Abstract
Problematic sexual behavior (PSB) among youth presents a persistent challenge within child welfare and multidisciplinary response systems. Because multidisciplinary responses shape investigation, referral, and service decisions, it is important to understand how professionals—including clinicians, law enforcement, child advocacy center staff, child welfare professionals, and prosecutors—perceive children who act out sexually on other children. This study examines how case and respondent characteristics jointly shape professional perceptions of PSB using vignette-based survey data from 363 multidisciplinary team members across Texas. Respondents assigned to vignettes involving a 14-year-old child displaying PSB were more likely than those assigned to a 10-year-old to agree the child was old enough to accept responsibility, knew the behavior was wrong, and should face criminal consequences. Political ideology, professional role, and geographic area of practice were each independently associated with these judgments. Mental health professionals and liberal-identifying respondents were less likely to endorse criminal consequences, while court/legal professionals and conservative-identifying respondents showed opposing patterns. Suburban practitioners were more likely than rural practitioners to endorse responsibility and knowledge of wrongdoing, while urban practitioners were less likely to endorse criminal consequences. Findings highlight the need for developmentally informed, coordinated approaches to training and intervention across multidisciplinary systems.
Keywords
Introduction
Child sexual abuse (CSA) is a serious, persistent public health issue. Global estimates suggest that approximately 20% of girls and 5–10% of boys experience some form of sexual abuse during childhood, though prevalence varies substantially across regions and measurement approaches (Stoltenborgh et al., 2011). More recent analyses indicate that rates of sexual violence against children may be higher than previously documented, with significant variation by age, sex, and geographic context (Cagney et al., 2025; Gewirtz-Meydan & Finkelhor, 2020). Although the consequences faced by child victims can be mitigated through evidence-based interventions, many victims are never identified and therefore do not receive clinical assessment to determine their needs or connect them with appropriate therapeutic services (Johnson, 2022).
This lack of treatment is particularly salient when the perpetrator is another youth. Due to underreporting, estimates of youth-perpetrated CSA are unreliable; however, recent data indicate that between one-third and 70% of all CSA cases are youth-perpetrated (Finkelhor, 2009; 2015; Gewirtz-Meydan & Finkelhor, 2020; Hackett, 2014; Shawler et al., 2020).
When sexual behavior of one child harms another, it is critical to support the safety and well-being of all children involved through evidence-based services and developmentally appropriate responses. Recent decades reflect a developing shift in the classification—and ultimately the treatment—of children who act out sexually on other children. Clinicians and other professionals are increasingly moving away from labeling these children criminally as sexual offenders and toward a focus on treating the behavior itself. This shift has resulted in growing attention to problematic sexual behavior (PSB). According to the National Center on the Sexual Behavior of Youth (NCSBY, n.d.), PSB can be defined as “…deviations from normative or typical sexual behavior. Problematic sexual behaviors are a set of behaviors that are non-normative, are considered unacceptable by society, and can cause impairment in functioning.”
Despite growing attention to PSB, relatively little is known about how practitioners, particularly those tasked with responding to these cases, perceive children who exhibit PSB. This gap matters because professional perceptions directly shape how cases are investigated, what services are allocated, and whether children are approached as youth in need of developmental support or as offenders in need of punishment. Research has examined how adults perceive and respond to children’s sexual behaviors in international contexts, including Sweden (Friedrich et al., 2000) and Australia (Ey & McInnes, 2017; Larsson & Svedin, 2002; Marriage et al., 2017). Related vignette-based work—though focused on adult perpetrators rather than youth with PSB—has shown that gendered assumptions shape perceptions of sexual abuse more broadly: male victims are viewed as less credible and more culpable, and female perpetrators are evaluated more favorably than male perpetrators (Rogers & Davies, 2007). More recently, studies by Shawler and colleagues (2020), Mangold et al. (2021), and Theimer et al. (2023) have begun to document how American adults and professionals perceive PSB cases specifically. However, experimental research systematically examining how both case characteristics and respondent characteristics jointly shape professional perceptions of child-initiated PSB remains limited—a gap the present study seeks to address.
Beyond case characteristics, individual and contextual characteristics of the professionals evaluating these cases may also shape their judgments—including their political orientation, professional role, gender, race, and the geographic context in which they practice. The literature review that follows addresses each of these factors in turn and provides the basis for the present study’s examination of how case and respondent characteristics may jointly influence professional perceptions of PSB.
Problematic Sexual Behavior (PSB)
Across mental health, child welfare, advocacy, and juvenile justice contexts, professionals have increasingly moved away from labeling children who engage in these behaviors as “sexual offenders” and toward developmentally informed terminology that focuses on the behavior rather than the child’s identity. This shift reflects growing recognition that adult-based sexual offender frameworks are ill-suited to children and that PSB is more appropriately understood as a behavioral concern within the context of youth development (Chaffin et al., 2008; ATSA Association for the, 2023). This shift has resulted in the increased use of the term PSB and a move away from terminology that labels youth as offenders. The definition commonly used in clinical and advocacy contexts originates with Chaffin et al. (2008): PSB can “include any child-initiated behaviors involving sexual body parts (i.e., genitals, anus, buttocks, or breasts) and are developmentally inappropriate and/or potentially harmful to themselves or others.”
