

Relatedly, research questions about the long-term effects of childhood maltreatment can be framed more inclusively, rather than categorically as to whether objective or subjective experiences of maltreatment matter more or less than one another. Psychopathology may develop as an outcome of identifiable instances of abuse and/or neglect during childhood, and psychopathology may also develop as a function of subsequent or cumulative lived experiences of victimization and the transactional relationships between experiences, appraisals, and well-being, as the DP perspective has explained since its inception (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Cicchetti, Reference Cicchetti, Cicchetti and Carlson1989; Masten, Reference Masten2006). The development of psychopathology following childhood maltreatment varies according to individual differences, various dimensions of maltreatment (e.g., timing, frequency, chronicity), and multifinality between maltreatment and outcomes (Cicchetti & Rogosch, Reference Cicchetti and Rogosch1996). It is likely that a wide range of experiences that occur within the context of maltreatment (including allegations alone) may be harmful to children and families (Merritt, Reference Merritt2021; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022).

In considering research questions on the long-term consequences following maltreatment, studies also need to increasingly include PTSD in adults as a mental health outcome of interest, as traumatic stress is a natural sequela following exposure to traumatic events (Narayan et al., Reference Narayan, Lieberman and Masten2021). PTSD is itself a grave form of psychological damage, with well-documented functional and structural abnormalities in the brains of children and adults with PTSD (Harnett et al., Reference Harnett, Goodman and Knight2020). We are concerned about statements like the following because of their potential harm to victims of childhood maltreatment:

Suggesting that retrospective “subjective” measures of child maltreatment are not likely to identify actual harm associated with maltreatment minimizes the painful and damaging experiences of victims and the associations between their traumatic experiences and real and observable injury and suffering. Whether intentional or not, this perspective represents a form of victim-blaming that may hinder potential treatment and recovery (Herman, Reference Herman2003; Poletti et al., Reference Poletti, Gebhardt and Raballo2022). Furthermore, PTSD is often best reported by the traumatized individuals themselves, as many physiological, cognitive, and internalizing aspects of this disorder make it difficult for outsiders to observe. Adopting our recommended inclusive approach, assuming that adults’ retrospective self-reports of maltreatment are valid, then it follows that associations found between self-reported maltreatment and self-reported outcomes (e.g., PTSD symptoms) are likely to be valid, particularly if they hold after controlling for factors affecting reporting biases.

Recommendations for more diverse samples and researchers

Developmental psychopathology research on child maltreatment is urgently in need of large, rigorous, multi-method, prospective longitudinal samples that are not predominantly White. Ideally, these samples would have prospective data on childhood maltreatment gathered from multiple sources (e.g., parents, home observations, CPS reports, court records, investigator case conferences, child self-report) and retrospective self-reported data on maltreatment that enables inquiry into topics of discordance in reporting formats, and how they relate to lived experiences. Many of the longest-running prospective studies on child maltreatment have drawn conclusions from predominantly White samples, such as the Dunedin Study (Poulton et al., Reference Poulton, Moffitt and Silva2015), the E-Risk Study (Moffitt et al., Reference Moffitt2002), the MLSRA (Sroufe et al., Reference Sroufe, Egeland, Carlson and Collins2005) and the Widom sample (Widom, Reference Widom1989). Aligned with the APA (2023b)’s standards to clearly communicate when samples are not adequately representative of minoritized populations, published conclusions about the effects of prospective-documented or retrospectively reported maltreatment should be clearly qualified if made from predominantly White samples that do not reflect the proportionality of White versus minoritized children in the child welfare system (Detlaff et al., Reference Detlaff, Weber, Pendleton, Boyd, Bettencourt and Burton2020; Merritt, Reference Merritt2021).

In contrast to other longitudinal studies of predominantly White samples, the extensive array of large, multi-method, multilevel research studies with maltreated children from Mt. Hope Family Center over the past several decades is an exemplar of conducting rigorous research with families from diverse racial and ethnic backgrounds (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Brown et al., Reference Brown, Rogosch, Shacklewood and Cicchetti2023; Cicchetti, Reference Cicchetti2016). Notably, the Maltreatment Classification System (MCS), originally developed by Dante Cicchetti and colleagues, is one widely used approach to classifying children’s maltreatment that standardizes ratings on several dimensions of the experience [e.g., subtype(s), frequency, severity, chronicity, developmental period(s), perpetrator(s), etc.]. As part of the MCS, independent coders rate these dimensions based on caseworkers’ narratives, rather than directly relying on the caseworkers’ labels of maltreatment, which widely vary between professionals (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993). While this process still relies on child welfare information, it promotes standardization and inter-rater reliability, and it eliminates many biases.

