Differential Response in Child Protective Services: A Literature Review Version 2 A Project of the Children’s Bureau US DHHS, ACF, ACYF CFDA NUMBER: 93:670 2008-2013 November 2011 This literature review was funded by a cooperative agreement with the Children’s Bureau, U.S. Department of Health and Human Services as part of the Quality Improvement Center on Differential Response in Child Protective Services (QIC-DR) This product expresses the views of the authors, not the views of the Children’s Bureau. Differential Response in Child Protective Services: A Literature Review November 2011 Table of Contents I. INTRODUCTION ..................................................................................................................................... 1 What Is Differential Response? ................................................................................................................ 1 The Scope of the Literature Review ......................................................................................................... 3 II. HISTORY OF DIFFERENTIAL RESPONSE ........................................................................................ 4 Development of Differential Response..................................................................................................... 4 Contributing Events in the History of Differential Response ................................................................... 8 III. IMPLEMENTATION OF DIFFERENTIAL RESPONSE................................................................... 12 Core Elements......................................................................................................................................... 12 Implementation Barriers and Strategies .................................................................................................. 20 IV. SYNTHESIS OF DIFFERENTIAL RESPONSE EVALUATIONS.................................................... 23 Evaluation Design, Methodology, and Results ....................................................................................... 25 Summary of Differential Response Evaluations ..................................................................................... 37 V. DISCUSSION AND CONCLUSIONS.................................................................................................. 39 What Has Been Accomplished?.............................................................................................................. 39 What Is Known? ..................................................................................................................................... 39 Where Are the Gaps? .............................................................................................................................. 40 Conclusion .............................................................................................................................................. 42 REFERENCES ........................................................................................................................................... 43 Differential Response in Child Protective Services: A Literature Review November 2011 Note: This Literature Review has been updated since it was first published in June 2009 to include the latest information and research available about differential response. I. INTRODUCTION Differential response has emerged as part of a larger movement in child welfare that emphasizes family-focused, strengths-based approaches to supporting child and family well-being. Over the past 18 years, many child welfare administrators and others interested in the delivery of child welfare services have become increasingly concerned that an investigative response to child maltreatment reports was inflexible and adversarial and did not provide sufficient services to meet family needs. Some have also been concerned that investigation is usually an intrusive intervention, and one that is not responsive to families’ circumstances. The inability of families to access appropriate services or to accept offered services has also been linked to additional occurrences of maltreatment (Waldfogel, 1998). A central tenet of differential response is that many children and families brought to the attention of child protective services (CPS) can be better served using a supportive and collaborative approach that is free of the constraints and stigma of an investigation (Huebner, Durbin, & Brock, 2009). Therefore, many state and local child protective services systems have been redesigned to include a differential response approach. This literature review presents a brief history and description of the practices that define differential response, as well as findings from several evaluation studies. Peer-reviewed articles, project reports, and conference presentations have also been reviewed, but policy manuals and other differential response-related jurisdictional documents have not been included. What Is Differential Response? By redesigning the ways in which CPS can respond to screened-in reports alleging child maltreatment, differential response approaches create more flexibility for agencies and their staff. Differential response has also been referred to as dual track, multi-track, and multiple response system or MRS. Differential response-organized CPS systems typically have two pathways to serve families: 1) an investigation pathway and 2) a non-investigation pathway. The non-investigation pathway has also been called alternative response, family assessment response, and similar titles by varying jurisdictions. In differential response-organized CPS systems, the investigation pathway is usually restricted to those referrals in which the child has been severely maltreated, or in which it appears that there is imminent risk for further abuse. In general, cases in which there appears to be a potential for involvement of the judicial system, such as serious physical abuse or sexual abuse, are assigned to this pathway. The investigation pathway typically requires the collection of forensic evidence in order to substantiate the maltreatment allegation. Jurisdictions may also decide to screen other types of cases, such as those involving institutional abuse or children with certain disabilities, into the investigation pathway. The non-investigation pathway is usually identified for cases that have been initially classified as low or moderate risk. A key characteristic of the non-investigation pathway is that it focuses on engaging the family in an assessment process to determine what they need to provide for the Version 2 1 Differential Response in Child Protective Services: A Literature Review November 2011 children’s safety and well-being, and the provision of services to meet those needs. It usually does not include a formal decision as to whether the specific allegations were founded or not, and the family caregivers are not classified as perpetrators, since no maltreatment is substantiated. Some CPS systems include a pathway that provides referrals for community-based services for families whose reports have been screened out by the agency (Child Welfare Information Gateway, 2008), although this is not classified as a differential response pathway in this literature review. Based on a review of existing state differential response approaches, Merkel-Holguin, Kaplan, and Kwak (2006) identified core components of differential response. These included: 1) The use of two or more discrete responses of intervention; 2) The creation of multiple responses for reports of maltreatment that are screened in and accepted for response; 3) The determination of the response assignment by the presence of imminent danger, level of risk, and existing legal requirements; 4) The capacity to re-assign families to a different pathway in response to findings from initial investigation or assessment (e.g., a family in the alternative response pathway could be re-assigned to the investigation pathway if the level of risk of the child is found to be higher than originally thought); 5) The establishment of multiple responses is codified in statute, policy, and/or protocols; 6) Families in the assessment pathway may refuse services without consequence as long as child safety is not compromised; 7) No formal determination of maltreatment for families in an assessment pathway, and services offered to such families without any such determination; and 8) No listing of a person in an assessment pathway as a child maltreatment perpetrator in the state’s central registry. In related work, Kaplan and Merkel-Holguin (2008) derived a set of core values common to what is, in practice, a diverse set of approaches. They found that, for the non-investigation pathway, there was a focus on:       Family engagement versus an adversarial approach; Services versus surveillance; Labeling as “in need of services/support” versus “perpetrator”; Being encouraging with families versus threatening; Identification of needs versus punishment; and A continuum of response versus “one size fits all.” Family engagement is often used as an umbrella label for these values and represents a significant change for many CPS agencies. These core values hold that families will be more receptive to services when they feel less under CPS surveillance and experience less stigma, and they can engage more positively in change when they are identified as having both strengths and needs rather than as being perpetrators. The differential response approach also recognizes all Version 2 2 Differential Response in Child Protective Services: A Literature Review November 2011 families’ unique strengths and needs, and addresses these in an individualized manner rather than with a “one size fits all” approach. The Scope of the Literature Review The goal of this review is to provide a synthesis of the existing literature regarding differential response. Articles, reports, book chapters, and other documents that present descriptive and evaluative information on child welfare policies, models, structures, protocols, and practices related to differential response were assembled from relevant public and organizational websites, researchers and practitioners in the field, and the authors themselves, as appropriate. Bibliographies and reference lists of documents received were then scanned for further references, and those documents were obtained as well. Slide show presentations from relevant conferences were also included when they provided substantive data or other information not available elsewhere. Using this process, a bibliography containing more than 127 reports, articles, books, and presentations on differential response was generated. The subset of this literature containing evaluation findings for differential response approaches was synthesized and incorporated into this review. Part of the first phase of the review was the preparation of a structured summary for each document. These summaries are included in an annotated bibliography posted online at www.differentialresponseqic.org. Child welfare and CPS-related literature indirectly related to differential response provided valuable context and support for the documents included in this review, but are not directly addressed here, in the interest of maintaining the focus on differential response. Version 2 3 Differential Response in Child Protective Services: A Literature Review November 2011 II. HISTORY OF DIFFERENTIAL RESPONSE Child welfare, as an area of public policy, began in the late 19th and early 20th centuries, with child safety at home and in the workplace as its primary focus (O’Neill Murray & Gesiriech, 2008). In the United States and other parts of the world, during the first half of the 20th century, society’s views on children changed dramatically. Initially considered the property of their families and precluded from governmental intervention, children became recognized as in need of protection from parental maltreatment. Protection primarily included removal of children from their parents’ care. In the late 1970s and early 1980s, a deeper understanding of parent-child attachment increasingly influenced child welfare policy and practice (Bowlby, 1969). Strategies for preventing and ameliorating child maltreatment shifted toward empowerment of the family and engagement of the family’s strengths as resources for the child. At the same time, courts began to play a larger role in situations in which the governmental agency undertook legal custody of a child or ordered services. Because agencies needed to show significant evidence of maltreatment when seeking legal custody or ordering services, CPS began to focus on obtaining evidence that would be accepted by the court (Hardin, 1996). While parents are still legally and socially responsible for their children’s care and safety, governments legally mandate that state, local, and tribal public agencies create child protective services to respond to allegations of harm and protect children. In carrying out its mandate, the CPS system must balance the protection of children’s physical and emotional safety with recognition of families’ unique strengths and needs. The CPS system must tailor a response to families that builds on their capacities in meeting their safety and service needs. Many practices, protocols, and policies have been developed, implemented, and evaluated, in an attempt to simultaneously meet the unique needs of each child and family involved in the child protection system and achieve consistently high standards for system performance. One of the more recent approaches is differential response. First introduced in the early 1990s, differential response has spread nationally and internationally in an effort to address the growing recognition that families’ differing circumstances and needs necessitate distinct responses. While having common philosophical and strategic foundations, various differential response approaches have developed at state and county levels, with unique trajectories and some variation in motivations, purposes, and impetuses. This chapter reviews the chronological development of various differential response approaches. It then explores some of the critical events that provided the backdrop for their development. Development of Differential Response In 1993, Florida and Missouri each passed legislation to make their CPS systems more responsive to varying levels of risk and the unique service needs of families and children reported to the agency. This new approach was intended to be efficient and effective in addressing maltreatment while minimizing caseworker burnout and turnover. The Family Assessment Response System (FARS) in Missouri was initially piloted in 14 counties, and implemented statewide by 1999 (Loman & Siegel, 2004a; Siegel & Loman, 1997; Siegel & Loman, 2000). This approach has served as a model for differential response in other states. Florida also initiated differential response statewide in 1993, but a 1996 evaluation found its implementation to be inconsistent across districts. In 1998, the mandating legislation was eliminated and differential response was discontinued in the state. A subsequent six month pilot Version 2 4 Differential Response in Child Protective Services: A Literature Review November 2011 project in three Florida counties in 2008 yielded somewhat positive results but also made recommendations for changes (Florida Department of Children and Families, 2009). A statewide workgroup was proposed in 2011 to re-establish alternative response and propose changes to the model. Other states that introduced differential response in the l990s included Oklahoma, Virginia, and Washington. In 1998, the Oklahoma Department of Human Services implemented a statewide non-investigation pathway as part of the CPS system (Zielewski, Macomber, Bess, & Murray, 2006). Virginia piloted differential response in five local departments from 1997 to 1999, and implemented the approach statewide in 2002 (Virginia Department of Social Services [VDSS], 2007). The Washington state legislature funded statewide implementation of its Alternative Response System in 1997 and the system has since been integrated into a broader child welfare system redesign (Washington State Department of Social and Health Services, 2008). Some states took steps toward differential response implementation that were not fully realized. The Texas Flexible Response model was piloted in San Antonio in 1997. Though the evaluation found that assessments took less worker time than investigations and that child safety was not jeopardized, the state chose to implement investigations for all cases, in lieu of Flexible Response (Chipley, Sheets, Baumann, Robinson, & Graham, 1999). In 1997, the Delaware legislature passed a statute allowing the implementation of differential response, but the state Department of Services for Children, Youth, and Their Families made a policy decision to investigate all screened-in reports, not implementing differential response. West Virginia implemented differential response on a pilot basis with the intent of going statewide, but financial resources limited implementation to five counties (Costello, 1998). Arizona used the Family Builders Program between 1997 and 2004 in 10 of the state’s 15 counties, as an alternative response to facilitate response by community-based service providers to low priority reports of suspected abuse or neglect (Merkel-Holguin et al., 2006). The program was subsequently discontinued, primarily due to funding and service capacity challenges. Additional differential response initiatives developed during the last 10 years include those in Alaska, Kentucky, Louisiana, North Carolina, and Tennessee. Based on the Missouri model, Alaska’s differential response effort began as a pilot in the Matanuska-Susitna Valley and was expanded in 2001 to Nome and Anchorage (Merkel-Holguin et al., 2006). In 2009, the legislation to continue to fund differential response was not passed and it ended. Kentucky’s Multiple Response System began as the focus for the state’s Title IV-B Family Services Plan for 20042009, and was implemented statewide beginning in 2004 (Huebner, 2005). In response to increased referrals, the Louisiana legislature developed guidelines for a multiple pathway approach to child protection and piloted it in the Jefferson and Orleans townships with low-risk families. In 2001, as a response to the state’s federal Child and Family Services Review, the North Carolina legislature adapted child welfare statutes to allow for a Multiple Response System demonstration in 10 counties. By 2006, North Carolina expanded the initiative statewide (North Carolina Department of Health and Human Services [NCDHHS], 2004). In 2005, the Tennessee General Assembly passed legislation authorizing implementation of the MultipleLevel Response System over a 5-year period, with a plan to implement it statewide by 2011 (Merkel-Holguin et al., 2006). Version 2 5 Differential Response in Child Protective Services: A Literature Review November 2011 Minnesota developed a 20-county differential response demonstration project beginning in 2001, which was implemented statewide by 2004. Evaluation findings reported that families, CPS workers, and community-based service providers found the approach to contribute to more positive working relationships; that families in the non-investigation pathway received more needed services than those in the investigation pathway; and that families in the noninvestigation pathway were less likely to have new maltreatment reports (Institute of Applied Research, 2006; Johnson, Loman, & Siegel, 2005; Loman & Siegel, 2004b; Minnesota Department of Human Services, 2005; Siegel & Loman, 2002; Siegel & Loman, 2003). Minnesota’s demonstration was unique in that it was supported with an initial contribution by the McKnight Foundation. This provided participating counties the funds to address initial service needs for families served through the non-investigation pathway, without stretching the county child welfare budget. Some jurisdictions also implemented child-welfare, child-protection, and family-support reforms that addressed many of the same needs as differential response but did not reflect all of the core components (Merkel-Holguin et al., 2006). In Iowa and South Dakota, assessment was used in all CPS responses to balance concerns regarding child safety with the desire for a less adversarial approach. Massachusetts provided private, community-based services to families with unsubstantiated reports through the pilot project Connecting Families in six locales. In Michigan, disposition processes were structured to allow for five differing categories of service provision based on a structured decision-making model. In New Mexico, the state created what it called a “differential response” unit for families whose reports were screened out. North Dakota implemented an assessment approach for all families regardless of risk level, and eliminated the investigation pathway. In Wisconsin, a grantee process allowed for local variations on approaches that used multidisciplinary community approaches to address child maltreatment. In California, a statewide Child Welfare Redesign effort brought together a group of 60 experts and stakeholders for several gatherings and a lengthy planning process, in response to a legislative mandate (Schene & Oppenheim, 2005). At the state level, three possible pathways were recommended.    