This terminological and conceptual shift is not without structural complications, however. Formal system involvement through child welfare investigation or juvenile justice referral often serves as the primary pathway through which families are connected to evidence-based PSB treatment (Letourneau et al., 2009). When cases are handled outside the formal system, families may not receive intervention, particularly given well-documented barriers to voluntary treatment access including cost, transportation, and stigma (Aarons et al., 2011; Garland et al., 2005; Kazdin et al., 1997). This tension between avoiding harmful labeling and ensuring meaningful access to services remains an unresolved challenge in PSB response.
Estimating the prevalence of PSB in the general population is challenging, as definitions, measurement approaches, and population samples vary considerably across studies. Although a precise incidence rate is difficult to obtain, PSB does not appear to be rare: between one-third and 70% of CSA cases are estimated to be youth-perpetrated (Finkelhor, 2009; Gewirtz-Meydan & Finkelhor, 2020; Shawler et al., 2020), and these figures likely underestimate the true scope given the significant underreporting of child-on-child sexual behavior. Where specific prevalence estimates have been offered, they must be interpreted carefully. For example, Allen (2017) found that approximately 20–25% of children in a clinical sample had sexual behavior concerns, but only 6.9% of that sample displayed interpersonal sexual behavior, which is the type most directly relevant to PSB cases that come to professional attention. These figures highlight the importance of distinguishing between the broader universe of sexual behavior concerns and the interpersonal cases most likely to come to the attention of coordinated child-serving systems, including Children’s Advocacy Centers (CACs) and multidisciplinary teams (MDTs). CACs coordinate investigation, assessment, advocacy, and service referrals for child maltreatment cases, including but not limited to child sexual abuse; however, CAC involvement in PSB cases varies, and not all centers have formal protocols for responding to youth with PSB, particularly adolescents (Theimer et al., 2023).
There are important nuances between PSB and sexual offending based on age, jurisdiction, and case characteristics. Some youth sexual behaviors may be illegal depending on local statute—including behaviors not atypical for adolescents, such as consensual sexual activity between similarly-aged teens that violates statutory requirements—while other behaviors that cause developmental harm may not meet the legal threshold for a chargeable offense. Additionally, youth may engage in illegal behavior and not be adjudicated. Importantly, problematic sexual behavior and illegal sexual behavior are overlapping but distinct categories: some youth who display PSB may never be adjudicated or formally charged, while some legally prohibited sexual behaviors may not meet clinical definitions of PSB depending on the ages involved, developmental context, consent, coercion, and jurisdictional statute. For the purposes of this study, PSB refers to any developmentally inappropriate sexual behavior, adjudicated or not. The present study also focuses specifically on sibling-initiated PSB, a context that has been historically under-researched and subject to particularly stigmatizing labels, such as “sibling incest,” that can exacerbate negative family responses and create barriers to treatment-seeking (Finkelhor, 1980; Meiselman, 1978; Taylor et al., 2021).
Perceptions of PSB
Public and professional perceptions of youth displaying PSB are shaped by policy, media, and family dynamics. The Adam Walsh Child Protection and Safety Act (AWA) of 2006 established a federal framework requiring juveniles aged 14 or older convicted of qualifying sexual offenses to register as sex offenders, in some cases for life. However, implementation varies considerably across states, with many having yet to fully integrate the AWA into their procedures and state-specific laws differing substantially from federal requirements (see https://www.juvenilecompact.org/age-matrix for current state-by-state information). Such legislation blurs distinctions between adult offending and developmentally-influenced juvenile behavior, reinforcing punitive rather than rehabilitative approaches—an orientation in direct contrast to the goals of the juvenile justice system (Bernard & Kurlychek, 2010). Placement on a public registry during childhood can impose lasting stigma, disrupting education, peer relationships, and future opportunities, and research indicates that registration laws may even deter prosecution as prosecutors avoid cases with mandatory lifetime registration consequences (Letourneau et al., 2009). These policies shape public opinion by equating children with PSB to the predatory behavior of adult offenders—a perception that is inaccurate and undermines rehabilitation and early intervention.
Public opinion remains overwhelmingly punitive, with one study finding that over 69% of participants—including many professionals—supported public registry requirements for youth who have committed sexual offenses without considering the context of the offense (Salerno et al., 2010; Salerno et al., 2010). This is problematic because allocation of treatment is shaped by perceptions, meaning the risks and needs of some youth will not be met. Research consistently demonstrates that youth who receive diversion and treatment are substantially less likely to reoffend than those who do not, and that recidivism rates for sexual offenses among youth have declined markedly in recent decades (Caldwell, 2016; Lussier & McCuish, 2024). Moreover, committing a sexual offense as a child is not a strong predictor of adult sexual offending, and broader patterns of antisocial or nonsexual offending are often more robust indicators of continued risk (Terry, 2012; Caldwell, 2016). Such attitudes reflect how legislation intended to protect children can perpetuate moral panic and stigma, inadvertently restricting youth access to therapeutic resources and reducing their likelihood of prosocial development (Terry, 2012).
