Many child protective services (CPS) agencies face a large volume of CPS reports, increasingly complex cases, and strained resources. Moreover, there is a growing recognition that many CPS cases do not require a traditional investigation. Thus, many States have developed practices and policies to differentiate how particular types of cases are handled (U.S. Government Accounting Office, 1997). Referred to as alternative response, differential response, dual track, or family assessment, these efforts at system reform promote new practices that affect how agencies respond to certain reports of maltreatment. Generally, investigation responses involve a more forensic approach and include processes for determining if child maltreatment occurred or if a child is at risk of child maltreatment. Alternative responses are characterized by an emphasis on an assessment of the needs of families and children with little emphasis on determining if a specific incident or condition of maltreatment occurred (U.S. Department of Health and Human Services, 2003a).
This analysis examines case-level data reported in 2002 to the National Child Abuse and Neglect Data System (NCANDS) by six States that offered both alternative response and traditional investigation as part of their child welfare services. The objective of this study was to compare the children in each State who were referred to alternative response systems with those referred to traditional investigations. Comparisons were made in terms of response characteristics, the circumstances of their reported maltreatment, and their subsequent reports and dispositions.
This study examined the following research questions:
- What are the characteristics of children who received an alternative response?
- How are the circumstances of the reported maltreatment related to the chances that a child receives an alternative response or an investigation response?
- How do outcomes differ between children who receive an alternative response and children who receive an investigation response?
Case-level data from the 2002 NCANDS Child File were used as the basis for the analysis. These case-level data are submitted on a voluntary basis in a common record format to the Federal Government by State CPS agencies. The submissions are a rich source of information about children who are the subjects of child maltreatment investigations, including data about the investigations, child demographics, types of maltreatment, and services.
Data from six States — Kentucky, Minnesota, Missouri, New Jersey, Oklahoma, and Wyoming — including 313,838 reported children, were used to create the data set for this research. These States were chosen because case-level data, including alternative response dispositions, were provided in sufficient numbers. Several new variables were derived, and some were recoded into categorical variables. Most importantly, the report disposition was used to create a new variable, called response type. Reports with dispositions of substantiated, indicated, unsubstantiated, or closed with no finding were coded as investigation response.
Dispositions of alternative response — victim and alternative response — nonvictim were coded as alternative response. All other dispositions were coded as other response; these cases were later excluded from analyses. The final data set was composed of 313,838 children in 6 States. Because of major differences in State policy and implementation, and the level of use of alternative response as reported to NCANDS, data from each State were each analyzed separately.
Even though some States have been implementing alternative response systems for several years, no large-scale, multistate, data-based research has been undertaken on the children who have received alternative response as compared with children who have received investigations. Findings from this research show that in some ways these groups are quite similar; however, some differences can be identified. This study compared children who were referred to alternative response systems with those referred to traditional investigations in terms of their own characteristics, the circumstances of their reported maltreatment, and their subsequent reports and dispositions.
Figure 1 shows children who were the subject of a report by the type of response that their report received in each of the six States represented in the data set. The number of children who were included in a maltreatment report, as well as the proportion of children who received an alternative response, varied across these States. Kentucky, Missouri, New Jersey, and Oklahoma all had large number of children who were subjects of a maltreatment report (between 60,000 and 80,000), but in Kentucky and Oklahoma, only 27 and 20 percent, respectively, were referred to alternative response, while in Missouri and New Jersey, 64 and 71 percent, respectively were referred to alternative response. Minnesota (26,344) and Wyoming (4,355) had much smaller number of children who were subjects of a report. In Minnesota, 20 percent were referred to alternative response, while in Wyoming, 58 percent were. Variation was noteworthy between States and reinforces the need to examine the States separately.