Path 1 serves families that are experiencing problems but do not meet statutory definitions of maltreatment, and are subsequently linked to services outside the CPS system. Path 2 serves families in which risk of future harm is low or moderate, children are deemed safe, and the family is likely to engage voluntarily with CPS and community service providers to improve child well-being. Path 3 serves families in which children are determined to be unsafe and risk of future harm is moderate or high. This path resembles the more traditional CPS response, in which actions may be taken with or without the family’s consent, and court involvement may occur. The state funded an initial 11 counties to test this multiple pathway approach. An additional 32 counties joined their counterparts in a Breakthrough Series Collaborative project funded by Casey Family Programs from 2002 through 2006. The project’s goal was to improve the participating counties’ organizational capacity in implementing this three-path model (Casey Version 2 6 Differential Response in Child Protective Services: A Literature Review November 2011 Family Programs, 2007). While many counties in California refer to their implementation of this three-path model as differential response, a finding is made as to whether or not the allegation of maltreatment is substantiated for Path 2 families that receive a “non-investigative” assessment response. In 2004, the Ohio Supreme Court Subcommittee on Responding to Child Abuse, Neglect, and Dependency recommended the implementation of differential response in the state. Planning began in June 2007, with implementation in 10 pilot counties launched in July 2008. A final report and evaluation from the pilot was published in April 2010. In 2009, a six county consortium was selected implement and evaluate differential response as part of the National Quality Improvement Center on Differential Response in Child Protective Services (QIC-DR). Ohio has continued expansion efforts with ten counties implementing in fall 2010 and 8 more counties implementing in summer 2011. Illinois and Colorado both began implementing a pilot as part of the QIC-DR in 2010; Illinois’ pilot is statewide, while Colorado’s consists of a five county consortium. Hawaii’s differential response was implemented statewide in 2005 (Walters, Beatty, Livingston, & Lau, 2007), and Nevada began implementing a differential response pilot project in three counties in 2007 before expanding nearly statewide by January 2009 (Siegel, Filonow, & Loman, 2010). Connecticut and Vermont spent time planning and preparing before implementing differential response (Connecticut Department of Children and Families, 2009; Vermont Department for Children and Families, 2007). The New Jersey State League of Municipalities published a document supporting implementation of differential response and announcing the intent of the state Department of Children and Families to fund a four-county pilot project (Ryan, 2007). In 2007, New York State enacted legislation to amend its social service law to allow for the initiation and implementation of what they have termed “family assessment response” as an alternative means of addressing child maltreatment reports. Since they began in 2007, one third of all the counties in the state of New York have trained staff and implemented differential response, with plans to expand statewide (Ruppel, et al., 2011). Advocacy groups in Maryland promoted the differential response approach as a way to address racial disproportionality in the state’s child welfare system (Advocates for Children & Youth, 2008), and it was authorized in the 2008 legislative session (Maryland Department of Legislative Services, 2008). However, the state Department of Human Resources reported that, due to the need for substantive additional public and community resources, the model would not be implemented (Maryland Department of Human Resources, 2008). Version 2 7 Differential Response in Child Protective Services: A Literature Review November 2011 Figure 1 shows the number of states implementing differential response either statewide or on a pilot basis between 1993 and 2011, based on the literature available. Figure 1. Growth of Differential Response Implementation Over Time 20 18 Number of States 16 14 10 10 12 10 5 8 6 2 4 1 2 0 1 2 2 1 2 5 4 4 6 6 6 6 6 6 8 9 10 12 13 10 Statewide Pilot 2 5 7 2 8 5 4 4 5 8 6 6 1993199419951996199719981999200020012002200320042005200620072008200920102011 Year Contributing Events in the History of Differential Response The previously described approaches to diversifying CPS response took place during a period when many discussions on CPS reform were taking place. This section describes the larger CPS context in which differential response developed prior to the 1990s, as well as some of the events that provided the context for development of the differential response approaches described. Beginning with the Child Abuse and Treatment Act (CAPTA) in 1974, followed closely by the Indian Child Welfare Act and Titles IV-B, Subparts 1 and 2, that support Child Welfare Services and Promoting Safe and Stable Families efforts at the state and tribal levels, the federal government has been a leader in addressing child abuse and neglect. The Adoption Assistance and Child Welfare Act of 1980, commonly known as Public Law 96-272, provided federal funding incentives to states to maintain children safely in their own homes or, if removal was necessary for safety reasons, to reunify them as quickly as possible. This law led to the widespread development during the late 1980s and early 1990s of family preservation and reunification programs across the country, encouraging the shift to a more family-centered child welfare practice that emphasized the family’s role in helping plan for services to meet its needs. This shift continued in the late 1990s and into the 21st century, with the increased use of family group decision making and other family engagement approaches to child welfare decision making and service planning (Pennell & Burford, 2000; Wheeler & Johnson, 2003). The Version 2 8 Differential Response in Child Protective Services: A Literature Review November 2011 reauthorization of CAPTA in 2003 required that child protective services agencies develop triage procedures to refer children who were not at imminent risk of harm to community organizations or voluntary preventive services. The Harvard Executive Session, a group of child welfare administrators, practitioners, policymakers, and experts, gathered for a series of meetings between 1994 and 1997 to discuss the need for child welfare system reform (Waldfogel, 1998). This group developed recommendations regarding needed changes in child maltreatment reporting and investigation practices, and identified five major system reform issues:  Overinclusion -- Families may be inappropriately reported to CPS for reasons such as personal vindictiveness or lack of resources.  Capacity -- The number of families referred exceeds the system’s ability to respond.  Underinclusion -- Paradoxically, two groups of families are often underincluded in reports. One group consists of high-risk families screened into CPS systems but not adequately served due to lack of resources. The second group consists of families that voluntarily request services without a screened-in report, but are denied because they are not assessed as high-risk, and are thus ineligible for assistance until actual abuse is indicated.  Service Orientation -- CPS workers try to both keep children safe and keep families intact, while conveying the belief that they, rather than the families themselves, know what is best for children and families. This approach can be confusing.  Service Delivery -- Many families do not receive the services they need. Closer collaboration between CPS and community-based service-providers was recommended, along with increased services for families whose primary language was not English. The Executive Session proposed a new paradigm for child protection in which the public CPS agency shares responsibility with partners in the community in order to respond to children and families in a more flexible, differentiated approach. This new paradigm would differ from current CPS practice in three ways: increased cross-agency service planning, links with informal helpers, and the application of differential response. In 2000, the U.S. Department of Health and Human Services’ Office of the Assistant Secretary for Planning and Evaluation and the Administration for Children and Families’ Children's Bureau (2003), initiated the National Study of Child Protective Services Systems and Reform Efforts. This comprehensive study included a review of state CPS policies, a survey of local CPS practices, and a literature review (Walter R. McDonald & Associates, Inc., 2001). The study identified 20 states providing both investigative and alternative responses, and compared their differential response approaches in terms of policy, implementation, model purpose, pathway definition, and case type. In reporting on local-level differential response practices, the study estimated that almost two-thirds of agencies provided some form of differential response, and that the majority had differential response in place for more than 5 years. The literature review Version 2 9 Differential Response in Child Protective Services: A Literature Review November 2011 conducted to complement the findings from the field (which included documents produced between 1996 and 2001), indicated that differential response was a general direction that CPS practice and policy might take in the “next few years.” In 2003, the Child Welfare League of Canada (Trocmé, Knoke, & Roy, 2003) published a book of articles describing research, policy and emerging models related to differential response implementation in Canada, Australia, England, and the United States. Editors asserted the “tremendous potential and promise” of the approach, while advocating “thoughtful and wellconceived applied research programs to provide a critical evaluation of effectiveness and outcomes” (Dudding, 2003, p. vi). This volume explored the themes that CPS should differentiate levels of family risk, provide diverse and responsive services, and be grounded in effective community-based networks of formal and informal resources. Cultural issues related to effective CPS practice were also addressed, as they related to First Nations children and families, and the French-speaking areas of Canada (Clavel, Cadieux, & Roy, 2003; Foxcroft & Blackstock, 2003). Shusterman, Hollinshead, Fluke, and Yuan (2005) conducted the first multistate analysis of caselevel differential response data from the National Child Abuse and Neglect Data System (NCANDS), as part of a research contract funded by the Office of the Assistant Secretary for Planning and Evaluation for secondary analysis of NCANDS data on topics of current policy interest. Data from 2002 for eight states were analyzed, which included information on 313,838 children, 140,072 of whom were served in the non-investigation pathway. The study found that jurisdictions were referring children to the non-investigation pathway at a stable or increasing rate, and that children served through that pathway were not at any more risk for subsequent reports or victimization than children who received investigations. In 2005, in response to the interest in and proliferation of differential response around the country, the American Humane Association published a double issue of Protecting Children, followed by a seminal, descriptive study entitled the National Study on Differential Response in Child Welfare (Merkel-Holguin et al., 2006), in collaboration with the Child Welfare League of America. A few years later, Kaplan and Merkel-Holguin (2008) revisited the study findings and identified several key, next steps. A growing number of states also began reporting data on differential response to NCANDS, thereby contributing to our understanding of the differential response landscape from a national, quantitative perspective (Ortiz, Shusterman, & Fluke, 2008; Shusterman & Ortiz, 2008) In December 2010, Congress passed, and the President signed, the Child Abuse Prevention and Treatment Reauthorization Act (CAPTA) of 2010. One of the major changes is an increased focus on State and local differential response systems, including new grant purposes and State plan requirements. The Act inserts differential response provisions throughout CAPTA with references to providing an alternative approach to children not at risk of imminent harm. Among the references in CAPTA, differential response provisions are also included in the State assurances of procedures that differentiate severity for appropriate referral, the use of Basic State Grant funding to improve child protective services, requirements to identify policies and procedures around the use of differential response, and the provision of annual State data on the number of families that received a differential response. Differential response is not explicitly Version 2 10 Differential Response in Child Protective Services: A Literature Review November 2011 defined in CAPTA, and States may determine how best to develop procedures to meet the new requirements. In November 2011, the US map of differential response (Figure 2) shows a wide range of statewide and regional implementations, as well as several states in a planning and preparation phase. Figure 2. US Map of Differential Response Implementation Status Version 2 11 Differential Response in Child Protective Services: A Literature Review November 2011 III. IMPLEMENTATION OF DIFFERENTIAL RESPONSE This chapter discusses the literature regarding differential response practices and protocols. The literature is reviewed in terms of the proposed typology of key components and in terms of more detailed description of practices in the field. Examples are also given of state implementation strategies. Core Elements As discussed in Chapter I, one typology of core elements of differential response includes the following.        There are two or more discrete responses to screened-in reports. Assignment protocols and criteria are based on assessment of risk, danger or other requirements. There is the capacity to reassign families to another pathway. The various responses are codified in statute, policy, or protocols. Families may refuse services in the non-investigation pathway; There is no formal determination of whether maltreatment has occurred in the noninvestigation pathway. Caregivers are not determined to be perpetrators and are not listed in a central registry. The extent to which these elements appear in jurisdictions that have implemented differential response is discussed below. Two or more pathways for families screened into the CPS system . A core element of differential response is that it reorganizes CPS to include two or more pathways to serve families that have been screened-in and accepted by the CPS system. At the time of the 2006 national study, 15 states were found to have this component in place (Merkel-Holguin et al., 2006). The literature shows that some jurisdictions have more than one non-investigation pathway:   Olmsted County, Minnesota, includes a separate pathway for families assessed as dealing with domestic violence issues (Sawyer & Lohrbach, 2005a). Hawaii’s differential response approach includes both a Family Strengthening Services pathway that requires family participation, and the Voluntary Case Management Services pathway, in which the family may decline services (Walters et al., 2007). Although a pathway that addresses reports that have been screened-out is not part of the proposed typology of differential response, several states have established more formal responses to screened-out reports. Merkel-Holguin et al. (2006) reported that seven of the 15 differential response states provided screened-out families with referrals to community-based services. Three states’ strategies are as follows:  Kentucky includes a pathway referring families to law enforcement when the alleged perpetrator is not the child’s caretaker. Version 2 12 Differential Response in Child Protective Services: A Literature Review November 2011   Many of California’s counties are implementing models described in the literature as differential response but that provide a non-investigation response only to screened-out families (Bagdasaryan, Furman, & Franke, 2008; Conley & Berrick; 2008; Schene & Oppenheim, 2005). Path 1 links families that do not meet statutory definitions of maltreatment to services outside the CPS system. Similarly, Michigan has implemented a five-path approach, but none of these pathways meets the criteria of a non-investigative response for screened-in families (Michigan Department of Human Services, 2008). Minnesota began implementing the Parent Support Outreach Program (PSOP) in 2005 in 28 counties. The program provides supportive services, on a voluntary basis, to families with children under age 10 who are screened out of the system (Loman et al., 2009). Assignment protocols and criteria. The decision regarding whether to screen a report in or out is commonly made at the point of initial receipt. An intake worker may make a decision to screen out the report, recommend an investigation by the agency, or refer the report to an assessment pathway. Supervisory review and approval is often required. States have differing criteria for how cases can be referred to the non-investigation pathway. The U.S. Department of Health and Human Services (2003) found that 69% of reporting local jurisdictions always required the worker to make an initial safety and risk assessment and determine which pathway to direct the case to. County-administered differential response programs were more likely than state-administered agencies to use formal safety and risk assessment instruments, but were less likely to use a formal domestic violence assessment tool. Other jurisdictions ask the reporter a series of questions and review the agency records to determine if the alleged perpetrator or child victim is known to the agency. As would be expected, there is some evidence of regional variation in assignment to pathways within jurisdictions (Huebner et al., 2009). One example of a pathway selection protocol is the “RED team” (review, evaluate, and direct) approach, used in Olmsted County, Minnesota since 1999 (Sawyer & Lohrbach, 2005a). In this approach, an experienced social worker makes an initial decision about whether a report requires a CPS response. When a report is screened-in as meeting the statutory threshold for public intervention, it is taken to the RED team, which is composed of representatives from a spectrum of child welfare-related disciplines and includes social workers from intake, assessment, investigation, and ongoing intervention. The team can refer a report to the investigation pathway, the Family Assessment Response or non-investigation pathway, or a pathway specifically for families identified as being involved in domestic violence. The RED team’s use of group decision making was found to build agency capacity and result in more consistent and reliable decisions over time. Colorado also uses a RED team approach in making pathway assignments (Drury, et al. in press). The Kentucky