Caregiver involvement in treatment is a cornerstone of evidence-based approaches to PSB, yet some interventions do not actively engage caregivers—a gap that can limit treatment effectiveness. CAC professionals often assess caregiver reactions early in the process to gauge family readiness for intervention (Taylor et al., 2021), and parents who respond with empathy and willingness to participate foster an environment conducive to healing, while disbelief or denial can heighten perceived risk and need for formal intervention (Chaffin, 2008; Theimer et al., 2023). Many caregivers report discomfort discussing sexual health, which may limit their ability to establish healthy boundaries and communicate about appropriate behavior (Shtarkshall et al., 2007), and providing accurate, nonjudgmental sexual education can serve as both a preventive measure and a protective factor following incidents.
Research has shown that case characteristics, particularly the age and gender of the child displaying PSB, influence how these cases are perceived. Age appears to have the strongest effect, with an inverse relationship between age and perceived trustworthiness (Burt & Estep, 1981) and perceived age increasing the likelihood of victim blaming (Rogers et al., 2011). Developmental maturity serves as a heuristic for moral and legal responsibility. As children age, they are perceived as capable of intent and thus more blameworthy, even though adolescents’ cognitive and emotional regulation capacities are still forming well into late adolescence (Rogers & Davies, 2007). Perceived age can also shape attributions of blame, as Rogers et al. (2011) found that girls who appeared to have undergone puberty were blamed more harshly for sexual victimization than prepubescent girls, reflecting adult-oriented biases about sexuality and gender roles that can extend to judgments of children (Esnard & Dumas, 2013).
While prior research found 93% of youth who committed sexual offenses were male (Fox, 2017), these gender differences are most pronounced among adolescent populations; school-age children show a more even gender distribution, with some research suggesting PSB may actually be more common among girls at preschool ages (Allen, 2017; Silovsky & Niec, 2002). This gender by age interaction suggests that the salience of gender as a perceptual cue may vary depending on the developmental stage of the child, and that perceptions documented in prior research may not generalize equally across age groups.
Multidisciplinary Team Members’ Perceptions of PSB
Due to the complex nature of PSB, multiple agencies often handle a single case—including schools, law enforcement, child welfare agencies, CACs, district attorneys, and clinicians—and their coordinated response has been shown to improve outcomes for affected youth and families (Hackett et al., 2016; Masson & Hackett, 2004). Law enforcement and child protective services are often among the first systems to respond to children who exhibit PSB, making their initial framing of the situation consequential for how the case proceeds. Clinicians’ perceptions are equally important, as misconceptions that children with PSB are predatory, sexually deviant, or likely to become adult offenders may contribute to stigmatizing responses that inhibit treatment engagement and healthy development (Chaffin et al., 2008). Applying adult-oriented frameworks to children’s cognitions may result in services that do not match the child’s actual risks and needs—a mismatch that itself increases the likelihood of future offending (Terry, 2012). Instead, language that centers the behavior, rather than the identity of the child, supports more accurate assessment and opens pathways to evidence-based treatment (Chaffin & Bonner, 1998).
Beyond professional roles, a range of individual and contextual characteristics may shape how MDT members evaluate PSB cases. Political orientation has been linked to variation in punitive attitudes toward juvenile offending, with conservative orientations associated with greater emphasis on accountability and liberal orientations tending toward rehabilitative and contextual explanations of behavior (Salerno et al., 2010). These ideological differences have been observed even among professionals with specialized training, suggesting that formal education may not fully mitigate ideologically driven variation in case interpretation (Letourneau et al., 2009). Respondent gender may also matter, as research on hypothetical cases involving various forms of sexual abuse has documented gender differences in attributions of responsibility and punitiveness (Davies & Rogers, 2009; Muniz et al., 2021; Muniz & Powers, 2022).
Prior research demonstrates that professional decision-making in child maltreatment cases is shaped by racial bias, with evidence of differential reporting, substantiation, and interpretation of risk across racial groups. Studies have documented racial disproportionality at multiple stages of the child welfare process, including reporting, investigation, and placement decisions, with Black children experiencing higher rates of system involvement than White children (Palusci & Botash, 2021). Research further suggests that implicit and structural biases may influence how professionals assess risk, make referrals, and interpret family circumstances, contributing to disparities in case outcomes (Harris, 2020). These patterns indicate that race may shape not only case trajectories but also the perceptual frameworks through which professionals evaluate children and families. Finally, the geographic context of professional practice may independently shape PSB-related judgments through differential access to training and evidence-based resources, variation in community norms around youth sexuality and accountability, and differences in how CAC systems and partner agencies are organized across urban, suburban, and rural settings. Rural communities often have distinct professional cultures and resource environments compared to urban and suburban settings, which may influence how professionals conceptualize and respond to cases involving children. Understanding how these individual and contextual factors shape professional judgment, alongside case characteristics, is essential for identifying sources of variability in MDT responses and developing more consistent, evidence-based approaches to PSB.