In addition to broadening the diversity of study samples in maltreatment research, it is also vital to broaden the diversity of scholars engaged in this research. The identities and lived experiences of researchers affects the questions we ask, the measures we choose, and the interpretations we make from our research. Therefore, it is important for the field to prioritize inclusion of scholars of color from under-represented backgrounds, as well as to commit to disseminating anti-racist scholarship (APA 2023a; Briggs et al., Reference Briggs, Hanson, Klika, LeBlanc, Maddux, Merritt, Palusci, Panlilio, Roygardner, Schelbe, Stormer, Valentino, Vaughan-Eden and Barboza2023; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022). Efforts to promote the work of scholars with diverse social identities and lived experiences will lead to richer and more DEI-informed research endeavors into the developmental psychopathology of maltreatment.

Conclusions and guidance for the future of developmental psychopathology

We have an opportunity as DP researchers of childhood maltreatment to build on the foundation of knowledge from renowned scholars who conducted decades-long studies and dispelled the notions that childhood maltreatment was not harmful nor common. We are well-positioned to move forward armed with more DEI-informed approaches for interpreting existing data and planning future research endeavors. Novel research questions should be framed with the assumption that long-term psychopathology (as well as other health outcomes and resilient functioning) develop from multiple dimensions of the maltreatment experience (Cicchetti & Rogosch, Reference Cicchetti and Rogosch1996; Cicchetti, Reference Cicchetti2016). These dimensions include the characteristics of the adversity itself, and the ways in which individuals interpret their lived experiences and perceptions of harm through their unique cultural lenses. Interpretations may change over time as result of subsequent traumatic and therapeutic experiences.

We are not the first to make the point that court records are not a pure gold standard nor to suggest that their objectivity may be impacted by racism (e.g., Smith et al., Reference Smith, Ireland, Thornberry and Elwyn2008). As Hollis (Reference Hollis2013, p. 3) asserted, “The experience of maltreatment—not the presence of a substantiated allegation of maltreatment—influences a child’s healthy development.” Nor are we the first (by any stretch) to point out how the inequities and disproportionalities in the child welfare system impact children and families. We advocate for increased dialogue and collaboration between psychology researchers and our colleagues in the fields of social work, sociology, economics, medicine, and public health who have been investigating this issue. We stand with them in calling for change in child welfare practices.

In our view, the field of child maltreatment research must move forward with the perspective that both prospective and retrospective assessment formats provide valuable, meaningful, and informative insights. We think this perspective will be particularly helpful for understanding the legacy of historical trauma on intergenerational processes related to child maltreatment. If retrospective assessments of maltreatment were presumed to be invalid, that would preclude understanding of the effects of trauma from previous generations in many cultures where prospective documentation rarely exists. Many cultures rely on elders’ narratives, families’ story-telling, and other oral history methods of documenting historical victimization and oppression (Hanna, Reference Hanna and Detlaff2021; Henry et al, Reference Henry, Carnochan and Austin2014; Tajima et al., Reference Tajima, Day, Kanuha, Rodriquez-JenKins and Pryce2022).

It is important for the future of maltreatment research to consider the earliest lessons from the DP perspective: Many dynamic, transactional factors influence the emergence of developmental phenomena such as abuse and neglect and the pathways following them (Barnett et al., Reference Barnett, Manly, Cicchetti, Cicchetti, S. and oth1993; Cicchetti, Reference Cicchetti, Cicchetti and Carlson1989). Ongoing transactions occur between children and their relationships, communities, and societies; and individual differences shape lived experiences and how they are internalized to influence developmental trajectories (Cicchetti & Rizley, Reference Cicchetti, Rizley, Rizley and Cicchetti1981; Masten, Reference Masten2006; Narayan et al., Reference Narayan, Lieberman and Masten2021). The presence of maltreatment alone does not signal psychopathology or maladjustment. Many individuals experience adversity without significant consequences – resilience is not the exception but the rule – so lack of associations between maltreatment and negative outcomes should be further examined for resilience processes at play (Egeland et al., Reference Egeland, Carlson and Sroufe1993; Narayan et al., Reference Narayan, Lieberman and Masten2021). Interventions must be offered to those with highest clinical need, regardless of how maltreatment was assessed (Newbury et al., Reference Newbury, Arseneault, Moffitt, Caspi, Danese, Baldwin and Fisher2018).

Since maltreatment is relatively common (CDC, 2023), it is also important that interpretations about its effects be written in a way that validates lived experiences (APA, 2023a, b). Researchers need to remember that consumers of our research findings may be individuals who experienced the adversities that we write about, and they may have been treated harmfully or unjustly by the social systems designed to prevent harm and injustice. We recommend that disclosures of childhood maltreatment be treated as true, regardless of whether they occurred in childhood or adulthood. The absence of corroborating evidence to support individuals’ disclosures can be viewed as a flaw in the system, indicating a need to obtain better evidence rather than to question the validity of the disclosure. Maltreatment research has evolved over the past 40 years to sharpen and deepen understanding of what we know about childhood abuse and neglect, development, lived experiences, and memory. We have new opportunities to acknowledge the long history of racism in the child welfare system. It is time to move forward with more DEI-informed approaches that begin from a place of respecting individual narratives and self-reports of lived experience as conveying a fundamental truth that is not recoverable in any other way.