intake process involves screening and assessing children’s risk and safety using a risk matrix (Huebner, 2005). The risk assessment process, termed Continuous Quality Assessment (CQA), consists of a checklist of risk factors and narrative screens with prompts to assist intake workers in assessing child and family functioning. After the intake call, case criteria are assessed and documented. The CQA rates risk and maltreatment using a 5-point scale (0-4). Version 2 13 Differential Response in Child Protective Services: A Literature Review November 2011 Reports meeting the criteria for abuse and neglect are assigned either to Families In Need of Services Assessment (FINSA) or to an investigation. Various assessment instruments are being used in the non-investigation and investigation pathways. In response to a worker-identified need for an appropriate assessment tool, Kirk (2008) adapted the existing North Carolina Family Assessment Scale (NCFAS) for use by jurisdictions offering differential response to help target services to families and to evaluate postservice outcomes. The NCFAS-G (for general services) assesses the family’s level of functioning along several dimensions, including family environment, parent capabilities, family interactions, family safety, child well-being, social and community life, self-sufficiency, and family health (Kirk). The NCFAS-G has been used and adapted in several jurisdictions, including Nevada (Weiser & Dreitzer, 2008), North Carolina, and San Mateo County, California (Kirk). Criteria for pathway assignment are based on department policy or enabling legislation (see Center for Child and Family Policy, 2006; Loman & Siegel, 2004b; Siegel et al., 2008). Generally, reports that are judged to represent low to moderate safety risk to the children are eligible for the non-investigation pathway. Reports that allege sexual abuse, serious physical abuse, such as broken bones or internal injuries, fatalities, or near-fatalities, or other behaviors that could result in a felony prosecution, or that involve families with a chronic history with the child protection agency, are usually investigated. However, rules mandating an investigation vary from state to state (Huebner et al., 2009; Office of Children’s Administration Research, 2000; Siegel & Loman, 1997;). Missouri law stipulates that reports involving fatalities, sexual abuse, severe abuse or neglect, serious abuse, or potentially criminal acts must be investigated, and most become co-investigations with law enforcement (Loman, 2005; Siegel & Loman, 1997). Minnesota’s criteria for investigation include the alleged maltreatment types of sexual abuse, abuse in a licensed facility, and “serious harm” (Sawyer & Lohrbach, 2005a). Illinois restricts eligibility for the non-investigative response to families with no previous reports of child abuse and neglect who are reported for inadequate food, shelter or clothing, environmental neglect, mental injury, medical neglect and inadequate supervision of children eight years and older (Kearney, et al., in press). Variations in the application of criteria have been found in several jurisdictions. Shusterman, Fluke, Hollinshead, and Yuan (2005) found that the non-investigation pathway was more likely to be identified as appropriate when reports were from non-professional or educational sources rather than from medical, social services, or legal/criminal justice reporters; when safety issues were less pressing; and when there were no allegations of sexual abuse. Shusterman and Ortiz (2008), in a multistate analysis using 2006 NCANDS data, found that the most frequent maltreatment type referred to the non-investigation pathway was neglect (45%). In Nevada, reports alleging improper supervision or educational, environmental, physical, or medical neglect are assigned to the non-investigation pathway (Weiser & Dreitzer, 2008). Educational neglect was found to be investigated only 6% of the time in Missouri; most reports of educational neglect were compounded with another type of maltreatment (Siegel & Loman, 1997). In Virginia, a family that has been referred to the non-investigation pathway three times prior to the current report must be investigated (VDSS, 2008). Pathway reassignment. All implementations of differential response allow for reassignment to the investigation pathway of a report referred to the non-investigation pathway Version 2 14 Differential Response in Child Protective Services: A Literature Review November 2011 (Costello, 1998; Huebner et al., 2009). In the Nevada differential response approach, if Family Resource Center staff members determine that a child is unsafe, that there is an immediate need for intervention, or that maltreatment allegations are not within the scope of differential response, the case is returned to the investigation pathway (Weiser & Dreitzer, 2008). The Alberta Response Model in Alberta, Canada, allows families to change response pathways if the perceived risk to the child or the family’s voluntary participation changes (Anselmo, Pickford, & Goodman, 2003). Some jurisdictions allow a report to be reassigned from the investigation pathway to the noninvestigation pathway if the case presents less risk to the children than initially believed. At the time of the Merkel-Holguin et al. study (2006), 7 of the 15 states identified as having differential response at that time reported that reassignment was possible from the investigation to the noninvestigation pathway. Some research has reported that such transfers happen less frequently than transfers from the non-investigation pathway to the investigation pathway (Loman and Siegel, 2004b). In Ohio’s 20008 alternative response pilot project, pathway changes were allowed for experimental families but not for control families for the purpose of the evaluation. Pathway changes were permitted to occur when the initial case assignment to alternative response was later found to be inappropriate, as in the case of discovered sexual abuse. This change only occurred in 92 families, 3.96% of the experimental group. However, due to inconsistent methods across counties for reporting these changes, this number may be an underreporting of actual pathway reassignments in the Ohio alternative response pilot (Loman et al., 2010). Pathways must be codified in statute, policy, and protocols. The implementation of differential response has been initiated through legislation in several jurisdictions. A comprehensive table documenting such enacted legislation was published in 2010 by the National Council of State Legislatures as part of the National Quality Improvement Center on Differential Response in Child Protective Services (NCSL, 2010) which includes the following examples:     In Kentucky, the legislative language allowing a differential response to child maltreatment reports through the FINSA pathway described previously was embedded in their approval of the Statewide Automated Child Welfare Information System (Huebner et al., 2009). In New York, the law excludes reports with allegations of sexual abuse, serious physical abuse, severe or repeated abuse, abandonment, and failure to thrive from eligibility for the non-investigative pathway (Ruppel, Huang, and Haulenbeek, 2011) The Minnesota differential response approach, Family Assessment Response, was developed, piloted, and implemented statewide, with the support of several pieces of legislation allowing counties to respond to less severe child maltreatment reports without a disposition (Loman & Siegel,, 2005). In 1994, the Missouri legislature passed a bill allowing the state Division of Family Services to implement a “flexible response system,” and authorized an initial pilot in 14 counties (Loman and Siegel, 2004a). Version 2 15 Differential Response in Child Protective Services: A Literature Review November 2011  North Carolina’s mandate for developing the state’s Multiple Response System came from the state Senate in 2001, after the findings of the federal Child and Family Services Review were released (NCDHHS, 2004). After an initial pilot project from 1997 to 1999, the Virginia House passed legislation in 2000 that mandated the establishment of statewide differential response, and a yearly evaluation of its outcomes (VDSS, 2008). Families may refuse services. A family’s ability to accept or refuse offered services after an assessment, as long as this refusal does not compromise child safety, is a core element of differential response. The purpose of this practice is to create a non-adversarial partnership among CPS, service providers, and family members in order to maximize commitment to improved child and family outcomes. However, it remains questionable whether services conducted within CPS can truly be voluntary (Conley, 2007). If a family refuses to accept services in the non-investigation pathway, most states will reassign the case to the investigation pathway or open a court case while working with the family. For example, Hawaii’s non-investigation pathway for moderate-risk families is explicitly called Voluntary Case Management (Walters et al., 2007). In this pathway, a family that agrees to participate receives a comprehensive assessment and ongoing case planning that includes both identified risk concerns and a plan to address them. Services are provided at no cost to the family and, if the outcomes identified in the case plan are successfully achieved, the case does not result in a CPS finding. If the family chooses not to participate or does not complete the services as recommended, the case is routed back to child welfare services, possibly prompting an investigation and a court-ordered service plan. No formal determination of maltreatment. In an investigation, a systematic effort is typically made to obtain evidence regarding the alleged abuse, and some level of legal proceedings is initiated if alleged abuse is substantiated or indicated. For cases in a noninvestigation pathway, however, family members are not labeled as perpetrators, nor are children labeled victims, and services are provided without a formal substantiation of the alleged maltreatment (Child Welfare Information Gateway, 2008). Merkel-Holguin et al. (2006) identified 15 states that did not require a substantiation disposition for families routed to their non-investigation pathway. Some jurisdictions, such as Delaware, Iowa, South Dakota, Texas, and Westchester County, New York, decided not to implement differential response because legislative mandate, department policy, or agency protocol required that all maltreatment reports receive a disposition of “substantiated” or “unsubstantiated.” In Yuan’s (2005) outline of potential policy issues related to the implementation of differential response, she questioned whether assessment of service needs should be separated from a determination of abuse or neglect in both the investigation and non-investigation pathways. No registration of caregivers in the central registry. When a maltreatment report is substantiated, the name of the alleged perpetrator is often entered into a central registry and made available for future use by public agencies. A USDHHS (2003b) study found that all participating states had a central registry for perpetrators, but varied greatly in the information they stored there. Ten states maintained information regarding all reports, 15 had unique Version 2 16 Differential Response in Child Protective Services: A Literature Review November 2011 guidelines for what was stored, and 23 had central registries that included only substantiated reports. In a non-investigation pathway, names are usually not entered into a central registry. Merkel-Holguin et al. (2006) found that 14 of 15 states indicated that they did not include the name of the alleged perpetrator in differential response cases in the state’s central registry. Additional Features of the Non-Investigation Pathway In addition to the core elements discussed previously, several other features of the noninvestigation pathway are critical to the implementation and sustainability of the approach. These include:       Engaging families; Being culturally relevant; Matching services to needs; Being flexible; Providing training and supervision; and Maintaining community partnerships. Engaging families. One of the distinguishing features of the non-investigation pathway is that parents and/or caregivers are engaged in the planning and achievement of the safety and well-being of their children, rather than treated as alleged perpetrators. The exact balance of family and agency power and roles, and whether engagement is a means to an end or an end in itself, remains open to interpretation. Connolly (2005) characterized an orientation to family support as an understanding of the social difficulties experienced by families and a willingness to provide intensive family-focused services. English, Wingard, Marshall, Orme, and Orme (2000) concluded successful engagement of families was dependent on the families’ recognition that their behavior was problematic and their willingness to engage in building solutions. They also noted the importance of staff members having sufficient skills to support families in recognizing their challenges and continued risk for child maltreatment, and to engage them in planning and resolving issues. Lohrbach et al. (2005) identified family engagement as central to “participatory social work practice,” and asserted that family assets, resources, and capacities were central to the effective resolution of child maltreatment issues. Among the strategies that jurisdictions use in their efforts to engage families were openly sharing information and professional knowledge; exploring family strengths; services driven by family- and agency-identified needs; and maximum family input into case planning. Based on these guidelines, Olmsted County, Minnesota, introduced family group decision making, case planning, and rapid response planning conferences into their child protective development plans in 1997. The type of conference they use depends on the severity of the case. The initiation of this series of family involvement strategies has led to a policy shift toward strengths-based partnership building with families (Christenson, Curran, DeCook, Maloney, & Merkel-Holguin, 2008). North Carolina’s approach incorporated a set of strengths-based strategies to maximize the ability of CPS to effectively engage families and community partners. Child and family team meetings were integrated into the implementation of the state’s non-investigation pathway, as part of in-home services (Comer and Vassar, 2008). While California’s differential response Version 2 17 Differential Response in Child Protective Services: A Literature Review November 2011 initiative does not meet all of the core elements described previously, California’s Breakthrough Series Collaborative identified strategies to improve family engagement, including respectfully engaging families early in the process, using community partners to engage birth families, and including families in all aspects of decision making (Casey Family Programs, 2007). Being culturally relevant. Cultural relevance is an important part of most jurisdictions’ family engagement efforts. Several California counties made ethnic and cultural considerations a cornerstone of their policy and practice (California Social Work Education Center, 2003). One example is the Point of Engagement model initiated in the predominantly African-American and Latino neighborhood of South Central Los Angeles (Marts, Lee, McRoy, & McCroskey, 2008). The model emphasized the protective role that churches and other faith-based groups can play in supporting families, and engaged families with culturally competent domestic violence, substance abuse, and child welfare services. These strategies were determined to be successful at keeping many children out of the CPS system, and therefore contributed to reducing disproportionality in the child welfare system. Ethnic and cultural inclusion was also extensively explored in Canadian and New Zealand applications of the differential response approach (Clavel et al., 2003; Trocmé & Chamberland, 2003). In British Columbia, the overrepresentation of aboriginal children in the child welfare system led to the development of tribal-based child welfare agencies largely grounded in differential response principles and practice (Foxcroft & Blackstock, 2003). New Zealand’s outcomes-based differential response approach applied “family-led and culturally based” outcomes for each phase of their work (Connolly, 2007). Matching services to needs. In the non-investigation pathway, a determination is made of what services the family needs. Although provision of service is a key concept related to noninvestigation, literature on the provision of services is sparse. Zielewski et al. (2006) examined differential response at an urban and a rural site in both Kentucky and Oklahoma, and identified several key elements that influenced how families were referred to the non-investigation pathway. Infrastructure and service availability were found to vary a great deal across the four sites in the study. Factors such as transportation, timing, location, cost, and the quality of relationship between family and service provider influenced families’ success in accessing services. In particular, mental health and substance abuse services were identified as limited by being unavailable in rural areas and oversubscribed in urban areas. Workers were more likely to follow up when families were referred for more intensive services, and CPS workers, service providers, and families all agreed that being served through the noninvestigation pathway made families more willing to work with and communicate with workers and service providers. The study recommended more policy and practice guidance for caseworkers regarding how to proceed with supporting families post-assessment. Being flexible. One hallmark of differential response is having the flexibility of resources to meet families’ service needs. In Minnesota, the McKnight Foundation supported the initial implementation of differential response by providing funds to help defray additional staff hiring and training costs, cover needs identified in the course of assessments, and support the evaluation, including the provision of incentive stipends for families participating in the Version 2 18 Differential Response in Child Protective Services: A Literature Review November 2011 evaluation (Loman & Siegel, 2004b). In Ohio’s first round of differential response implementation, the 10 pilot counties were reimbursed $1,000 per family (up to a predetermined maximum number of families) for post-assessment services while the program evaluation was being conducted (Loman et al., 2010). In other jurisdictions, while financial incentives may not be provided to families, caseworkers seek to provide or arrange for services based on need, in a flexible manner. Providing training and supervision. Each jurisdiction must determine the best caseload standard and structure for the implementation of differential response. For example, the Center for Child and Family Policy (2004), in its evaluation of North Carolina’s differential response, recommended a 1:8 caseload standard for workers serving non-investigation pathway families, to address increased time demands related to working more closely with the families. The study also recommended the continuation of the same social worker in the non-investigation pathway as long as the family was involved with the agency, so that the demands of new cases did not interfere with workers’ capacity to serve families on their current caseload. Worker training and supervision has been raised as a critical component of successful differential response implementation (Child Welfare Information Gateway, 2008). Similar to other CPS and child welfare functions, high-quality, ongoing training must be provided for workers responsible for pathway assignments and implementation of assessments. Cross-training for all workers in the CPS system, including mandated reporters, community-based service providers, and court and law enforcement personnel was also recommended. Comer (2008) stressed the importance of well-designed training for systemic change. Critical components of such training include common structure and language; a respectful, safe training environment; participant