Decision Making in Child Protection: An International Comparative Study

Research Overview

This research article, supported by the Haruv Institute, Jerusalem, explores decision-making processes among child welfare professionals across different countries utilizing the Judgments and Decision Processes in Context model (JUDPIC).

Introduction

Child welfare professionals make critical decisions affecting children and families, such as removing a child from home or reunifying them with their biological family. These decisions have both short and long-term impacts. The study uses the JUDPIC model to understand factors associated with these decisions in cases of alleged child maltreatment, comparing practitioners from Israel, the Netherlands, Northern Ireland, and Spain. Decisions are often made under pressure, with limited resources and ambiguous information. The JUDPIC model suggests that professionals' judgments are based on case information, personal characteristics, and agency features, all within broader ecological, organizational, and cultural contexts.

Key Factors

  • Case characteristics (e.g., physical signs of abuse, parents’ explanations)

  • Professional characteristics (e.g., personal experiences, attitudes toward child removal)

  • Agency features (e.g., placement policies)

  • Ecological context of the family

  • Organizational context of the decision-making agency

  • Cultural contexts (e.g., public attitudes toward the protective system)

Prior Research

Studies show that case characteristics like ethnicity and poverty are associated with maltreatment substantiation and placement decisions. Practitioner characteristics, such as age, experience, and personal history, also influence decisions. Contextual factors, including country-specific welfare regimes and historical events, play a significant role.

  • Rivaux et al. (2008) & Dettlaff et al. (2011): Case characteristics (ethnicity, poverty, risk) are associated with maltreatment substantiation and placement decisions.

  • Drake et al. (2011) & Fallon et al. (2013): Practitioner decisions are associated with the child’s race.

  • Brunnberg & Pecnik (2007): Younger, less experienced, and childless workers perceive higher risk and are more likely to recommend placing the child in care.

  • Davidson-Arad et al. (2003): More experienced workers are less prone to implement removal recommendations.

  • Regehr et al. (2010): Workers who experienced previous traumas were less likely to assess a child as being at risk.

  • Jent et al. (2011): White and more educated workers were more inclined than others to classify physical injuries as abuse.

  • Gold et al. (2001): Demonstrated the impact of the larger country context by comparing decisions in Canada and Israel.

  • Brunnberg & Pecnik (2007): Croatian social workers were more likely than their Swedish peers to assess a situation as requiring child protection and to favor removal.

  • Gilbert, Parton, & Skivenes (2011): Between-country differences reflect variations in their welfare regimes.

Impact of Historical Events

Dramatic events can significantly impact protective practices. For example, the death of Peter Connolly in England in 2008 led to a 50% rise in applications to courts for Care Orders. Similarly, the Savannah case in the Netherlands led to increased public scrutiny and a refocusing on child protection in Germany after media coverage of child deaths.

Study Details

The study uses the JUDPIC model to compare judgments and decisions in cases of alleged maltreatment in Israel, the Netherlands, Northern Ireland, and Spain, examining the relationships between case characteristics (mother’s wish regarding removal), practitioner characteristics (child welfare attitudes), protective system context (the four countries), and dependent factors (substantiation of maltreatment, risk assessment, intervention recommendation).

Method

The study assessed a case vignette of alleged child maltreatment by professionals responsible for case assessments and recommendations in the four countries.

Vignette

The vignette describes a family with three young children, where a teacher reports concerns about Dana (7), who has unexplained physical marks and outbursts. The family lacks community ties and faces financial struggles. Participants received one of two versions of the vignette, varying the mother’s reaction to potential removal.

Sample and Procedure

The sample included 828 practitioners from the four countries, recruited to represent the decision-makers in their respective systems. Data were collected through questionnaires distributed at conferences, team meetings, and via email.

Measures

  • Personal and Professional Background: Demographic information.

  • Attitudes: The Child Welfare Attitudes Questionnaire, covering six content areas, assessed attitudes toward:

    • Removal from home of children at risk (overall alpha = 0.790

    • Reunification and optimal duration of alternative care (overall alpha = 0.735

    • Children’s participation in decisions (overall alpha = 0.740

    • Parents’ participation in decisions (overall alpha = 0.643

    • Foster care's ability to promote children’s development (overall alpha = 0.449

    • Residential care's ability to promote children’s development (overall alpha = 0.700

  • Maltreatment Substantiation: Assessed on a five-point scale for emotional neglect, physical neglect, emotional abuse, physical abuse, and sexual abuse.