Current Study
Taken together, the reviewed literature underscores that while awareness of PSB in youth has grown, empirical research examining professional perceptions of these cases, especially within coordinated multidisciplinary systems, remains limited. Misperceptions and biases among professionals can have profound implications: they can shape investigative trajectories, determine whether families are offered supportive versus punitive interventions, and influence whether children are labeled as offenders or in need of care. Understanding how key stakeholders conceptualize and respond to PSB is therefore vital for improving practice, policy, and outcomes for affected youth and families.
Across Texas CACs, more than 30,000 unique cases of child sexual abuse were documented annually between 2019 and 2022. Of these, an estimated 26% to 29% of alleged perpetrators were juveniles. However, these statistics likely underestimate the true prevalence of youth-initiated cases, as many incidents go unreported or are inconsistently categorized by CACs and allied systems. Some CACs may handle these cases internally, while others may divert them to mental health or juvenile justice agencies, creating substantial variability in documentation and service delivery. This inconsistency in both reporting and response reflects broader systemic uncertainty regarding how to conceptualize and address child-on-child sexual abuse. Professionals from different disciplines (e.g., law enforcement, mental health, child protection, prosecution, and medical fields) often approach such cases from distinct philosophical and procedural perspectives. These differing frameworks can lead to fragmented responses, inconsistent treatment pathways, and potential secondary harm to children and families. This variability may be further shaped by the geographic contexts in which professionals practice, as urban, suburban, and rural settings differ in their access to training, service infrastructure, and community norms around youth behavior and accountability.
The present study seeks to address this gap by examining how multidisciplinary team (MDT) members across Texas perceive cases of child-on-child sexual abuse involving PSB. Specifically, the study investigates how case characteristics (age and gender of both the initiator and the victim) and respondent characteristics (professional role, political ideology, gender, race, and geographic area of practice) interact to shape professionals’ agreement that the initiating child was old enough to accept responsibility for the situation, old enough to know their behavior was wrong, and deserving of criminal consequences. By empirically testing these variations using an experimental vignette design, this study contributes to a growing body of research on professional decision-making in sensitive child welfare contexts. Ultimately, this work aims to inform the development of training, policy guidance, and coordinated response frameworks that reduce bias, promote trauma-informed care, and support balanced approaches to accountability and rehabilitation for youth exhibiting PSB.
Methods
Participants
The target population included CAC staff and MDT members across Texas. CAC staff encompass a range of professionals affiliated with Children’s Advocacy Centers, including forensic interviewers, victim advocates, and case coordinators. MDT members include professionals from partnering agencies who collaborate on child abuse investigations and response, such as law enforcement, mental health providers, child protective services workers, prosecutors, and medical professionals.
Because MDT participation spans multiple decentralized agencies and no comprehensive sampling frame identifying all CAC-affiliated MDT members in Texas exists, sample representativeness could not be formally evaluated. In addition, neither CAC program reports nor publicly available National Children’s Alliance data include workforce demographic breakdowns at the state level. However, the occupational categories represented in the sample are consistent with MDT composition described in CAC program documentation, and the heterogeneity of the workforce is reflected in the range of professional roles captured. A convenience sample was therefore obtained by soliciting participation through the Texas CAC membership organization. Recruitment materials were distributed by the state membership organization to the executive directors of all 70 CACs in Texas, 23 Statewide MDT Facilitator (SMDTF) members, and a listserv of approximately 220 MDT members. CAC executive directors were asked to forward the survey link to their MDT members. Because survey distribution was handled by CAC personnel rather than the research team directly, the extent to which the subject matter of the survey was described uniformly in recruitment communications is unknown.
A total of 723 individuals consented to participate. Of these, 682 answered the first substantive survey question, 431 answered the second, and 363 completed the full survey. This pattern of early attrition—where the largest drop occurred within the first two questions—may reflect the sensitive nature of the subject matter, uncertainty about survey relevance, or other unobserved factors. Four additional respondents had fully missing demographic data, consistent with dropout prior to reaching that section. The final analytic sample consisted of 363 respondents, with analytic samples ranging from 325 to 326 across models following listwise deletion of cases with missing values on any predictor. This study received institutional review board (IRB) approval by the IRB board at The University of Texas at El Paso.
Descriptive Statistics for Sample (n = 363)
Note. Profession categories reflect recoded values. An additional 19 respondents who initially identified as ‘other’ were recoded based on their described role; 4 whose responses could not be classified were excluded from profession-based analyses (n = 355).