opportunities to make emotional connections to the material; opportunities for reflection and application; and skills practice. The experiences and knowledge of workers have also been leveraged to construct differential response training programs in North Carolina and Alameda County, California (Comer & Vassar, 2008; Berrick et al. 2008). Maintaining community partnerships. Differential response approaches vary in the extent to which private agencies serve families referred to the non-investigation pathway. Some states, such as Florida (The Florida Senate, 2003) Kentucky (Huebner et al., 2009), and Washington (Office of Children’s Administration Research, 1998), implemented differential response approaches in which responsibility for serving families referred to the non-investigation pathway was shared between public and private agencies. In these cases, the public agency retained case planning and management responsibility but either the assessment or service delivery was contracted to private providers. In Nevada, all service responsibilities for the noninvestigation pathway were transferred to non-public provider agencies, relying on contracts with local family resource centers (FRCs). The system of FRCs was established by the state legislature to work with state and county agencies primarily to help families and individuals access needed services and support (Siegel et al., 2008). The FRCs hired differential response staff members, who responded when CPS assigned a report to the family assessment response pathway. Huebner et al. (2009) found that in Kentucky, the original plan to transfer all case management responsibility for non-investigation pathway families to private agencies met with considerable resistance from the public CPS agency and resulted in either the non-investigation Version 2 19 Differential Response in Child Protective Services: A Literature Review November 2011 pathway being underused or responsibility being shared. In Illinois, a state with a history of joint public-private partnerships in child welfare, a paired model was implemented. Under this model, family assessment responses are assigned to a two person team consisting of a DR worker from CPS and a caseworker from a private agency. Cases are turned over to the private agency caseworker if no immediate safety concerns are discovered during the initial assessment, ending CPS involvement with the family (Kearney, K., et al., in press). Schene (2005) highlighted the importance of both determining the availability and accessibility of needed services, and effectively coordinating between CPS and community-based service providers. Zielewski et al. (2006) identified factors that affect the success of this connection, including the quality of communication between service providers and CPS and the availability, accessibility, and capacity of community-based services. Though not directly differential response, a San Mateo County, California, survey of community-based service providers reported that they were interested in working with the child welfare agency, but that significant gaps existed in the service array, organizational capacity, and staff training needs (Sphere Institute, 2006). Some jurisdictions used differential response to successfully build community capacity and collaboration. In Olmsted County, Minnesota, CPS integrated domestic violence service providers, law enforcement and probation, and the courts (Sawyer & Lohrbach, 2005b) into their team approach to CPS reports. The Ohio Alternative Response System places emphasis on building and maintaining constructive relationships between and among state agencies, county agencies, community providers and line workers, supervisors, and families (Kaplan, Shannon, & Loman, 2008). Differentiation of cases by severity was also found to positively affect work being done by other community partners involved with the same families. For example, a Missouri study found that the partnership between CPS and law enforcement was enhanced by the clear identification of more serious cases for the investigation pathway (Loman, 2005). New Zealand’s outcome-focused approach relied heavily on partnerships with non-governmental and community-based service providers (Connolly, 2007).While not directly differential response, several examples of successful Canadian community-based service models were presented for consideration in the literature (Anselmo et al., 2003; Brubacher & Narayan, 2003). Implementation Barriers and Strategies While differential response has many philosophical and practical advantages as an approach to addressing child maltreatment, it also presents challenges to CPS systems, which must make major shifts in practices and resource allocations to implement this approach. Primary barriers to successful differential response implementation include professional and stakeholder concerns that not conducting investigations compromises the safety of children, resource allocation, consistent application of assessment and casework protocols and tools, obtaining community buy-in, and community capacity for service provision. For states such as Minnesota (Loman & Siegel, 2004b), evaluation findings, such as those discussed in Section IV of this report, that differential response did not put children in great danger were central to its acceptance. Version 2 20 Differential Response in Child Protective Services: A Literature Review November 2011 Insufficient financial and staff resources have also led to challenges in differential response implementation. One example was in North Carolina, where no additional resources were provided with the differential response-enabling legislation. Local staff reported struggling greatly with restructuring their caseloads to allow sufficient time to serve families in the noninvestigation pathway (Center for Child & Family Policy, 2004). Part of Minnesota’s success in differential response can be attributed to the support of the McKnight Foundation, which contributed significant funding to the state’s initial differential response pilot project (Loman & Siegel, 2004b). West Virginia implemented differential response on a pilot basis with the intent of going statewide, but financial resources limited full-scale implementation (Costello, 1998). Consistent use of assessment tools and the application of casework protocols has also been a challenge. Evaluations of the Missouri use of Structured Decision-Making (SDM) tools found inconsistent use and lack of staff training, which undermined the approach’s effectiveness (Siegel & Loman, 1997). Case reviews included as part of Virginia’s annual evaluation process found ongoing inconsistencies between articulated case protocols and practice at the local level (VDSS, 2006). Worker training on assessment tools and on the differential response approach was consistently recommended as one way to overcome this barrier. One major shift that takes place with differential response implementation is an increased need for CPS to obtain services from community-based service providers. Zielewski et al. (2006) found gaps in the availability of key services in both urban and rural areas in Kentucky and Oklahoma. The implementation of differential response in Nevada has also found an increased need for service provider capacity in order for the approach to be successful in the state, while simultaneously understanding that current financial realities make expansion of Family Resource Centers, the community-based agencies that provide DR services, unfeasible (Siegel et al., 2010). A commonly used strategy to overcome some of these problems has been to phase in differential response implementation (Bagdasaryan et al., 2008). Rather than immediately implementing differential response statewide, many states have first implemented it on a pilot basis in several counties (Loman & Siegel, 2004a; Loman & Siegel, 2004b; Siegel & Loman, 1997), and then expanded it statewide (e.g., Minnesota and Missouri). Aside from Minnesota and Missouri, states that used a multicounty pilot approach included Nevada (Siegel et al., 2010), North Carolina (Center for Child and Family Policy, 2006), Ohio (Carpenter, 2007), Texas (Chipley et al., 1999), Massachusetts, and Florida (Merkel-Holguin et al., 2006). Colorado has additionally opted for a multi-county implementation approach as a participating site in the QIC-DR. In addition to conducting an evaluation of its pilot project, Ohio chose to add an additional qualitative mechanism to study the pilot process, the result of which was the “Chronicle.” The Ohio chronicle documents the pilot project successes, barriers, challenges, and lessons learned throughout each of the phases of the project and from multiple perspectives. A number of challenges were documented during the implementation phase, including caseload demands, inconsistencies in pathway assignment decisions across counties, integration of alternative response into the statewide automated child welfare information system (SACWIS), internal staff buy-in and tension between alternative response and investigation response workers, worker discomfort and skill, and resource constraints (Carpenter, 2010). Version 2 21 Differential Response in Child Protective Services: A Literature Review November 2011 Kentucky, which has a state-administered system, implemented differential response statewide without a pilot process (Huebner, 2005; Huebner et al., 2009), and had difficulty in implementing it consistently. In all differential response approaches where implementation was accomplished in multiple sites, some cross-site differences in the actual implementation of the non-investigation pathway structure were found (Center for Child and Family Policy, 2006; Chipley et al., 1999; Huebner et al., 2009; Loman & Siegel, 2004b; Office of Children’s Administration Research, 1998; Siegel & Loman, 1997; Siegel et al., 2008; VDSS, 2008). This observation suggests the need to explore change management practices involved in implementing a systems change, and the effects of organizational culture on implementation (Bagdasaryan et al., 2008). Version 2 22 Differential Response in Child Protective Services: A Literature Review November 2011 IV. SYNTHESIS OF DIFFERENTIAL RESPONSE EVALUATIONS Differential response has been studied in many jurisdictions to evaluate its effectiveness and value as an approach to flexibly serving families that are screened-in to receive a CPS response. Differential response has been evaluated in 21 states and 1 Canadian province. Of these evaluations, 16 have been completed and 4 are ongoing, among which, Florida has both a completed evaluation of a discontinued differential response system, and an ongoing evaluation of its current pilot program. Initiatives similar to differential response have been evaluated in four other states. One of these is California, with at least six separate evaluation efforts involving 20 or more counties. Table 1 summarizes the status of the evaluations described in this chapter. Table 1.The Status of Differential Response and Related Initiative Evaluations Jurisdiction AK Alberta, Canada AZ CO FL HI IL KY LA MN MO NC NJ NV NY OH OH (SOAR) a TNb TX VA WAc WV a Jurisdiction CA 11 pilot counties 11 small counties Alameda Contra Costa Los Angeles Monterey MN (PSOP) MA NY (Westchester County) Differential Response (DR) Program Evaluations Year Initiated Evaluation Period Evaluation Status 1999 1999-2001 Completed 2002 2003-2004 Completed 1997 1999-2001 Completed 2010 2010-2013 Ongoing 1993; 2005 1996 Completed 2005 TBD In planning/preliminary results 2010 2010-2013 Ongoing 2000 2001-2005 Completed 1999 2001-2004 Completed 1997 2001-2006 Completed 1994 1994-2005 Completed 2001 2003-2004 Completed 2004 2004 Completed 2007 2007-2010 Completed 2007 2009-2010 Completed 2007 2007-2013 Completed, ongoing 2010 2010-2013 Ongoing 2006 2006-2010 Ongoing 1995 1998-1999 Completed 1997 2004-2008 Completed 1998 1996-2005 Completed 1995 1996-1997 Completed Related Initiative Evaluations Year Initiated Evaluation Period Evaluation Status 2003 2004 2002 2004 2005 2007 2005 2001 2003-2005 2004-2005 2002-2007 2005-2006 2005-2007 2006-2008 2005-2008 2004-2006 Completed Completed Completed Completed Completed Completed Completed Completed 2004 2004-2006 Completed Ohio (SOAR) refers to the “Six Ohio counties implementing Alternative Response” as one of the QIC-DR research and demonstration sites to differentiate between a previous study in Ohio. b Tennessee’s evaluation data is preliminary and the QIC-DR has not been given permission to share additional details at the time of this writing. Version 2 23 Differential Response in Child Protective Services: A Literature Review November 2011 c Washington’s implementation of differential response is currently included as part of the state’s child welfare redesign, and may include screened-out reports at the time of this writing. Considered as a whole, these evaluations are instructive on two levels. First, they suggest ways of implementing differential response that have been effective in specific contexts and that could be tested in other settings to determine whether the approach is transferable or unique to its original venue. They also identify some practices and implementation strategies that have been less effective than anticipated. Second, they provide insight into evaluation methodologies that can be used to determine whether differential response is effective and efficient. This section synthesizes the findings of these studies and examines their implications for future differential response evaluation efforts. The evaluations varied considerably in terms of scope and methodology. Some have not yet produced formal written evaluation reports but plans, methodology, and preliminary results were reported in informal ways. Consequently, the amount of information available about the studies was extremely varied. In an effort to develop a comprehensive literature review, all of the differential response evaluations identified in Table 1 were considered. Many evaluations included comparisons between families receiving different responses from the CPS agency. Table 2 indicates which studies included which types of comparison groups. To date, three completed studies (Minnesota, Ohio, and Onondaga New York) used random assignment of families eligible for non-investigation services into groups that received noninvestigation and investigation services. Version 2 24 Differential Response in Child Protective Services: A Literature Review November 2011 Table 2. Outcome Evaluation Methodologies Differential Response (DR) Program Evaluations Jurisdiction AK Alberta, Canada AZ CO FL HI IL KY LA MN MO NC NJ NV NYa OH OH (SOAR) TX VA WA WV DR Total Jurisdiction a CAb 11 pilot counties Alameda Contra Costa Los Angeles Monterey MA MN (PSOP) NY (Westchester County) Related Total Experimental Design with Random Assignment Quasi-Experimental Design Matched Site Matched Families  Natural Experiment (case review)   Pre-Post Data Comparison   (case review)   (sub-studies)          6 Experimental Design with Random Assignment  (case review) 4 3 Related Initiative Evaluations Quasi-Experimental Design Matched Site Matched Families        10 Natural Experiment      2 Pre-Post Data Comparison     0 0 3 7 2 The experimental design was followed in Onondaga County, while Tompkins County used a historical control group, using eligibility criteria and propensity scores to create a control group. b The 11 small-county study in California did not include an outcome component. Evaluation Design, Methodology, and Results Three major components of an evaluation -- process evaluation, outcome evaluation, and cost evaluation -- as implemented in the existing differential response evaluation literature, are summarized in the remainder of this chapter. For each component, evaluation methodologies employed and the results found are summarized. Version 2 25 Differential Response in Child Protective Services: A Literature Review November 2011 Process evaluation. Process evaluation methods are intended to address why specific outcomes were or were not achieved. Process evaluations of differential response generally include collection and analysis of demographic and case characteristics data on the various groups participating in the outcome analysis. Additionally, organizational characteristics of the participating agencies, including changes in differential response use and community contextual variables; experience and attitudes of workers involved in the study; and client, community, and worker satisfaction data are often collected and reviewed to identify factors that may have influenced the outcomes associated with differential response. Table 3 summarizes the types of process evaluation methodologies that have been used in differential response and related initiative evaluations conducted to date. As seen in Table 3, most of the evaluations used similar process evaluation methodologies. Among the most frequently used methods for understanding impacts on differential response implementation processes were analyses of the demographic and case characteristics of families. This information was primarily used to determine the similarities and differences between families served through investigation pathways and non-investigation pathways. These data were also used to describe the populations served and determine if families assigned to the two pathways met the appropriate criteria. In a few studies, these data were also used to examine whether serving families with specific characteristics (e.g., very young children or adolescents) through the non-investigation pathway was more or less effective. In Minnesota, Loman and Siegel (2004b) surveyed a sample of 690 workers. Responses indicated statistically significant differences between experimental and control group families. Families served through the non-investigation pathway were seen as more cooperative, realistic, and motivated, and as having higher self-esteem, suggesting that the family-engagement process inherent in differential response changed the dynamics of interaction between workers and family members. They also found that this more positive and less confrontational or investigative interaction may have inhibited these family members from reporting domestic violence issues. In Ohio, workers were also surveyed around their perceptions of family reactions to alternative response. Workers reported feeling that families reacted more positively to them, were more cooperative, and were more involved in the decision making process in a family assessment response than in a traditional investigation. Families themselves were also surveyed and their responses confirmed worker perceptions. Families receiving family assessment responses were more likely to have participated in services and reported greater satisfaction with how they were treated by their workers and with the services they received than families receiving investigation responses (Loman et al., 2010). Families surveyed in New York and Nevada’s differential response pilot who received a family assessment response also reported satisfaction with worker treatment and perceived workers as friendly, supportive and cooperative (Ruppel et al., 2011; Siegel et al., 2010). Loman and Siegel (2004b) also found that families served through the non-investigation pathway in Minnesota were more likely to receive services if they reported that their children had educational or mental health difficulties, or if the family was in financial stress, in poverty, had few social supports, or had inadequate housing. Families assigned to the control group, namely the investigation pathway, reported similar difficulties, but were less likely to receive services to Version 2 26 Differential Response in Child Protective Services: A Literature Review November 2011 ameliorate them. This study also found that families of color in the non-investigation pathway were more likely to receive services than were white families, while in the investigation pathway, white families were more likely to receive services. The authors attributed these differences to two factors. First, some allegations (such as neglect, which was associated with a higher percentage of families of color than with white families) were more likely to lead to services in the non-investigation pathway than in the investigation pathway. Second, families