When Is Maltreatment Confirmed? Using Unsupervised Machine Learning to Advance Understanding of Substantiated Child Maltreatment Report Dispositions
When Is Maltreatment Confirmed? Using Unsupervised Machine Learning to Advance Understanding of Substantiated Child Maltreatment Report Dispositions
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152135
- ISBN
- 9798383692103
- DDC
- 320
- 서명/저자
- When Is Maltreatment Confirmed? Using Unsupervised Machine Learning to Advance Understanding of Substantiated Child Maltreatment Report Dispositions
- 발행사항
- [Sl] : The University of North Carolina at Chapel Hill, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 127 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-02, Section: B.
- 주기사항
- Advisor: Putnam-Hornstein, Emily.
- 학위논문주기
- Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2024.
- 초록/해제
- 요약Policymakers rely on sources of administrative child protection system (CPS) data to inform policies that prevent harm to children and support families. One frequently used indicator contained in these sources is whether maltreatment allegations have been substantiated, or confirmed, by a CPS agency. For decades, there has been debate about how we should translate and use data about substantiation. In addition to case facts, the decision to substantiate likely reflects the policy and practice of CPS agencies. For example, substantiation may be required to refer families to services. It also may function as a pathway to certain outcomes, such as mandated court supervision, removal of a child, and placement of a caregiver on a central registry of child maltreatment perpetrators. In this three-paper dissertation, I use a data-driven approach and a source of population-based data from California. In paper I, I calculate rates of substantiated child maltreatment reports across county-years. I then test organizational theories by examining the association between county-level agency and substantiation rates. I expand on previous studies by adjusting for a more robust set of community characteristics and examining the relationship over several years. In paper II, I use model-based clustering to categorize substantiated reports into distinct, underlying typologies based on child and report characteristics. In paper III, I examine the distribution of clusters identified in paper II and the extent to which these clusters are explained by county systems. To explore why some counties have a higher rate of reports from a particular cluster, I employ regression analyses with substantiation typology as the outcome, adjusting for annual investigation count as well as report and community characteristics. Taken together, results of the papers demonstrate that substantiation rates are indicative of organizational context and that there are numerous typologies of substantiated cases. Furthermore, county jurisdictions vary in their propensities to substantiate typologies of cases, beyond their propensity to substantiate an average case. Given that substantiation operates as a pre-requisite to services and sanctions, research must consider how to ensure administrative indicators reflect sufficient dimensions of risk and need such that they are useful in identifying appropriate, effective responses.
- 일반주제명
- Public policy
- 일반주제명
- Social work
- 일반주제명
- Public health
- 일반주제명
- Statistics
- 키워드
- Child welfare
- 키워드
- Decision-making
- 키워드
- Family services
- 키워드
- Substantiation
- 기타저자
- The University of North Carolina at Chapel Hill Social Work
- 기본자료저록
- Dissertations Abstracts International. 86-02B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017163100
■00520250211152135
■006m o d
■007cr#unu||||||||
■020 ▼a9798383692103
■035 ▼a(MiAaPQ)AAI31484061
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a320
■1001 ▼aMcNellan, Claire Ruth.
■24510▼aWhen Is Maltreatment Confirmed? Using Unsupervised Machine Learning to Advance Understanding of Substantiated Child Maltreatment Report Dispositions
■260 ▼a[Sl]▼bThe University of North Carolina at Chapel Hill▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a127 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-02, Section: B.
■500 ▼aAdvisor: Putnam-Hornstein, Emily.
■5021 ▼aThesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2024.
■520 ▼aPolicymakers rely on sources of administrative child protection system (CPS) data to inform policies that prevent harm to children and support families. One frequently used indicator contained in these sources is whether maltreatment allegations have been substantiated, or confirmed, by a CPS agency. For decades, there has been debate about how we should translate and use data about substantiation. In addition to case facts, the decision to substantiate likely reflects the policy and practice of CPS agencies. For example, substantiation may be required to refer families to services. It also may function as a pathway to certain outcomes, such as mandated court supervision, removal of a child, and placement of a caregiver on a central registry of child maltreatment perpetrators. In this three-paper dissertation, I use a data-driven approach and a source of population-based data from California. In paper I, I calculate rates of substantiated child maltreatment reports across county-years. I then test organizational theories by examining the association between county-level agency and substantiation rates. I expand on previous studies by adjusting for a more robust set of community characteristics and examining the relationship over several years. In paper II, I use model-based clustering to categorize substantiated reports into distinct, underlying typologies based on child and report characteristics. In paper III, I examine the distribution of clusters identified in paper II and the extent to which these clusters are explained by county systems. To explore why some counties have a higher rate of reports from a particular cluster, I employ regression analyses with substantiation typology as the outcome, adjusting for annual investigation count as well as report and community characteristics. Taken together, results of the papers demonstrate that substantiation rates are indicative of organizational context and that there are numerous typologies of substantiated cases. Furthermore, county jurisdictions vary in their propensities to substantiate typologies of cases, beyond their propensity to substantiate an average case. Given that substantiation operates as a pre-requisite to services and sanctions, research must consider how to ensure administrative indicators reflect sufficient dimensions of risk and need such that they are useful in identifying appropriate, effective responses.
■590 ▼aSchool code: 0153.
■650 4▼aPublic policy
■650 4▼aSocial work
■650 4▼aPublic health
■650 4▼aStatistics
■650 4▼aIndividual & family studies
■653 ▼aChild maltreatment
■653 ▼aChild protection system
■653 ▼aChild welfare
■653 ▼aDecision-making
■653 ▼aFamily services
■653 ▼aSubstantiation
■690 ▼a0630
■690 ▼a0452
■690 ▼a0573
■690 ▼a0628
■690 ▼a0463
■71020▼aThe University of North Carolina at Chapel Hill▼bSocial Work.
■7730 ▼tDissertations Abstracts International▼g86-02B.
■790 ▼a0153
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17163100▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