Procedure
The survey instrument was developed in consultation with staff from the state CAC membership organization and members of the research team with prior experience conducting focus groups and qualitative coding with this population. Prior to data collection, the instrument was pilot tested with CAC staff to ensure clarity of language, absence of errors, and feasibility within the specified completion window. The survey was administered online via QuestionPro. Informed consent materials explicitly stated that the study concerned responses to children who exhibit problematic sexual behavior and indicated that participants would read and respond to a vignette involving children who act out sexually on other children. Participants were given two weeks to complete the survey and received one reminder prompt. The average completion time was 19 minutes. Because recruitment emails were distributed through intermediary CAC personnel rather than directly by the research team, we cannot determine the extent to which recruitment communications described the survey content prior to participants opening the consent form; this uncertainty should be considered when interpreting the pattern of early attrition observed. To incentivize participation, respondents were offered entry into a raffle for one of fifty $50 Amazon gift cards. Email addresses used for raffle entry were not linked to survey responses to preserve confidentiality.
Vignettes
To examine how case characteristics shape professional perceptions, participants were randomly assigned to one of eight hypothetical vignettes describing a case of sibling-initiated sexual contact. Vignettes varied systematically by the age and gender of the initiating child—the child described as having acted out sexually toward another child—and the gender of the other child involved, yielding a 2 (initiator age: 10 or 14 years old) × 2 (initiator gender) × 2 (other child gender) design. In all vignettes, the victim was five years younger than the initiator (i.e., age 5 or 9), and no threat or force was described. The five-year age difference between initiator and victim was held constant across all conditions (i.e., the 10-year-old initiator was paired with a 5-year-old victim, and the 14-year-old initiator was paired with a 9-year-old victim) to reflect a developmentally meaningful power differential sufficient to call into question consent and developmental appropriateness, while avoiding an age gap so extreme as to obscure the effects of initiator age and gender on professional judgment. The vignette text read as follows, with experimentally manipulated elements in italics: [Mary/Jason], who is [10/14] years old, is involved in a case investigation after a report was made regarding [her/his] acting out sexually on [her/his] little [sister/brother], who is [5/9] years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
The eight vignette conditions and their corresponding initiator/victim dyads are displayed in Appendix A. Random assignment resulted in modestly unequal cell sizes across conditions (range = 35–52 respondents per vignette), which was addressed analytically using regression modeling. Three binary predictor variables were derived from vignette assignment: age of initiator (1 = 14-year-old, 0 = 10-year-old), male initiator (1 = male, 0 = female), and opposite-sex dyad (1 = initiator and victim of different genders, 0 = same gender).
Measures
Dependent Variables
Three items assessed respondents’ perceptions of the initiating child. Each item was rated on a five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, 5 = strongly agree). Accordingly, mean values above 3 indicate overall agreement with the statement, whereas values below 3 indicate disagreement. This scaling should be considered when interpreting descriptive statistics and model estimates. Responsible was measured by the item “The older child was old enough to accept responsibility for the situation” (M = 3.64, SD = 1.06; 65% of respondents agreed or strongly agreed). Wrong was measured by the item “The older child was old enough to know what they did was wrong” (M = 3.83, SD = 0.95; 71% agreed or strongly agreed). Criminal was measured by the item “The older child should face criminal consequences” (n = 362, M = 2.85, SD = 1.09; 31% agreed or strongly agreed). The distributions of Responsible and Wrong were notably skewed toward agreement, while Criminal showed a slight lean toward disagreement.
Respondent Characteristics
Political ideology was assessed using a five-point item ranging from 1 (very liberal) to 5 (very conservative) and collapsed to three categories for analysis to preserve interpretability and avoid sparse cell counts: liberal, moderate, and conservative, with 16 respondents missing ideology data. Parental status was coded as parent (1) or non-parent (0) and geographic area of practice was coded as urban, rural, or suburban. Respondent gender was coded as female (1) or male (0) and race was coded as White (1) or non-White (0) due to small subgroup sizes that precluded stable estimation across individual gender and racial/ethnic categories; this is a limitation of the analysis.
Analytic Strategy
Three ordinal logistic regression models were estimated, one for each dependent variable, using Stata 18. Ordinal logistic regression was selected because the dependent variables are ordinal rather than continuous, and because regression modeling allowed simultaneous estimation of vignette and respondent predictors without requiring the collapsing of predictor categories. All models included vignette-level predictors, respondent-level predictors, and geographic area of practice, and were estimated with robust standard errors to account for potential heteroskedasticity. Results are reported as odds ratios (ORs) with 95% confidence intervals. Missing data were handled via listwise deletion. The proportional odds assumption—that the effect of each predictor is consistent across all response thresholds—was evaluated using generalized ordered logit models estimated via gologit2 (Williams, 2006) with the parallel lines option, which tests and retains equality constraints for predictors that satisfy the assumption. All 42 equality constraints were retained across all three models, confirming that the proportional odds assumption was satisfied and that the constrained regression estimates are appropriate.
To assess whether professional role added explanatory power beyond individual respondent characteristics, likelihood ratio tests compared each full model (including profession) to a reduced model excluding profession, estimated on identical analytic samples. Likelihood ratio tests were also conducted to assess whether geographic area added explanatory power beyond all other predictors. To examine whether the association between respondent gender and criminal consequences was confounded by professional role, a supplemental model excluding profession was estimated and the gender coefficient compared across nested models.