of color were more likely to accept services voluntarily (in the non-investigation pathway) than engage in services that were imposed involuntarily (in the investigation pathway). In Onondaga, New York, family caregivers whose families were given a non-investigative response were more likely to report that they had received some services compared to those receiving an investigation response (70 percent and 54 percent, respectively) (Ruppel et al., 2011). In Ohio, Loman et al. (2010) found that all experimental families, regardless of race, received increased rates of service provision, when compared to their control group counterparts. In contrast to Minnesota’s evaluation findings no difference was found in the number of services provided or the provision of direct services between white and African American families. However, some variation was found in the types of services provided between these groups. Whereas African American families received more child care, legal, and disability services, in addition to more household appliances and furniture, white families received more counseling, respite care, and other types of financial assistance. Version 2 27 Differential Response in Child Protective Services: A Literature Review November 2011 Table 3.Process Evaluation Methodologies Jurisdiction AK Alberta, Canada AZ CO FL HI IL KY LA MN MO NC NJ NV NY OH OH (SOAR) TX VA WA WV DR Total Jurisdiction CA 11 pilot counties 11 small counties Alameda Contra Costa Los Angeles Monterey MA MN (PSOP) NY (Westchester County) Related Total Describe Client Characteristics                15 Describe Client Characteristics        Differential Response (DR) Program Evaluations Describe Worker Describe Case Depict DR Assess Model Characteristics & Characteristics Use Fidelity Satisfaction                                                                     18 17 18 15 Related Initiative Evaluations Describe Worker Describe Case Depict DR Assess Model Characteristics & Characteristics Use Fidelity Satisfaction          8 8      5 Report Client Satisfaction/ Engagement   Describe Community Context                   13 7 Report Client Satisfaction/ Engagement Describe Community Context                      7 6 6 7      In Missouri, Siegel and Loman (1997) found that families in the investigation and noninvestigation pathways were “essentially identical” in terms of prior CPS history. The families were also “very similar” demographically, with both groups showing a decline in the percentage of two-parent families over the evaluation period, suggesting an increase in the complexity of families served, particularly in the more urban areas. The extent to which differential response was used and changes in use over time were also analyzed as part of the process evaluation in the evaluations considered. In North Carolina, for example, Center for Child and Family Policy (2006) found that in the pilot counties, the number of families served through the non-investigation pathway increased 10% between 2003 and 2005. They attribute this increase to the conclusion that “as DSS workers and supervisors feel more confident that they can serve families more effectively in the Family Assessment pathway Version 2 28 Differential Response in Child Protective Services: A Literature Review November 2011 without compromising the children’s safety, they will likely reserve the investigation pathway for cases of severe maltreatment” (p. 18). The authors also found that families for whom services were not recommended decreased by 10%. That is, differential response expanded the number of families that receive services on a voluntary basis. There were differences among the pilot counties in the way they implemented the differential response program. These differences make it difficult to examine changes in differential response implementation in detail. For example, because of the effort to “frontload” services for families served through the non-investigation pathway, some counties found that by the time a decision was made, the family no longer needed services, so they made a “services not recommended” decision. Other counties, for the same types of families, made a “services recommended” decision because the family needed and received services. This disparity resulted in adding a new decision option in 2006, “services provided, child protective services no longer needed.” Like North Carolina, Siegel and Loman (1997) found in Missouri that differential response seemed to expand the population served by the CPS system. That is, there were small but statistically significant increases over time in the reports of less severe physical abuse and unmet basic needs of children, both of which were generally not accepted as CPS cases prior to differential response. There was also a decrease in the reports of educational neglect, which somewhat offset the expansion of the service population due to less severe abuse and povertyrelated neglect cases. In general, this finding suggests that as CPS agencies attempted to meet the basic needs of children at risk of maltreatment and their families, the population they served may also have been broadened. The assessment of model fidelity is critical to understanding whether differential response was effective, since it is essential to know with considerable accuracy just what program was being implemented and evaluated. Model fidelity was examined in 21 of the evaluations reviewed. The core components of the differential response model were implemented fairly consistently in each site. However, there were some local adaptations in all sites. In Kentucky, Huebner et al. (2009) found that, statewide, the percentage of reports that met the criteria for the noninvestigation pathway and were accepted increased from 26% in 2001 to 34% in 2008, with statistically significant differences among the regions in terms of the percentage of eligible reports that were referred. Using administrative data for more than 39,000 cases that met the criteria in federal fiscal year 2008, the percentage of reports referred for assessment ranged from less than 20% (eastern Kentucky) to almost 60% (northern Kentucky). In one region, the percentage of cases that were assessed exceeded the percentage investigated. Of the reports that were investigated, the percentage substantiated also differed significantly among regions. Regions of Kentucky that had the highest percentages of investigated reports had the lowest percentage of substantiated reports. Virginia found similar regional differences in the use of the non-investigation pathway, as well as similar growth in its use over time (VDSS, 2008). Some of these differences among the regions may be due to differences among the local departments of social services in terms of the number of families with service needs, the number and adequacy of available local services (thus reducing worker likelihood of requesting services known to be unavailable), worker capability to assess and meet needs, and differences in the thoroughness with which workers enter data into Version 2 29 Differential Response in Child Protective Services: A Literature Review November 2011 the state’s information system. In Virginia, referrals to the non-investigation pathway ranged from 25% in western Virginia (which borders eastern Kentucky, the lowest referral region in that state) to 45% in the northern and central regions. The similarly low referral rates in the mountainous and poverty-stricken regions of western Virginia and eastern Kentucky may reflect both the inadequacy of available services and the local culture that resists dependence on “outsiders” for help. Kentucky’s experience in implementing differential response was instructive in other respects. The original program design was that families that met the criteria for an assessment would be referred to a private agency for assessment and subsequent case management. However, Huebner et al. (2009) reported that, “The use of community partners to complete assessment of low risk cases was never implemented” (p. 4). The private agencies were not prepared to conduct these assessments and saw it as an attempt by the CPS agency to “delegate” its work to them. Conversely, CPS personnel indicated that all of the screened-in reports, independent of which pathway they were assigned to, met the state legal criteria for child abuse and neglect, so the CPS agency needed to be involved in the case. Nevada’s alternative response model was also designed for family assessments to be conducted outside of CPS. Following the CPS screening process, cases assigned to a family assessment response are immediately referred to a local Family Resource Center (FRC). FRCs are agencies that were established in the mid-1990s to work with state and local agencies to assist families in accessing supportive services. In Nevada, FRCs are responsible for conducting the initial family assessment, including safety and risk assessments, case planning, service provision, and entering case data into the state child welfare system, UNITY. However, unlike Kentucky, feedback from families and FRC workers indicate that differential response has been implanted in Nevada with model fidelity, both in terms of protocol and service provision (Siegel et al., 2010). Loman and Siegel (2004b) found in Minnesota that each of the participating 20 counties adapted differential response to meet local conditions. In several counties, the organizational structure through which differential response was administered changed at least once. For example, both Hennepin (Minneapolis) and Ramsey (St. Paul) Counties changed at least once from administering differential response through specialized units to having all workers serve families in both the investigation and non-investigation pathways. In some small counties, in which “a single set of workers did everything” (p. 29), the same worker retained the case from the initial report through service delivery to case closing. In other counties, some had specialized units for their family assessment response (non-investigation) cases; others had the same workers serving families in both pathways. In still others, the initial worker who conducted the family assessment also provided ongoing services to the family, while others transferred the family to an ongoing services worker. In some counties, the initial worker served the family if its case remained open for a short period of time but transferred the family to another worker if the case remained open past a certain number of days. Used often, but with slightly less frequency, were surveys of worker characteristics and satisfaction with differential response (16 evaluations), determinations of client satisfaction and the extent to which participating families were rapidly and effectively engaged in an Version 2 30 Differential Response in Child Protective Services: A Literature Review November 2011 investigation or assessment, service planning and service delivery processes (12 evaluations), and a description of community factors that affected differential response implementation (10 evaluations). Workers were generally satisfied with the non-investigation pathway in terms of improved family engagement and cooperation. However, workers in some states expressed serious concerns about some characteristics of the approach, which were seen to affect implementation. For example, as discussed previously, in Kentucky, Huebner et al., (2009) found that both CPS and private agency personnel were uncomfortable with the decision to turn lowrisk cases over to private agencies for assessment and case management without CPS agency involvement. That aspect of the differential response system was never implemented because of these concerns. Client satisfaction reviews in 13 sites indicated uniformly that clients who participated in the non-investigation pathway were happier with it than with the investigation approach. They felt as if they were engaged in the assessment and service planning processes. For example, Siegel and Loman (1997) found that in Missouri, 81% of the families in differential response pilot areas were satisfied with their services, compared with 71% of families from the comparison areas that were not implementing differential response. Within the pilot areas, 88% of the families that received an assessment were satisfied with the services they received, compared with 57% of the families that were investigated. Similar results were found in the other jurisdictions that examined clients to determine their satisfaction (Center for Child and Family Policy, 2006; Costello, 1998). The 11 evaluations that looked at impact on community involvement and attitude toward CPS found that the number of agencies that cooperated with the CPS agency increased after differential response implementation (Loman, 2005; Siegel & Loman, 1997). Community context also affected the evaluation. For example, Loman and Siegel (2004b) found that prior local experience with differential response and local evaluations affected participation in the statewide experimental evaluation in Minnesota. Other sites found that the improved client attitudes toward CPS were reflected in increased cooperation of other agencies with CPS (Center for Child and Family Policy, 2006; Huebner et al., 2009). Outcome evaluation. Evaluating the outcomes of any intervention effort demands a comparison, in some manner, of intervention participants with non-participants. That is, an outcome evaluation must address the question, “Does this program work compared to doing something else -- usually the standard intervention -- or doing nothing?” Such a comparison is sometimes combined with an analysis of outcomes for the same group(s) before and after the intervention, referred to in this review as “pre/post data comparison.” The most rigorous method for determining the effectiveness of any social program in achieving its desired outcomes is an experimental design that randomly assigns eligible participants to either the innovative program being evaluated (the “experimental group”) or to standard services that they would otherwise have received if the service innovation were not available (the “control group.”) The outcome evaluation methodologies used, singly and in combination, in the differential response evaluations reviewed are summarized in Table 2. Of the completed differential response evaluations, Minnesota, Ohio, and New York initiatives were evaluated using an experimental design. However, in New York only Onondaga County Version 2 31 Differential Response in Child Protective Services: A Literature Review November 2011 implemented using an experimental design. In Minnesota, of the 20 pilot counties in Minnesota which participated in the evaluation, only 14 permitted implementation of an experimental design. Some counties had several years’ experience with implementing differential response before the statewide implementation began and were reluctant to make changes to an approach that they believed to be successful (Loman and Siegel, 2004b). Additionally, the three QIC-DR sites (Ohio, Colorado, and Illinois) are currently employing random assignment in their differential response systems so that each of their evaluations will use an experimental design. Quasi-experimental designs, in which groups of participants in the new service are matched with similar clients or families, or counties and other jurisdictions that implement the new service are matched with similar jurisdictions that do not, are used in social service evaluations more often than experimental design. The differential response evaluations in Alaska, Missouri (Siegel and Loman, 1997), Nevada (Siegel et al., 2010), North Carolina (Center for Child and Family Policy, 2006), and Texas (Chipley et al., 1999), and substudies in Kentucky (Huebner et al., 2009), and Tompkins County, New York (Ruppel et al., 2011) used a quasi-experimental design. Those in Alaska, Missouri, North Carolina, and Texas matched jurisdictions to the extent possible, while those in Kentucky and Nevada matched groups of client families on key characteristics. Of the seven evaluations of programs similar to differential response, those in Alameda County, California (Berrick et al., 2008), and Massachusetts (Merkel-Holguin et al., 2006) used an evaluation strategy that created groups matched on certain characteristics. A third evaluation strategy often used in social service evaluations is to conduct a “natural experiment” by aggregating data on clients who are referred to and participate in the new program compared to those who are eligible for the new service but are either not referred or elect not to participate. This method was the most commonly used method for outcome evaluations, as seen in Table 2. These two groups are usually compared using both original data and administrative data available from automated information systems. This approach has been used in 11 of the differential response evaluations and 7 evaluations of related programs. For example, it was used to evaluate differential response programs such as in Kentucky (Huebner et al., 2009), Virginia (VDSS, 2008), and Washington (Office of Children’s Administration Research, 2005). As may be seen from Table 2, the natural experiment was the most commonly used method for assessing outcomes. A pre/post data comparison approach involves collecting data on all of the participating populations for some specified period before differential response implementation, then comparing the pre-implementation data with similar data for those populations collected during the follow-up period (usually 6 months to a year, but sometimes up to 5 years) after differential response implementation and services have been received. This data collection strategy normally relies on administrative data compiled from the agency’s automated information system, although almost all differential response evaluations have used the administrative information system as a major source of data for the evaluation. This approach was used for the differential response in Alaska (Merkel-Holguin et al., 2006) and Missouri (Siegel and Loman, 1997), and with the evaluations of related initiatives in Massachusetts and Westchester County, New York (Merkel-Holguin et al., 2006). . Version 2 32 Differential Response in Child Protective Services: A Literature Review November 2011 The completed outcome evaluations of differential response approaches have examined the effects of participation in the non-investigation pathway on a number of child, family, agency, and community outcome variables. These variables (from completed evaluations) are summarized in Table 4. Table 4. Differential Response Outcome Evaluation Variables Version 2    2 4 1 1 1 8 Related Initiative Evaluations 8      7          Physical/Behavioral Health  Education  Time in Care Maltreatment Risk    3          1  9 3 1 1 5  2 Worker Retention Workload            7 2  8 Agency Outcomes 5     8 Community Outcomes           4 Job Satisfaction Parent Health/Skills    3 1 Family Outcomes Placement Moves Re-Report Severity          Removals Re-Report Number CA 11 pilot counties Alameda Contra Costa Los Angeles Monterey MA MN (PSOP) NY (Westchester County) Related Total Service Adequacy     Service Capacity      Positive View of CPS 8 Service Timeliness   5      Worker Retention 7      Workload            Job Satisfaction     Parent Health/Skills    Service Adequacy    Community Outcomes Service Capacity         Child Outcomes Jurisdiction Economic Hardship Physical/Behavioral Health  Education Time in Care Maltreatment Risk  Service Timeliness            16   Economic Hardship   Placement Moves    Agency Outcomes Family Outcomes Removals AK Alberta, Canada AZ FL HI KY LA MN MO NC NJ NV NY OH TX VA WA WV DR Total Re-Report Severity Jurisdiction Re-Report Number Child Outcomes Positive View of CPS Differential Response (DR) Program Evaluations  3 2  1 1 2 3 33 Differential Response in Child Protective Services: A Literature Review November 2011 Of the evaluations that included an outcome analysis, 23 examined, in one manner or another, the extent to which children reported as maltreated in the incident that led to the referral CPS were rereported during the follow-up period. Ten evaluations took this examination further and analyzed the relative severity or types of the alleged maltreatments to determine if the children and families were