  • Risk Assessments: Assessed the level of risk of physical and emotional harm to the child on a five-point scale.

  • Intervention Decisions: Participants recommended an intervention from six options, ranging from no intervention to foster placement without parental consent.

Analysis

Multivariate analyses of variance compared child welfare attitudes across countries, followed by one-way analyses of variance with post hoc comparisons. Cluster analysis identified subsets of participants with distinct child welfare attitudes. Multivariate analyses examined the relationship of substantiation of maltreatment and risk assessments as dependent variables. Chi-square tests examined differences in intervention recommendations.

Ethics

Participation was anonymous and voluntary, with procedures reviewed and approved by ethical boards.

Findings

The study examined the relationships between case characteristics, practitioner characteristics, and protective system context. The mother’s wishes regarding removal did not significantly impact judgments or decisions. Practitioner attitudes and country context had significant effects.

Child Welfare Attitudes Across Countries

Significant differences in child welfare attitudes existed between countries (F(18, 2463) = 34.82, p < 0.001). Northern Ireland practitioners opposed removal and favored reunification, while Spanish practitioners had opposing views. Dutch practitioners were similar to Northern Ireland in opposing removal but had lower support for reunification.

Northern Ireland practitioners strongly advocated for children’s and parents’ participation in decision-making, while Spanish, Israeli, and Dutch practitioners were less supportive of children’s participation. Spanish professionals had the least negative view of foster care, while Israelis had the least negative view of residential care.

Cluster Analysis of Attitudes

Two clusters were identified:

  • Against Removal: Strong attitudes against removal, favoring short/optimal placement duration, and negative views toward foster and residential care.

  • More Pro Removal: Less against removal, with more favorable attitudes toward foster care and residential treatment.

Countries differed in the frequency of each cluster ($\chi^2(3) = 77.18, p < 0.001). Northern Ireland had the largest group against removal (75.7%).

Maltreatment Substantiation

Significant differences existed between countries (F(15, 2292) = 16.81, p < 0.001) and between attitude clusters (F(5, 762) = 2.92, p < 0.05). Substantiation tended to be lower among practitioners against removal, except in Northern Ireland. Dutch professionals substantiated alleged maltreatment less than Israeli and Northern Ireland professionals.

Risk Assessments

Risk assessments were associated with country (F(6, 1620) = 4.16, p < 0.001) and attitude cluster (F(2, 809) = 6.21, p < 0.01). Practitioners against removal had lower risk assessments. Dutch practitioners made lower assessments regarding risk for physical harm compared to Spanish practitioners and lower assessments of risk for emotional harm compared to their Northern Ireland colleagues.

Intervention Recommendation

Significant differences existed between countries ($\chi^2(9) = 122.32, p < 0.001). Spanish practitioners were divided, with some recommending no additional services and others recommending removal. Dutch and Israeli practitioners recommended removal less frequently than Spanish practitioners.

Attitudes were associated with recommendations ($\chi^2(3) = 34.80, p < 0.001$$). Practitioners against removal recommended fewer placements.

Discussion

The study supports the JUDPIC model, highlighting the impact of practitioner attitudes and country context on child protection decisions. The lack of impact of the mother’s wishes is concerning, suggesting a need for training and policy guidelines. Practitioner attitudes significantly impact judgments and decisions.

Key Points

  • Client characteristics (mother's wish) did not impact decisions.

  • Practitioner characteristics (attitudes) significantly influenced judgments and decisions.

  • Country context significantly impacted all study variables.

Country-Specific Findings

Israel

Israeli practitioners had the lowest removal recommendations, consistent with previous findings. This reflects a cultural and professional stance favoring family preservation.

Spain

Spanish practitioners were less against removal, aligning with a tradition of out-of-home placement, particularly residential care.

Northern Ireland

Northern Ireland practitioners showed strong attitudes against removal but substantiated allegations more and recommended removal more than Israeli and Dutch practitioners. This reflects concerns about the public care system and the influence of human rights legislation.

Netherlands

Dutch practitioners, similar to Northern Ireland, had strong attitudes against removal, reflecting a family service orientation. They substantiated alleged maltreatment less than others.

Limitations

  • Vignette studies create an artificial decision environment.

  • Potential for cultural and language variations in interpreting the vignette.

  • Use of convenience sampling limits generalizations.

Implications for Practice and Policy

Practitioners and policymakers should be aware of how attitudes and country context impact decisions. Training should increase understanding of these influences, and policies should align with desired outcomes. Further international studies are needed to explore the formation of attitudes and their impact on judgments and decisions.