Results
Descriptive Statistics by Vignette Condition
Note. Means and standard deviations are unadjusted. Response scale: 1 = strongly disagree to 5 = strongly agree; 3 = neither agree nor disagree. Vignette condition indicates the age and gender of the initiating child and the gender of the victim (initiator age/victim gender). Criminal n = 362 due to one missing response; Responsible and Wrong n = 363. Vignette 5 (Male 14/Female 9) n = 43 for Responsible and Wrong, 42 for Criminal. Inferential comparisons are based on the ordinal logistic regression models presented in Table 3.
Ordinal Logistic Regression Models Predicting Professional Perceptions of Problematic Sexual Behavior (N = 325–326)
Note. OR = odds ratio. 95% confidence intervals in brackets. All models estimated with robust standard errors. Reference categories: age of initiator = 10-year-old; initiator gender = female; dyad = same-sex; political ideology = moderate; profession = CAC staff; geographic area = rural. Profession LR χ2(4) and Geography LR χ2(2) test the joint significance of those predictor sets against models estimated on identical samples. Generalized ordered logit sensitivity analyses confirmed the proportional odds assumption was satisfied for all predictors across all three models (all 42 equality constraints retained in each); estimates were numerically identical to the constrained models.
*p < .05 **p < .01 ***p < .001.
Likelihood ratio tests confirmed that professional role added significant explanatory power for Criminal, χ2(4) = 29.82, p < .001, but not for Responsible, χ2(4) = 5.78, p = .217, or Wrong, χ2(4) = 6.26, p = .181, indicating that professional role was more strongly associated with endorsement of criminal consequences than with the two age-related judgment items. Geographic area added significant explanatory power across all three models: Responsible, χ2(2) = 14.67, p < .001; Wrong, χ2(2) = 11.91, p = .003; Criminal, χ2(2) = 16.59, p < .001, indicating that where professionals work adds explanatory power beyond individual-level predictors.
Vignette Characteristics
Age of Initiator
Age of the initiating child was the strongest and most consistent predictor across all three models. Compared to respondents assigned a 10-year-old initiator, those assigned a 14-year-old initiator were significantly more likely to agree the child was old enough to accept responsibility (OR = 7.51, 95% CI [4.70, 12.01], p < .001), to know that their behavior was wrong (OR = 4.30, 95% CI [2.71, 6.83], p < .001), and should face criminal consequences (OR = 5.99, 95% CI [3.80, 9.44], p < .001). The magnitude of these effects was substantially larger than any other predictor in the models. Gender of initiator. Gender of the initiating child was not a significant predictor of any outcome after controlling for other vignette and respondent characteristics. Dyad composition. Opposite-sex dyads were significantly more likely than same-sex dyads to be perceived as knowing their behavior was wrong (OR = 1.70, 95% CI [1.10, 2.63], p = .018). Dyad composition was not significantly associated with judgments of responsibility or criminal consequences.
Respondent Characteristics
Political Ideology
Compared to moderate respondents, conservative respondents were significantly more likely to agree the child was old enough to accept responsibility (OR = 1.76, 95% CI [1.06, 2.92], p = .030) and to know their behavior was wrong (OR = 1.85, 95% CI [1.13, 3.03], p = .014), but did not differ significantly from moderate respondents in their agreement that the child should face criminal consequences. The pattern was reversed for liberal respondents: compared to moderate respondents, liberal respondents did not differ significantly in judgments of responsibility or whether the child knew their behavior was wrong, but were significantly less likely to agree the child should face criminal consequences (OR = 0.52, 95% CI [0.28, 0.95], p = .033). This asymmetric pattern suggests that ideological differences among professionals may be more strongly related to agreement that the child should face criminal consequences than to judgments of responsibility or if the child knew the behavior was wrong.
Professional Role
Court/legal professionals were consistently more likely than CAC staff to agree with all three outcomes: responsibility (OR = 2.24, 95% CI [1.23, 4.10], p = .008), whether the child knew their behavior was wrong (OR = 2.19, 95% CI [1.21, 3.99], p = .010), and criminal consequences (OR = 2.52, 95% CI [1.39, 4.57], p = .002). Mental health professionals did not differ significantly from CAC staff on responsibility or whether the child knew their behavior was wrong but were substantially less likely to agree the child should face criminal consequences (OR = 0.35, 95% CI [0.18, 0.67], p = .002). DFPS professionals did not differ significantly from CAC staff on responsibility or knowledge of wrongdoing but were significantly more likely to agree the child should face criminal consequences (OR = 1.93, 95% CI [1.04, 3.57], p = .036). Medical professionals did not differ significantly from CAC staff on any outcome.