reported for similar issues both before and after being served through the noninvestigation pathway (or investigation pathway if in the control group). Nine evaluations also considered post-program safety and risk of future maltreatment. In the Missouri extended followup study, Loman and Siegel (2004a) were able to combine 13 measures of risk that were associated with future re-reports into a single consolidated measure. Twelve evaluations considered the effects of participating in the non-investigation pathway on subsequent removal and placement of a child in out-of-home care. Three considered number of placement moves and six analyzed the effects of participation on the length of time in care. Only Massachusetts and Minnesota examined such child well-being issues as educational performance and physical or behavioral health. Regardless of which method was used, family participation in the non-investigation pathway was demonstrated to reduce the numbers of post-program reports for alleged child abuse and neglect, (though the differences were often modest), so that as several evaluators put it, child safety was not compromised as a result of families being served through the non-investigation pathway. This fundamental result was found in Alberta, Canada (Weiden, Nutter, Wells, & Sieppert, 2005), Alaska, Arizona, Kentucky (Huebner et al., 2009), Massachusetts, Minnesota (Loman & Siegel, 2004b), Missouri (Siegel et al., 1997), Nevada (Siegel et al., 2010), North Carolina (Center for Child and Family Policy, 2006), Ohio (Loman et al., 2010), Texas (Chipley et al., 1999), Virginia (VDSS, 2008), Washington (Office of Children’s Administration Research, 1999), and West Virginia (Merkel-Holguin et al., 2006).In Onondaga, New York, the risk of rereport was not significantly different across the two groups, but those results were based on a short follow-up period of six months (Ruppel et al., 2011).. Loman and Siegel (2004b) also found that the level of risk presented by a family influenced its likelihood of being rereported for child maltreatment during the follow-up period. Controlling for risk, the families in the non-investigation pathway had significantly lower recurrence rates than families that were investigated. Recurrence percentages were found to be lower for families in the non-investigation pathway for the three major racial groups (White, African American, and American Indian) in the study. This result was also found in Ohio as well as in Nevada where families served through the non-investigation pathway were found to have a 27% reduction in report recurrence than similar families receiving an investigation response over a 40-month period (Loman et al., 2010; Siegel et al., 2010). As may be seen in Table 4, 21 of the 26 differential response and related initiative evaluations considered some family outcomes. The most frequently examined of these was the adequacy of services to meet the family’s needs after the initial report. The majority of these sites found that services were provided to those served through the non-investigation pathway earlier than to those in the investigation pathway. Thirteen evaluations also considered the adequacy of services provided in meeting the assessed needs of the families involved. A number of the differential response evaluations were in states with large rural areas where service availability was a chronic Version 2 34 Differential Response in Child Protective Services: A Literature Review November 2011 problem. Studies of three county-level differential response-related initiatives (in California, Massachusetts, and Westchester County, New York) that considered service adequacy in the evaluations of initiatives similar to differential response were mostly in more heavily populated areas, although the 11 pilot counties in California included some large rural areas. In more urbanized areas, the primary issue of service adequacy was one of available capacity of existing services rather than the actual existence and accessibility of services that affect rural areas. All of these evaluations found that the availability of services increased under differential response. Minnesota’s alternative response program evaluations and the Parent Support Outreach Program (PSOP)evaluation examined family economic hardship. Both of these studies found that families participating in both the non-investigation pathway and PSOP were less stressed regarding their economic well-being a year after receiving services than were families in the investigation track (Loman & Siegel, 2004b; Loman et al., 2009). Four of the evaluations of related initiatives in California (ACTION Council of Monterey County and Monterey County Department of Family and Children's Services, 2008; Child and Family Policy Institute of California, 2006; Berrick et al., 2008) and Massachusetts (Merkel-Holguin et al., 2006) reviewed the effects of participating in the non-investigation pathway on families’ economic situations in terms of income, employment stability, housing adequacy, and homelessness. The evaluations in Alberta (Weiden et al., 2005), Massachusetts (Merkel-Holguin, 2006), the California 11-county pilot (Child and Family Policy Institute of California, 2006), and Alameda County (Berrick et al., 2008) were the only four that analyzed the effects of participation in the non-investigation pathway on parental social support, parenting skills, and related behavioral health issues. Agency outcomes were also considered in 13 of the studies. The most common agency outcome included in the differential response evaluations (examined to some extent in eight differential response studies and two evaluations of differential response-related initiatives) was the effect on worker job satisfaction. All found that, like families, workers were positive about the less confrontational and more service-oriented approach of the non-investigation pathway, and felt that it did not reduce child safety or increase risk. Studies for California (Child and Family Policy Institute of California, 2006), Florida (The Florida Senate, 2003), Hawaii (MerkelHolguin et al., 2006), Kentucky (Huebner et al., 2009), Ohio (Loman et al., 2010), and Nevada (Siegel et al., 2010) considered the effects of differential response on workload. The Kentucky study found some increase in the workload of staff who participated in the non-investigation pathway. Onondaga, New York found slightly higher rates of dissatisfaction with workload among alternative response workers compared with investigation response workers (Ruppel et al., 2011). Nine studies examined the effects of differential response or related efforts on community attitudes toward CPS. Eleven studies reviewed the effects of differential response implementation on the development and location of community-based services appropriate for meeting the needs of participating families. In general, the findings from both sets of analyses were positive. Cooperation and information sharing among the public CPS agency, law enforcement agencies, schools, and private providers improved because of differential response and related program efforts, particularly those in which non-investigation pathway staff were colocated in other agency facilities (Loman, 2005; Loman & Siegel, 2004b; Siegel & Loman, 1997). In most communities where the effects of differential response on service capacity were Version 2 35 Differential Response in Child Protective Services: A Literature Review November 2011 considered, the availability of services increased to some extent, limited by economic and other environmental factors (Weiden et al., 2005). Cost evaluation. Of the differential response evaluations completed to date, only Minnesota and Ohio’s have included a cost-evaluation component. The cost evaluation of the Minnesota program included a comparison between a sample of 752 experimental and control group families served between July 2001 and December 2002. For all these families, the total costs of services during two periods were computed. The first period was between the referral for the incident that brought the family into the study and the date of the last CPS contact related to that report (85 days on average). These costs -- the initial service costs -- were then compared with the total cost of services for the family from the end of the initial intervention period through the end of the follow-up, an average of 453 days. In that way, the total costs of serving the sample of families included in the cost-evaluation component of the study could be compared to determine if there were long-term cost differences in serving families in the investigation and non-investigation pathways (Loman & Siegel, 2004b). For these cases, the total cost of services was computed, including all purchased services identified by local CPS agency bookkeepers and the cost of staff time reported in the state information system. For the sampled cases in the experimental group, an average of $1,131.80 was expended during the initial service period and $804.44 during follow-up, for a total of $1,936.24. For control group cases, the opposite expenditure pattern emerged. During the initial period, $593.45 was expended on average, but $1,537.68 was expended during the follow-up period, for a total of $2,131.13. In other words, these data indicated that services for families in the non-investigation pathway were more expensive initially but result in reduced service costs for participating families in the long run. The authors of the study stated the basic conclusion of the cost study: “Savings achieved by experimental families during the follow-up period more than offset investment costs incurred during the initial contact period” (Loman & Siegel, 2004b, p. 162). The fundamental finding remained unchanged when the authors conducted further analyses of cost data excluding data from Ramsey County, which had considerably higher costs than other counties for unexplained reasons, and for the six counties that did not participate in the experimental study. These findings were also found in a follow-up several years later. They also explored the cost-effectiveness of the differential response system in Minnesota. Using outcome and cost data, they concluded that the mean cost of achieving the goal of avoiding recurrence of child abuse and neglect reporting was $398 less for the experimental group than for the control group. If Ramsey County data were excluded from the calculation, the difference would have been $966 on average (Loman & Siegel, 2004b). Using a representative sample of families from the experimental and control groups, Ohio’s cost evaluation focused on direct service and indirect (administrative) costs incurred to both groups during a period that ranged from ten to 15 months. The evaluation found that indirect costs for experimental families were higher at an average of $940 per family, versus $732 for control families, for the initial assessment period. This is consistent with assumptions that workers would spend more time with these families who may not have had a substantiated case had they received an investigation Version 2 36 Differential Response in Child Protective Services: A Literature Review November 2011 response. However, when adding indirect costs from any cases opened as the result of a subsequent report and any subsequent home removals, costs were higher for the control families at an average of $266 versus $145 for experimental families. The total indirect costs during the 15-month maximum follow-up period were $1,085 for families in the experimental group versus $998 for the control group (Loman et al., 2010). In examining direct service costs, such as any spending on concrete services such as food, clothing, or out of home placement, Ohio’s cost evaluation found that in the initial assessment period costs for control families averaged $99 versus $194 for experimental families. Again, these figures are consistent with evaluation finding that families receiving family assessment responses received more concrete, often poverty-related, services than their counterparts in the control group. However, when costs incurred for subsequent cases were examined, average costs of direct services to control families increased to $136 while decreasing for experimental families to $48. Total costs, direct and indirect, for experimental families were more expensive by $92 in the follow-up period. However, Loman et al. (2010) note that there is evidence that the possibility for net savings to be achieved in a family assessment response in the long-term, as was the case in Minnesota, as subsequent reports and home removals are reduced over a longer follow-up period. Summary of Differential Response Evaluations The process evaluation results were generally consistent across the different jurisdictions. Both families and program staff were more satisfied with the non-investigation pathway than with the investigation pathway. Families were more easily and fully engaged, usually earlier in the service delivery process. However, in all of the multijurisdictional sites that were evaluated, there was considerable variation both in the extent to which differential response was implemented and in the attitudes toward this approach. These differences were especially marked in Kentucky and Virginia, with the extremely rural Appalachian regions of both states using the non-investigation pathway considerably less than other parts of the states. In the evaluations that examined client and case characteristics as factors that might affect program outcomes, the most consistent finding was that degree of risk, including prior CPS history, was the primary differentiating factor between more and less successful outcomes in terms of reduced re-referrals to CPS. Although the results from all of these studies are generally positive with regard to the outcomes achieved by the non-investigation pathway participants, the extent to which this was the case varied considerably. In general, the outcome evaluations of the programs in Minnesota, Missouri, and North Carolina found that there was no negative impact on child safety and that the number of re-referrals within the timeframe studied was less for families served through the noninvestigation pathway than for those who were investigated. While these differences were statistically significant, they were also relatively small. Fewer and less clear differences were found in the other state studies included in the analysis. To date, only three differential response evaluations that used an experimental design based on random assignment of participating families have been completed, with three more to come from the QIC-DR sites. Several additional evaluations used a quasi-experimental design. Those in Version 2 37 Differential Response in Child Protective Services: A Literature Review November 2011 Alaska, Missouri, North Carolina, and Texas compared counties or other jurisdictions that implemented differential response with similar jurisdictions that did not. The evaluation in Nevada matched groups of client families on key characteristics. The remaining evaluations included in this review used a “natural experiment” design in which the evaluators analyzed aggregated data on clients who were referred to, and participated in, the non-investigation pathway, compared with data on those who were eligible for the new service but were either not referred, elected not to participate, or participated in different service delivery models. These groups were compared using both data collected specifically for the evaluation and administrative data available from automated information systems. Version 2 38 Differential Response in Child Protective Services: A Literature Review November 2011 V. DISCUSSION AND CONCLUSIONS In this conclusion, we will summarize the accomplishments of the field and what is known about differential response, and identify important gaps in our knowledge base. What Has Been Accomplished? Having started in Florida and Missouri in 1993, differential response is now being implemented at some level in 17 states, and another 6 have plans to implement it in the near future. Many states that started with pilot projects in selected counties have subsequently implemented differential response statewide. Although some states have chosen to discontinue differential response, the overall number of families served through a non-investigation pathway has seen steady growth. The field has attained greater definitional clarity as it has developed, including recent and ongoing attempts to refine the list of core characteristics of the non-investigation pathway. Such activities help the research and practice fields by providing a common language for discussion and planning. There remain some differences of opinion on such issues as how the noninvestigation pathway is defined (negatively by the absence of legal investigation or positively by the support of and partnership with families), whether engagement in CPS can be truly voluntary, and if the non-investigation pathway is limited to reports screened-in to the CPS system, or those screened-out upfront. These are all important debates that will help shape differential response in the future. What Is Known? Program descriptions and evaluations have addressed essential questions concerning best practices, and the effects of participating in the non-investigation pathway on child safety and permanency, family well-being, and agency functioning. Important issues of cost-effectiveness were addressed in one differential response evaluation so far. These program reports and evaluations provided substantial knowledge for guiding current practice, and help identify implementation issues that will influence both future differential response development and research opportunities. Process. One of the impressive findings from the evaluation synthesis was the consistency on key measures related to family engagement, in spite of the great cross-site variations in geography and implementation. This indicates a certain robustness of the differential response approach, one that is tolerant of operational and circumstantial variances. Another common finding was the reported increase in cooperation between CPS and partner agencies following the implementation of differential response, a core goal for most differential response approaches. Outcomes. The premier and most consistent finding of these evaluations, regardless of evaluation methodology, was that differential response does not result in increased harm to children. It must be noted, however, that differential response is used with children who have been assessed as less at risk. Indeed, the randomized control studies to date have randomly assigned children assessed to be at lower risk of future maltreatment. There is some evidence of a modest benefit to safety, as indicated by fewer rereports. Several studies, such as the Version 2 39 Differential Response in Child Protective Services: A Literature Review November 2011 evaluations of Missouri’s Family Assessment and Response program and Minnesota’s alternative response program, found modest reductions in subsequent removal from the home, while other evaluations did not. One study examining educational outcomes found no effect, based on reports by parents. In most studies, parents in the non-investigation pathway had a more favorable attitude toward CPS services, were more likely to report being positively engaged, and were more likely to receive services earlier in the process. Only the evaluation of Minnesota’s program examined family economic hardship, and found families participating in the non-investigation pathway were less stressed regarding their economic well-being a year after receiving services than were their counterparts that had received investigations. Among the evaluations that looked at agency staffing issues, most found greater worker satisfaction under the non-investigation pathway than under the investigation pathway. Workers also reported that participating in the non-investigation pathway was an effective strategy for meeting client needs. Contrary to the expectations of some jurisdictions, workload reduction was not found, and some increase in workload was found for non-investigation pathway social workers in a few jurisdictions. Cost. According to Minnesota’s cost evaluation, long-term CPS costs were found to be lower for the