Demographic Characteristics
Female respondents were significantly more likely than male respondents to agree the child was old enough to accept responsibility (OR = 2.00, 95% CI [1.16, 3.44], p = .012) and to know their behavior was wrong (OR = 1.89, 95% CI [1.14, 3.13], p = .013). However, respondent gender was not significantly associated with whether the child should face criminal consequences. A supplemental model excluding professional role indicated that the association between respondent gender and agreement that the child should face criminal consequences may be partly attributable to differences in professional-role composition: female respondents are overrepresented in mental health and CAC roles, which in this sample were associated with lower agreement that the child should face criminal consequences. Once professional role is controlled, respondent gender has no independent effect on agreement that the child should face criminal consequences. Lastly, White respondents were significantly less likely than non-White respondents to agree the child should face criminal consequences (OR = 0.59, 95% CI [0.37, 0.94], p = .027), while race was not significantly associated with responsibility or knowledge judgments. Respondent parental status was not significantly associated with any outcome.
Geographic Area
Compared to rural respondents, suburban respondents were significantly more likely to agree the child was old enough to accept responsibility (OR = 2.28, 95% CI [1.27, 4.09], p = .006) and to know their behavior was wrong (OR = 2.21, 95% CI [1.26, 3.86], p = .005), with a similar but non-significant trend for criminal consequences (OR = 1.63, 95% CI [0.97, 2.72], p = .062). Urban respondents did not differ significantly from rural respondents on responsibility or knowledge of wrongdoing but were significantly less likely to agree the child should face criminal consequences (OR = 0.50, 95% CI [0.30, 0.85], p = .011). These findings suggest that professional practice context—beyond individual ideology, role, and demographics—adds explanatory power for judgments about PSB cases.
Discussion
This study examined how professional and situational factors shape perceptions of children displaying problematic sexual behavior (PSB). Using an experimental vignette design and ordinal logistic regression, findings indicate that judgments regarding responsibility, knowledge of wrongdoing, and criminal consequences vary systematically by characteristics of both the case and the respondent. Age of the initiating child, political ideology, professional role, and geographic context were particularly salient, underscoring that professional responses are shaped not only by case features but also by broader disciplinary and sociocultural contexts (Chaffin, 2008; Letourneau et al., 2009).
Vignette Characteristics
Consistent with prior research (Burt & Estep, 1981; Rogers & Davies, 2007), age of the initiating child was the strongest predictor across all outcomes. Respondents were more likely to attribute responsibility, awareness of wrongdoing, and support for criminal consequences to a 14-year-old compared to a 10-year-old, reflecting developmental assumptions about increasing cognitive and moral capacity (Casey et al., 2011). Notably, agreement that a child understands wrongdoing did not uniformly translate into support for criminal consequences, suggesting these judgments are conceptually distinct.
Gender of the initiating child was not significantly associated with any outcome after controlling for other factors. Although prior research has documented gender-based differences in perceptions of sexual offending (Fromuth et al., 2002; Howell et al., 2011), null findings here suggest such biases may be attenuated in structured professional evaluations or may operate earlier in the case process—during recognition or reporting—rather than within formal assessment contexts (Sahlstrom & Jeglic, 2008). Opposite-sex dyads were associated with greater agreement that the child knew the behavior was wrong, though not with responsibility or criminal consequences, possibly reflecting heteronormative assumptions about sexual behavior.
Respondent Characteristics
Political ideology was differentially associated with outcomes. Conservative respondents were more likely to attribute responsibility and knowledge of wrongdoing, whereas liberal respondents were less likely to endorse criminal consequences—demonstrating that ideological differences are more strongly tied to punitive responses than to perceptions of developmental capacity (Salerno et al., 2010).Professional role was most strongly associated with criminal consequences. Mental health professionals were less likely, and court/legal professionals more likely, to endorse criminal consequences, consistent with differing institutional mandates (Allen, 2017; Masson & Hackett, 2004). Differences in responsibility and wrongdoing judgments were comparatively smaller, suggesting greater cross-disciplinary agreement on recognizing behavior than on determining consequences. Apparent gender differences among respondents were attenuated after accounting for professional role, suggesting workforce composition may explain observed effects and that gender-based biases may operate earlier in case processing—during reporting or referral—rather than within structured evaluations. Lastly, geographic context remained significant across models even after controlling for ideology and profession. Suburban respondents were more likely to attribute responsibility and awareness, while urban respondents were less likely to endorse criminal consequences, suggesting professional judgments are shaped by local environments through mechanisms that warrant further investigation.
Policy and Practice Implications
The variation documented here is patterned by profession, ideology, and geography in ways that point to specific intervention targets. The divergence between court/legal and mental health professionals is particularly instructive. Court professionals were substantially more likely to endorse criminal consequences—not necessarily because of bias, but because legal frameworks assign accountability in ways that may not adequately distinguish child initiators from adult offenders. Where statutes treat PSB primarily as a criminal matter, they may constrain professional discretion in ways misaligned with the evidence base. Legislative reform that codifies diversion, deferred prosecution, and treatment referral as preferred responses—rather than formal adjudication—may be the most direct lever for reducing this variation (Letourneau et al., 2009, 2018; Borduin et al., 2021). Public health frameworks offer an additional model, prioritizing early identification, prevention, and family support over reactive legal intervention, with promising implementation in countries including Australia (New South Wales Ministry of Health, 2022). Training that helps court professionals understand recidivism evidence (Caldwell, 2016; Lussier & McCuish, 2024) and the range of legally available responses may further reduce over-reliance on formal criminal processing.