non-investigation pathway than for the investigation pathway, with an average follow-up for 3.6 years per family (Siegel & Loman, 2006). However, another cost evaluation from Ohio found slightly higher costs for the non-investigation pathway in a variable follow-up period that had a 15-month maximum (Loman, Filonow & Siegel, 2010). Additional long-term cost evaluations are needed to demonstrate cost effectiveness of differential response-organized CPS systems. Where Are the Gaps? While program descriptions and evaluation research have addressed many important questions about differential response, many questions remain to be answered, and those answers would considerably strengthen the knowledge base supporting differential response policy and practice. Which aspects of differential response implementation are most effective in improving outcomes for children and families? While it is clear that participation in a noninvestigation pathway can produce positive outcomes, little is known about specifically what makes this effective. For example, the strategies for family engagement, the criteria and protocol for pathway assignment, and the array of services offered may vary across approaches in different locations. Each aspect of implementation may affect outcomes differently. Increased documentation of the interactions that occur between workers and families would be useful. Evaluations that systematically compare diverse differential response implementation approaches would help identify the most effective practices. This would be a major contribution to future differential response policy and practice. Are the positive effects on families due to assignment to a non-investigation pathway or due to actual service provision? Existing research tells us little as to whether the positive effects of differential response result from the positive family engagement that occurs in a nonVersion 2 40 Differential Response in Child Protective Services: A Literature Review November 2011 investigation pathway, or from the provision of needed services. If the simple avoidance of an investigation and engagement with a social worker creates a positive effect, future differential response implementations should focus on appropriate pathway assignment so that as many families as possible can benefit from this approach. If services are determined to be the major predictor of positive outcomes, policies should focus on improving service quality for families participating in either pathway. How do the criteria for assignment into the investigation or non-investigation pathway influence effectiveness of the non-investigation pathway? States and counties vary considerably in their policies regarding case characteristics that make a family suitable for the non-investigation pathway. As a result, the characteristics of families assigned to the noninvestigation pathway may vary substantially across jurisdictions. For example, some may have non-investigation pathway families with much lower overall risk levels than other jurisdictions. This selectivity may result in different overall levels of effectiveness across such sites, but it is largely unstudied. How does participation in the non-investigation pathway differentially affect families with different demographic, social, or cultural characteristics? While certain jurisdictions have addressed the effects of participating in the non-investigation pathway related to race and culture (Hawaii and Ohio) or differences in maltreatment type (Olmsted County, Minnesota), more exploration of the differential impact of participating in the non-investigation pathway related to child, family, or case characteristics is warranted. Does assignment to a non-investigation pathway affect child and family well-being beyond basic safety? In the non-investigation pathway, caseworkers work with families to address factors such as financial issues, parent mental health, substance abuse, or parenting practices. These factors are addressed in order to prevent future child maltreatment and increase child safety. However, families and children may experience additional benefits. Further exploration of a wider range of child and family outcomes, including mental and physical health, education, and labor force participation would contribute to a greater understanding of differential response benefits. What is the total cost-effectiveness of differential response when costs to other service and support systems are considered? Existing differential response cost-effectiveness studies were limited to a comparison of CPS costs in the investigation and non-investigation pathways. However, since service provision in the non-investigation pathway often includes services from other agencies, some of which may not be paid for by CPS, it would be important to know whether differential response decreases or increases costs among affected service agencies as a whole. Partnerships between CPS and other agencies are an important part of the differential response strategy, and future partnerships would be made easier if there were solid evidence that it ultimately reduced costs to partners as well as CPS. Will the key findings for child and family outcomes hold up under more rigorous evaluation designs? To date, three experimental design evaluations have been completed (Minnesota, Ohio, and Onondaga, New York), with three additional in planning phases for the Version 2 41 Differential Response in Child Protective Services: A Literature Review November 2011 QIC-DR sites. Results from experimental design have their own limitations, but are the most scientifically robust and therefore very valuable to the field. What is the impact on the CPS system as a whole when multiple pathways are incorporated into the agency’s response to allegations of maltreatment? There is some evidence that the addition of a non-investigation pathway can result in more families being served by the CPS system. Additional research could be conducted to assess the impact of these pathways on the total number of families screened-in for a response; the numbers of families that receive services, regardless of pathway; the proportion of families whose reports were substantiated when additional pathways were added; and the overall impact on the number of children considered victims of maltreatment. In other words, while research has focused on the immediate issue of maintaining child safety, not much is known regarding the impact on the community when CPS takes a differential response approach. Long-term impacts on the agency itself are also, as yet, relatively unstudied. Will the type of workers recruited to work in CPS change? Will worker retention rates increase? What will be the impact on the ongoing service units of the agency? In other words, will fewer families require ongoing in-home services, if significant numbers receive services upfront? Will foster care use be significantly reduced? Conclusion The questions identified above provide a rich road map for future differential response research. We recommend further research into the actual activities and services in the non-investigation pathway. Specific features of the approach, such as the length of time a worker is engaged with the family, the focus of the engagement on the caregivers or the children, further service provision by child welfare and other agencies to families that receive a non-investigation response, and caseload and workload of the CPS agency, have not been fully explored. We also recommend that, if non-investigation is to become a model approach which can be replicated, additional documentation on actual interactions with families be gathered. Given a more solid understanding of the detailed features of the intervention known as non-investigation, additional evaluation studies will prove to be useful. Such research will considerably strengthen the knowledge base for differential response policy and practice for the coming years. For More Information: For the most up-to-date resources on differential response, visit the National Quality Improvement Center on Differential Response in Child Protective Services website at www.DifferentialResponseQIC.org. Version 2 42 Differential Response in Child Protective Services: A Literature Review November 2011 REFERENCES ACTION Council of Monterey County and Monterey County Department of Family and Children's Services. (2008, June). Partnership and innovation: A program and data review of P2S in Monterey County. Salinas, CA: Authors. Retrieved from http://mcdses.co.monterey.ca.us/reports/downloads/P2S_PrgramandDataReview_C ombinedReport_6-20.pdf Advocates for Children & Youth. (2008). Alternative response key to reducing child welfare disparities. Retrieved from http://www.acy.org/articlenav.php? id=246 Anselmo, S., Pickford, R., & Goodman, P. (2003). Alberta response model: Transforming outcomes for children and youth. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 98-104). Ottawa, ON: Child Welfare League of Canada. Bagdasaryan, S., Furman, W., & Franke, T. (2008). Implementation of California’ differential response model in small counties. Protecting Children, 23(1 & 2), 40-56. Barber, J., & Knoke, D. (2003). Evaluating the implementation of assessment tools in the Australian child protection system. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 49-63). Ottawa, ON: Child Welfare League of Canada. Berrick, J. D., Bryant, M., Conley, A., de Elizalde, L., Garcia, V., & Greer, A., … Price, A. (2008). Differential response and alternative response in diverse communities: An empirically-based curriculum. Berkeley, CA: University of California at Berkeley, Center for Child and Youth Policy. Retrieved from http://www.csulb.edu/projects/ccwrl/Differential%20Response%201023.pdf Bowlby, J. (1969). Attachment and loss: Vol. 1: Attachment. New York: Basic Books. Brubacher, M. D., & Narayan, J. (2003). Community based child welfare services in Guelph and Wellington County. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 89-97). Ottawa, ON: Child Welfare League of Canada. Brunson, L., & Bouchard, C. (2003). Mobilizing communities to prevent child abuse and neglect: A cultural shift in child protection. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 75-88). Ottawa, ON: Child Welfare League of Canada. Version 2 43 Differential Response in Child Protective Services: A Literature Review November 2011 California Social Work Education Center. (2003). Early intervention and differential response: Decision points where fairness and equity can be addressed and evaluated. Berkeley, CA: Author. Cameron, G., Freymond, N., & Roy, C. (2003). Avenues for positive innovations in Canadian child welfare: Lessons from the Partnerships for Children and Families Project and international jurisdictions. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 14-31). Ottawa, ON: Child Welfare League of Canada. Child Abuse Prevention and Treatment Act (CAPTA) Reauthorization Act of 2010 (P.L. 111320, 12/20/10). Carl Vinson Institute of Government. (2009). Differential response/family support services: Policy analysis and recommendations. Athens, GA: Author. Carpenter, C. (2007, Spring). Alternative response. Children, families, and the courts: Ohio bulletin, 3(3), 1-13. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-ar-ohiochild-law-bulletin.pdf Carpenter, C. (2010). Process perspectives: Chronicling Ohio’s alternative response pilot project. Englewood, CO: American Humane Association. Retrieved from http://www.americanhumane.org/assets/pdfs/children/oh-alternative-response-section4.pdf Casey Family Programs. (2007). Implementing differential response in California: Promising practices and lessons learned. Seattle, WA: Author. Retrieved from http://www.casey.org/Resources/Publications/BreakthroughSeries_DifferentialResp onse.htm Center for Child and Family Policy. (2004, April). Multiple response system (MRS) evaluation report to the North Carolina Division of Social Services (NCDSS). Raleigh, NC: Sanford Institute of Public Policy, Duke University. Retrieved from http://www.ncdhhs.gov/dss/mrs/docs/MRS%20Evaluation%20Report%202004.pdf Center for Child and Family Policy. (2006, June). Multiple response system (MRS) evaluation report to the North Carolina Division of Social Services (NCDSS). Raleigh, NC: Sanford Institute of Public Policy, Duke University. Retrieved from http://www.ncdhhs.gov/dss/publications/docs/mrs_eval_rpt_6_30_06_all_combined. pdf Child and Family Policy Institute of California. (2006). Child welfare services system improvements: 11 county pilot implementation evaluation: Initial assessment phase, July 2003 to June 2006. Sacramento, CA: California Department of Social Services, Version 2 44 Differential Response in Child Protective Services: A Literature Review November 2011 Children and Family Services Division. Retrieved from http://www.cfpic.org/children/pdfs/11_County_Eval_Phase1.pdf Child Welfare Information Gateway. (2008). Differential response to reports of child abuse and neglect. Washington, DC: U.S. Department of Health and Human Services. Retrieved from http://www.childwelfare.gov/pubs/issue_briefs/differential_response Child Welfare Services Stakeholder Group. (2002). Part V: Report of the early intervention and differential response strategies. In CWS Redesign: Conceptual Framework (46-69). Sacramento, CA: Author. Retrieved from http://www.cdss.ca.gov/cdssweb/res/cws/pdf/progrpt2002.pdf Chipley, M., Sheets, J., Baumann, D., Robinson, D., & Graham, J. C. (1999). Flexible response evaluation. Austin, TX: Texas Department of Protective and Regulatory Services. Christenson, B., Curran, S., DeCook, K., Maloney, S., & Merkel-Holguin, L. (2008). The intersection between differential response and family involvement approaches. Protecting Children, 23(1 & 2), 88-95. Clavel, G., Cadieux, L., & Roy, C. (2003). The inclusive approach of the Outaouais Centres Jeunesse. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 112-118). Ottawa, ON: Child Welfare League of Canada. Comer, D. (2008, November). The six principles of partnership: A foundation for differential response [PowerPoint slides]. Presentation at the American Humane Association Conference on Differential Response, Columbus, OH. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/the-sixprinciples.pdf Comer, D. P., & Vassar, D. (2008). Six principles of partnership: Building and sustaining system-wide change. Protecting Children, 23(1 & 2), 96-104. Conley, A. (2007). Differential response: A critical examination of a secondary prevention model. Children and Youth Services Review, 29, 1454-1468. Conley, A., & Berrick, J.D. (2008). Implementation of differential response in ethnically diverse neighborhoods. Protecting Children, 23(1 & 2), 30-38. Conley, A. & Berrick, J.D. (2010). Community-based child abuse prevention: Outcomes associated with differential response program in California. Child Maltreatment,15(4), 282-292. Version 2 45 Differential Response in Child Protective Services: A Literature Review November 2011 Connecticut Department of Children and Families. (2009). CT DCF differential response system (DRS) status of work matrix. Hartford, CT: Author. Retrieved from http://www.ct.gov/dcf/lib/dcf/drs/pdf/drs_status_of_work_matrix.pdf Connolly, M. (2005). Differential responses in child care and protection: Innovative approaches in family-centered practice. Protecting Children, 20(2 & 3), 8-20. Connolly, M. (2007, November). Practicing for outcomes: Differential response model, New Zealand style [PowerPoint slides]. Presentation at the American Humane Association Conference of Differential Response, Long Beach, CA. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-drmkeynote-presentation.ppt Connolly, M. & Smith, R. (2010). Reforming child welfare: An integrated approach. Child Welfare, 89(3), 9-21. Costello, T. (1998). West Virginia Family Options Initiative: Final pilot evaluation report. Unpublished report. Dhillon, A. (2005). Keeping families together and safe: A primer on the child protection-housing connection. Washington DC: Child Welfare League of America. Retrieved from http://www.cwla.org/programs/childprotection/childprotectionhousing.pdf Drury, I., Winokur, M., Batchelder, K. & Mackert, M. (in press). Decision point: Screening practice as the foundation for differential response. Protecting Children. 26(4). Dudding, P. (2003). Foreword. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 32-48). Ottawa, ON: Child Welfare League of Canada. English, D., Fluke, J. D., & Yuan, Y-Y. T. (2003). Alternative response to child protective services investigations in the United States. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 64-74). Ottawa, ON: Child Welfare League of Canada. English, D. J., Wingard, T., Marshall, D., Orme, M., & Orme, A. (2000). Alternative responses to child protective services: Emerging issues and concerns. Child Abuse & Neglect, 24(3), 375-387. Florida Department of Children and Families (2009) Florida’s Alternative Response Pilot Final Summary Report. Retrieved from: http://centerforchildwelfare.fmhi.usf.edu/kb/trdiver/ARS_Final_Report_4-3-09.doc.pdf Version 2 46 Differential Response in Child Protective Services: A Literature Review November 2011 Florida Senate. (2003, November). Retention of protective investigators phase II. Interim Project Report 2004-112. Tallahassee, FL: Author. Foxcroft, D., & Blackstock, C. (2003). Usma: Cherished ones, precious ones, the children: A First Nations approach to child, family, and community well-being. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 105-111). Ottawa, ON: Child Welfare League of Canada. FRIENDS National Resource Center for Community-Based Child Abuse Prevention. (2007, September). Alternative response systems: Learning tool 13. Chapel Hill, NC: Author. Retrieved from http://www.friendsnrc.org/joomdocs/ars.pdf Hardin, M. (1996, Winter). Responsibilities and effectiveness of the juvenile court in handling dependency cases. The Future of Children, 6(3), 111-125. Retrieved from http://www.princeton.edu/futureofchildren/publications/docs/06_03_08.pdf Hilker, K. (2010). Tennessee Department of Children’s Services: Multiple response system 2010 preliminary evaluation. Murfreesboro, TN: Tennessee Center for Child Welfare, Middle Tennessee State University. Huebner, R. A. (2005, August 21). Program evaluation of the multiple response system: Kentucky. Unpublished report. Huebner, R. A., Durbin, L., Brock, A. (2009, April 3). Program evaluation of the multiple response system Kentucky Department for Community Based Services. Unpublished report. Institute of Applied Research. (2005, April). Effectiveness and costs of the new paradigm: Alternative response in Minnesota. Presentation at the 15th National Conference on Child Abuse and Neglect, Boston, MA. Retrieved from http://www.iarstl.org/papers/MnPresBostonApril05.pdf Institute of Applied Research. (2006, November). Poverty, chronic neglect, and differential response. Presentation at the National Conference on Differential Response, San Diego, CA. Johnson, C., Sullivan Sutton, E., & Thompson, D. M. (2005). Child welfare reform in Minnesota. Protecting Children, 20(2 & 3), 55-60. Jones, H., Chant, E., & Ward, H. (2003). Integrating children’s services: A perspective from England. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 119-131). Ottawa, ON: Child Welfare League of Canada. Version 2 47 Differential Response in Child Protective Services: A Literature Review November 2011 Kaplan, C., & Merkel-Holguin, L. (2008). Another look at the national study on differential response in child welfare. Protecting Children, 23(1 & 2), 5-21. Kaplan, C. & Rohm, A. (2010) Ohio alternative response pilot project: Final report of the AIM team. Englewood, CO: American Humane Association. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-dr-ohiosection1-aim-final-report.pdf Kaplan, C., & Schene, P. (2008, November). Primer on differential response: Take two! [PowerPoint slides]. Presentation at the American Humane Association Conference on Differential Response, Columbus, OH. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-dr-confpresentation2008.pdf Kaplan, C., Shannon, C., & Loman, T. (2008, December). What a difference a year makes: Charting the past, present & future of the Ohio alternative response project [PowerPoint slides]. Presentation before the Subcommittee on Child Abuse, Neglect, and Dependency, Columbus, OH. Retrieved from http://www.iarstl.org/papers/OHarPres20081208.pdf Kearney, K., Fuller, T., Jones, W. & McEwen, E. (in press). Putting it all together: Lessons learned from implementing differential response in Illinois. Protecting Children. 