The MDT model represents the most promising structural mechanism for bridging these professional divides. CAC-based MDTs bring together forensic interviewers, law enforcement, prosecutors, child welfare investigators, and clinicians to coordinate responses reflecting both legal accountability and child well-being (Theimer et al., 2023). When functioning effectively, the MDT channels professional differences productively—so that legal expertise and developmental knowledge inform a shared response rather than producing inconsistent outcomes depending on who has first contact. Geographic variation in judgments suggests this coordination may be especially important in rural and suburban contexts where informal professional networks may carry more influence than standardized protocols.
Training is a necessary complement. Addressing ideological variation requires explicit engagement with myths driving punitive responses—including overestimation of recidivism risk and conflation of PSB with adult offending (Chaffin & Bonner, 1998; Fix et al., 2023). Shared cross-disciplinary training, rather than profession-specific continuing education, may reduce the siloing that allows divergent frameworks to persist. Consistent adoption of non-stigmatizing, behaviorally descriptive language across systems—PSB rather than sexual offending, initiating child rather than perpetrator—further shapes how professionals interpret behavior and assign consequences, promoting more developmentally appropriate responses.
Limitations, Future Directions, and Conclusions
Several limitations should guide interpretation. First, the convenience sample was drawn only from Texas MDT members, limiting generalizability to jurisdictions with different policies, training, and community norms around PSB. Second, participant attrition raises the possibility of self-selection bias. Although the consent materials disclosed that the study concerned professional responses to children displaying PSB, the early drop-off suggests some respondents may have been uncomfortable with the topic or vignette content. Third, the self-report vignette design limits ecological validity; actual investigative, referral, or treatment decisions may differ from stated perceptions. Finally, the study could not fully examine intersectional factors such as race, ethnicity, socioeconomic status, culture, or religion. Given concerns about racialized assumptions regarding criminality and sexuality (Fix et al., 2023), future research should incorporate race as a vignette-level manipulation and examine how these factors shape professional judgment across diverse MDTs.
Addressing misconceptions and biases is essential to ensuring that children who exhibit PSB receive rehabilitative rather than punitive responses. A trauma-informed, developmentally grounded approach recognizes that these children are still developing, and that punitive measures may reinforce stigma, limit engagement in treatment, and worsen outcomes. Coordinated responses that balance accountability with care can better support both the child exhibiting PSB and the child who was harmed. At the system level, the findings highlight the need for stronger alignment across child welfare, legal, and clinical systems. Shared training, clearer communication, and consistent language may reduce fragmentation and improve case handling across disciplines. Judgments of responsibility and consequences are influenced not only by case characteristics but also by respondent characteristics such as professional role and political ideology. These findings underscore the importance of continued efforts to align professional responses with developmental science while addressing the contextual factors that shape decision-making. Ultimately, this research contributes to ongoing efforts to reframe how professionals understand children who engage in PSB. Viewing PSB as a developmental and behavioral concern, rather than solely a criminal issue, supports more consistent, evidence-based, and compassionate responses.
Footnotes
Ethical Considerations
This study was approved by the institutional review board at the University of Texas at El Paso on 2/23/2023 (IRB protocol # 2025976-2).
Consent to Participate
All participants provided written informed consent to participate in the study.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by an award from the Administration for Children and Families (ACF), U.S. Department of Health and Human Services. The contents are solely the responsibility of the authors and do not necessarily represent the official views of the Administration for Children and Families, the state membership organization, or the U.S. Government.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Appendix
Display of Vignette Conditions
Vignette
Conditions
Initiator/Victim Dyad
Vignette 1
Mary, who is 14 years old, is involved in a case investigation after a report was made regarding her acting out sexually on her little sister, who is 9 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Female 14/Female 9
Vignette 2
Mary, who is 10 years old, is involved in a case investigation after a report was made regarding her acting out sexually on her little sister, who is 5 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Female 10/Female 5
Vignette 3
Mary, who is 14 years old, is involved in a case investigation after a report was made regarding her acting out sexually on her little brother, who is 9 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Female 14/Male 9
Vignette 4
Mary, who is 10 years old, is involved in a case investigation after a report was made regarding her acting out sexually on her little brother, who is 5 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Female 10/Male 5
Vignette 5
Jason, who is 14 years old, is involved in a case investigation after a report was made regarding him acting out sexually on his little sister, who is 9 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Male 14/Female 9
Vignette 6
Jason, who is 10 years old, is involved in a case investigation after a report was made regarding him acting out sexually on his little sister, who is 5 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Male 10/Female 5
Vignette 7
Jason, who is 14 years old, is involved in a case investigation after a report was made regarding him acting out sexually on his little brother, who is 9 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Male 14/Male 9
Vignette 8
Jason, who is 10 years old, is involved in a case investigation after a report was made regarding him acting out sexually on his little brother, who is 5 years old. Child confirmed in forensic interview. There was no evidence of threat or force involved.
Male 10/Male 5