26(4). Kelley, D., Konopka, R., & Baker, C. (2009). Differential response. In Slack, K.S., Jack, K.M., & Gjertson, L.M. (Eds.) Child maltreatment prevention: toward an evidence-based approach (47-52). Madison, WI: Institute for Research on Poverty, University of Wisconsin-Madison. Retrieved from http://www.irp.wisc.edu/research/WisconsinPoverty/pdfs/ChildMaltreatment-Final.pdf Kirk, R. S. (2008). Development and field testing of a family assessment scale for use in child welfare practice settings utilizing differential response. Protecting Children, 23(1 & 2), 71-87. Larsen-Rife, D. & Brooks, S. (2009). The importance of family engagement in child welfare services. Davis, CA: Northern California Training Academy. Retrieved from http://academy.extensiondlc.net/file.php/1/resources/LR-FamilyEngagement.pdf Lawrence, N. & Snyder, E. (2009, Spring). Multiple Response System and System of Care: Two policy reforms designed to improve the child welfare system. Durham, NC: Center for Child and Family Policy, Duke University. Retrieved from http://childandfamilypolicy.duke.edu/pdfs/news/PolicyBrief_mrssoc.pdf Lemon, K., D'Andrade, A. & Austin, M. (2005). Understanding and addressing disproportionality in the front end of the child welfare system. Berkley, CA: Bay Area Social Services Consortium. Retrieved from http://cssr.berkeley.edu/bassc/public/DISPRO_PDF.pdf Version 2 48 Differential Response in Child Protective Services: A Literature Review November 2011 Lohrbach, S., Sawyer, R., Saugen, J., Astolfi, C., Schmitt, K., Worden, P., et al. (2005). Ways of working in child welfare: A perspective on practice. Protecting Children, 20(2 & 3), 93100. Loman, L. A. (2005, May). Differential response improves traditional investigations: Criminal arrests for severe physical and sexual abuse. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-dr-severe-a buse-mopdf.pdf Loman, L. A. (2006). Families frequently encountered by the child protection services: A report on chronic abuse and neglect. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/FEfamiliesChronicCAN.pdf Loman, L.A. (2007, November). Poverty, child neglect, and differential response. Presentation at the American Humane Association Conference on Differential Response, Long Beach, CA. Loman, L. A., Filonow, C.S., & Siegel, G. (2010). Ohio alternative response evaluation: Final report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/pc-dr-ohiosection2-final-evaluation-report-1.pdf Loman, L.A., Shannon, C., Sapokaite, L., & Siegel, G. (2009). Minnesota parent support outreach program evaluation: Final report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/PSOPFinalReport.pdf Loman, L. A., & Siegel, G. L. (2004a, February). Differential response in Missouri after five years: Final report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/MODiffResp2004a.pdf Loman, L. A., & Siegel, G. L. (2004b, November). Minnesota alternative response evaluation: Final report. Saint Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/ARFinalEvaluationReport.pdf Loman, L. A., & Siegel, G. L. (2005). Alternative response in Minnesota: Findings of the program evaluation. Protecting Children, 20(2 & 3), 78-92. Loman, L. A. & Siegel, G. (2005, March). Alternative response research in Missouri, Minnesota, and Virginia: Findings in six areas [PowerPoint slides]. Saint Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/CAPresentation.pdf Marshall, S.K., & Charles, G. (2010). Comparing differential responses within child protective services: A longitudinal examination. Child Welfare, 89(3), 57-77. Version 2 49 Differential Response in Child Protective Services: A Literature Review November 2011 Marts, E. J., Lee, E.-K. O., McRoy, R., & McCroskey, J. (2008). Point of engagement: Reducing disproportionality and improving child and family outcomes. Child Welfare, 87(2), 335358. Maryland Department of Human Resources. (2008). Analysis of the FY 2010 Maryland executive budget, 2009. Retrieved from http://mlis.state.md.us/2009RS/budget_docs/All/Operating/N00B__DHR_Social_Services.pdf Maryland Department of Legislative Services. (2008). Maryland General Assembly fiscal and policy note: House bill 262. Retrieved from http://mlis.state.md.us/2008rs/fnotes/bil_0002/hb0262.pdf Merkel-Holguin, L., Kaplan, C., & Kwak, A. (2006, November). National study on differential response in child welfare. Englewood, CO: American Humane Association and Child Welfare League of America. Retrieved from http://www.americanhumane.org/assets/pdfs/children/pc-2006-national-studydifferential-response.pdf Michigan Department of Human Services. (2008). Children’s protective services investigation process. Retrieved from http://www.michigan.gov/dhs/0,1607,7-124-5452_7119_7194159484--,00.html Miller, D.W. (2008, Fall). Alternative response in child welfare. The Advocate: The Newsletter of Division 37, 31(3), 4-6. Washington, DC: American Psychological Association. Retrieved from http://www.apa.org/divisions/div37/AdvocateFall2008.pdf Minnesota Department of Human Services. (2005, April 4). Minnesota's child welfare report for 2003: Bulletin #05-68-03. Saint Paul, MN: Author. Retrieved from http://www.dhs.state.mn.us/main/groups/children/documents/pub/dhs16_141408.pdf Monterey County Department of Family and Children's Services. (2006). Differential response phase I: A report on the Monterey County differential response planning process. Salinas, CA: Author. Retrieved from http://mcdses.co.monterey.ca.us/reports/downloads/DR_Report_Final.pdf National Center for Adoption Law and Policy. (2010). Recommendations for a differential response statutory/rule framework in Ohio. Columbus, OH: Author. Retrieved from http://www.law.capital.edu/adoption/AR/Section%203%20Recommendations%20fo r%20Rule%20Statute%20Framework.pdf National Child Welfare Resource Center for Family-Centered Practice. (2001). Best practice/next practice: Family-centered child welfare [Newsletter]. Washington, DC: Author. Retrieved from http://www.hunter.cuny.edu/socwork/nrcfcpp/downloads/newsletter/BPNPSpring01.pdf Version 2 50 Differential Response in Child Protective Services: A Literature Review November 2011 National Conference of State Legislatures. (2010). Differential response approach: State enacted legislation [Table]. Retrieved from http://www.ncsl.org/?tabid=17800 National Quality Improvement Center on Differential Response in Child Protective Services. (2009, September). Differential response in child protective services: A legal analysis. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/differential-response-in.pdf National Quality Improvement Center on Differential Response in Child Protective Services. (2009, March). Information summit on disproportionality: Final report. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/disproportionality-info-summitreport-final.pdf National Quality Improvement Center on Differential Response in Child Protective Services. (2009, February). Information summit on policy and practice: Final report. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/policypractice-information-summit-feb-09.pdf National Quality Improvement Center on Differential Response in Child Protective Services. (2009, July). Information summit on prevention: Final report. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/prevention-infosummit-report-feb-09.pdf National Quality Improvement Center on Differential Response in Child Protective Services. (2009, March). Information summit on research and evaluation: Final report. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/research-evaluation-reportfinal.pdf National Quality Improvement Center on Differential Response in Child Protective Services. (2009). Online survey of state differential response policies and practices findings report. Englewood, CO: Author. Retrieved from http://www.differentialresponseqic.org/assets/docs/qic-dr-findingsreportjun09.pdf North Carolina Department of Health and Human Services. (2004). North Carolina's multiple response system of child protective services. Raleigh, NC: North Carolina Division of Social Services, Family Support and Child Welfare Section. Retrieved from http://www.ncdhhs.gov/dss/mrs/docs/MRSReport2004.pdf Office of Children's Administration Research. (1998). Alternative response systems evaluation progress report. Seattle, WA: Unpublished report. Office of Children's Administration Research. (1999). Alternative response systems evaluation progress report. Seattle, WA: Unpublished report. Version 2 51 Differential Response in Child Protective Services: A Literature Review November 2011 Office of Children's Administration Research. (2000). Alternative response systems evaluation progress report. Seattle, WA: Unpublished report. Office of Children’s Administration Research. (2005). Alternative response systems program progress report. Seattle, WA: State of Washington Department of Social and Health Services, Children’s Administration Practice Improvement Division, Office of Children’s Administration Research. O’Neill Murray, K., & Gesiriech, S. (2008). A brief legislative history of the child welfare system. Washington, DC: The Pew Commission on Children in Foster Care. Retrieved from http://pewfostercare.org/research/docs/Legislative.pdf Ortiz, M. J., Shusterman, G. R., & Fluke, J. D. (2008). Outcomes for children with allegations of neglect who receive alternative response and traditional investigations: Findings from NCANDS. Protecting Children, 23(1 & 2), 57-70. Paxson, C. & Waldfogel, J. (2003). Welfare reforms, family resources, and child maltreatment. Journal of Policy Analysis and Management, 22(1), 85-113. Pennell, J., & Burford, G. (2000). Family group decision making: Protecting children and women. Child Welfare, 79(2), 131-158. Provincial and Territorial Directors of Child Welfare. (2003). New directions in child welfare. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 113). Ottawa, ON: Child Welfare of Canada. Richardson, J. (2008). Differential response literature review. Urbana, IL: Children and Family Research Center, University of Illinois at Urbana-Champaign. Ruppel, J., Huang, Y., & Haulenbeek, G. (2011). Differential response in child protective services in New York State: Implementation, initial outcomes, and impacts of pilot project. Report to the Governor and Legislature. New York State Office of Children and Family Services. Ryan, K. M. (2007). Differential response: Supporting families in crisis. Trenton, NJ: New Jersey State League of Municipalities. Retrieved from http://www.njslom.com/featart0407.html Sawyer, R., & Lohrbach, S. (2005a). Differential response in child protection: Selecting a pathway. Protecting Children, 20(2 & 3), 44-53. Sawyer, R., & Lohrbach, S. (2005b). Integrating domestic violence intervention into child welfare practice. Protecting Children, 20 (2 & 3), 62-77. Schene, P. (2001, Spring). Meeting each family’s needs: Using differential response in reports of child abuse and neglect. Best Practice/Next Practice: Family-Centered Child Welfare. Version 2 52 Differential Response in Child Protective Services: A Literature Review November 2011 Washington, DC: National Child Welfare Resource Center for Family-Centered Practice. Retrieved from http://www.hunter.cuny.edu/socwork/nrcfcpp/downloads/newsletter/BPNPSpring01.pdf Schene, P. (2005). The emergence of differential response. Protecting Children, 20(2 & 3), 4-7. Schene, P., & Kaplan, C. (2007, November). Getting started with differential response: Fundamentals and first steps [PowerPoint slides]. Presentation at the American Humane Association Conference on Differential Response, Long Beach, CA. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/c-dr-conf2007.pdf Schene, P., & Oppenheim, S. (2005). Choosing the path less traveled: Strengthening California families through differential response. What Works Policy Brief. Sacramento, CA: Foundation Consortium for California’s Children and Youth. Retrieved from http://www.chhs.ca.gov/initiatives/CAChildWelfareCouncil/Documents/Strengtheni ngCAFamiliesthroughDifferentialResponse.pdf Shusterman, G. R., Fluke, J. D., Hollinshead, D. M., & Yuan, Y-Y. T. (2005). Alternative responses to child maltreatment: Findings from NCANDS. Protecting Children, 20(2 & 3), 32-42. Shusterman, G. R., Hollinshead, D., Fluke, J. D., & Yuan, Y-Y. T. (2005, July). Alternative responses to child maltreatment: Findings from NCANDS. Washington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Retrieved from http://aspe.hhs.gov/hsp/05/child-maltreatresp/report.pdf Shusterman, G., & Ortiz, M. J. (2008, November). Using quantitative methods to guide qualitative research in differential response. Presentation at the American Humane Association Conference on Differential Response in Child Welfare, Columbus, OH. Siegel, G. L., & Loman, L. A. (1997, November). Missouri family assessment and response demonstration: Final evaluation report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/MO%20FAR%20Final%20Reportfor%20website.pdf Siegel, G. L., & Loman, L. A. (2000, January). The Missouri family assessment and response demonstration impact evaluation: Digest of findings and conclusions. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/MoFamAssess.pdf Siegel, G., & Loman, L. (2002). Alternative response evaluation first annual report: Executive summary. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/MnARexecsum.pdf Version 2 53 Differential Response in Child Protective Services: A Literature Review November 2011 Siegel, G., & Loman, L. (2003). Minnesota alternative response evaluation second annual report: Executive summary. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/mnARExecSum2.pdf Siegel, G. & Loman, L. (2006). Extended follow-up study of Minnesota’s Family Assessment Response: Final report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/FinalMNFARReport.pdf Siegel, G. L., Loman, L. A., Cline, J., Shannon, C., & Sapokaite, L. (2008, November). Nevada differential response pilot project: Interim evaluation report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/Nevada%20Differential%20Response%20Pilot%20ProjectInterim%20Report%20November%202008.pdf Siegel, G. L., Filonow, C. S., & Loman, L. A. (2010, December). Differential response in Nevada: Final evaluation report. St. Louis, MO: Institute of Applied Research. Retrieved from http://www.iarstl.org/papers/NevadaDRFinalReport.pdf Sphere Institute. (2006). Implementing differential response: An assessment of community organizations' capacity and interest. Burlingame, CA: Author. Retrieved from http://www.sphereinstitute.org/publications/DR_Report_Final.pdf Stanley, T. (2007). Risky work: Child protection practice. Social Policy Journal of New Zealand, 30, 163-177. Retrieved from http://www.msd.govt.nz/documents/about-msd-and-ourwork/publications-resources/journals-and-magazines/social-policy-journal/spj30/30pages163-177.pdf Thompson, D., Siegel, G. L., & Loman, L. A. (2008). The Parent Support Outreach Program: Minnesota’s early intervention track. Protecting Children, 23(1 & 2), 23- 29. Trocmé, N., & Chamberland, C. (2003). Re-involving the community: The need for a differential response to rising child welfare caseloads in Canada. In N. Trocmé, D. Knoke, & C. Roy (Eds.), Community collaboration and differential response: Canadian and international research and emerging models of practice (pp. 32-48). Ottawa, ON: Child Welfare League of Canada. Trocmé, N., Knott, T. & Knoke, D. (2003). An overview of differential response models. Toronto, ON: Centre of Excellence for Child Welfare, University of Toronto. Retrieved from http://dev.cecw-cepb.ca/files/file/en/DifferentialResponse4E.pdf U.S. Department of Health and Human Services and Children's Bureau. (2003, April). National study of child protective services systems and reform efforts: Review of state CPS policy. Washington, DC: Author. Retrieved from http://aspe.hhs.gov/hsp/cps-status03/statepolicy03/chapter5.htm Version 2 54 Differential Response in Child Protective Services: A Literature Review November 2011 Vermont Department for Children and Families. (2007, November 1). Report and recommendations to the legislature: Act 77, an act relating to the child abuse registry and sex offender registry requirements. Waterbury, VT: Author. Virginia Department of Social Services. (2006, December 15). Evaluation of the differential response system. Richmond, VA: Author. Retrieved from http://www.dss.virginia.gov/files/about/reports/children/cps/allother/2006/differentialresponsesystem_annualreport_2006.pdf Virginia Department of Social Services. (2007, December). Evaluation of the differential response system. Richmond, VA: Author. Retrieved from http://www.dss.virginia.gov/files/about/reports/children/cps/allother/2007/differentialresponsesystem_evaluation_annualreport_2007_12-07.pdf Virginia Department of Social Services. (2008, December). Evaluation of the differential response system. Richmond, VA: Author. Retrieved from http://www.dss.virginia.gov/files/about/reports/children/cps/allother/2008/differentialresponsesystem_evaluation_annualreport_2008_12-08.pdf Waldegrave, S. & Coy, F. (2005). A differential response model for child protection in New Zealand: Supporting more timely and effective responses to notifications. Social Policy Journal of New Zealand, 25, 32-48. Retrieved from http://www.msd.govt.nz/documents/about-msd-and-our-work/publicationsresources/journals-and-magazines/social-policy-journal/spj25/25-pages32-48.pdf Waldfogel, J. (1998, Spring). Rethinking the paradigm for child protection. The Future of Children, 8(1), 104-119. Waldfogel, J. (2001). Protecting children in the 21st century. Family Law Quarterly, 34(3), 311328. Walter R. McDonald & Associates, Inc. (2001, March). National study of child protective services systems and reform efforts: Literature review. Rockville, MD: Author. Retrieved from http://aspe.hhs.gov/hsp/protective01 Walter R. McDonald & Associates, Inc. (2003). National study of child protective services systems and reform efforts: Site visits report. Rockville, MD: Author. Retrieved from http://aspe.hhs.gov/hsp/CPS-status03/site-visits/index.htm Walters, J., Beatty, D., Livingston, A. H., & Lau, W. (2007). Hawaii's differential response system: Practical implementation strategies that led to successful outcomes [PowerPoint slides]. Presentation at the Children’s Bureau Conference for Agencies and Courts in Arlington, VA. Retrieved from http:cfsr.dhshawaii.net/documents/download/33 Version 2 55 Differential Response in Child Protective Services: A Literature Review November 2011 Washington State Department of Social & Health Services. (2008, March). Report to the legislature: Consideration of a differential response in Washington State's child protection system. Olympia, WA: Author. Weiden, T., Nutter, B., Wells, L., & Sieppert, J. (2005). Alberta response model implementation evaluation phase 1: Baseline data. Calgary, AB: Centre for Social Work Research and Professional Development, University of Calgary. Weiser, B., & Dreitzer, D. (2008, November). Partnerships and pitfalls: Nevada differential response program [PowerPoint slides]. Presentation at the American Humane Association Conference on Differential Response, Columbus, OH. Retrieved from http://www.americanhumane.org/assets/pdfs/children/differential-response/partnershipsand-pitfalls.pdf Westat, Inc. (2009). Recent trends in local child protective services practices. Rockville, MD: Author. Retrieved from http://aspe.hhs.gov/hsp/09/TrendsinCPS/index.pdf Wheeler, C. E., & Johnson, S. (2003). Evaluating family group decision making: The Santa Clara example. Protecting Children, 18(1 & 2), 65-69. Wright, M., Tickler, S., & Vernor, K. (2008). California child welfare services: eleven-county pilot project evaluation report. Santa Rosa, CA: The Results Group. Retrieved from http://www.cwda.org/downloads/11CountyPilot2008.pdf Yuan, Y-Y. T. (2005). Potential policy implications of alternative response. Protecting Children, 20(2 & 3), 22-31. Zielewski, E. H., Macomber, J., Bess, R., & Murray, J. (2006). Families’ connections to services in an alternative response system. Washington, DC: Urban Institute. Version 2 56