The study team reviewed states’ policy documents seeking guidance used within the state regarding the circumstances under which agencies will or will not provide alternative response. In New Jersey, where child welfare assessments are provided to families with a wide range of problems such as homelessness, domestic violence, child or parent substance abuse, their policies did not explicitly identify circumstances under which alternative response could not be provided. Other States take a more restrictive approach by aggressively extracting reports for which alternative response cannot be considered before identifying those for which it can. Alternatively, these differences may result from the style adopted by the State for describing its alternative response system. For example, in Wyoming — where 58 percent of reported children were referred to alternative response — the policy described only the circumstances precluding alternative response, not those that allow it.
States’ use of alternative response varied considerably. During 2002, referrals to alternative response ranged from 20 and 71 percent across these six States. Viewing these trends over 5 years, it appears that States were generally either increasing or maintaining levels of alternative response referrals. The use of alternative response appears to have an impact on the numbers of both victims and nonvictims identified by these States when comparing 1998–2002 disposition data. In general, the use of alternative response resulted in a decrease in the numbers of victims and nonvictims identified by States using alternative response. The impact of the alternative response system on States’ victim identification ranged from a 6 percent decrease (in a State in which the program is being piloted) to a 36 percent decrease. The impact of alternative response systems on the number of nonvictims identified generally was reflected by a decrease of nonvictims (ranging from 18% to 57%); however, the rate of nonvictims in Oklahoma rose by 30 percent.
This aspect of the research examined whether demographic characteristics — such as age, sex, race, and ethnicity — distinguished children who were referred to alternative response from those who received traditional investigations. Older children were more likely to receive an alternative response than were younger children. However, race and ethnicity did not strongly distinguish between children who received an alternative response and those who received an investigation. Some differences were evident in the likelihood of alternative response for children of different races; however, the variation in the number of children of different races within the CPS population — and the population as a whole — render these differences difficult to interpret. In general, the sex of the child was not a major factor in determining whether a child received alternative response or investigation, although in Missouri and Wyoming, boys were more likely to be referred to alternative response than were girls.
In most of the States, prior victimization was related to a decreased likelihood of alternative response. In Minnesota and Missouri, this difference was quite dramatic — none of the children with prior victimization received an alternative response. While researchers identified no written policies in these States prohibiting alternative response in cases with prior victimization, it is possible that such guidance existed outside the statutes or policy manuals reviewed. In Oklahoma and Kentucky, only 7 percent and 16 percent (respectively) of the children with prior victimization received an alternative response. In New Jersey only, children with prior victimization were equally likely to receive alternative response. It is plausible that concerns about a family’s responsiveness to system interventions, or about the chronic nature of problems, may factor into decisions about what response track is appropriate.
How are the circumstances of the reported maltreatment related to the chances that a child receives an alternative response or an investigation response?
State policy regarding criteria for the use of alternative response is based on the circumstances of the report, particularly on the type or severity of the alleged maltreatment. Previous research has shown that some circumstances of alleged maltreatment are related to whether a child receives an investigation or an alternative response, while others are not. Earlier studies have not reported on whether the other characteristics of the maltreatment report — such as number of children included in the report, living arrangement, or the presence of risk factors such as family violence or caretaker substance abuse — were associated in any way with a referral to alternative response. Findings from this study confirm that most of these variables are relevant in distinguishing which track is chosen.
Source of Report. Alternative response more often resulted from referrals from parents, relatives, friends, schools, or the children themselves. Referrals from social workers, medical personnel, legal, or criminal justice sources were less likely to be referred to alternative response.
Maltreatment Type. The connection between maltreatment type and referral to alternative response in each State was strong, but varied across States. For two States in this study — Missouri and Wyoming — all children referred to alternative response had the same maltreatment characteristics. In Missouri, all were children with no reported maltreatment. In other circumstances these reports might have been screened out, or referred to other agencies. In Wyoming, all had been reported for “other” forms of abuse, i.e. not the four types of maltreatment referred to in federal statutes: abuse, neglect, sexual abuse or emotional maltreatment. “Other” forms of abuse and neglect recognized by states vary, but may include, for instance, educational neglect and medical neglect. In the other States — Kentucky, Minnesota, New Jersey, and Oklahoma — a portion of children with all different maltreatment types were referred to alternative response. Only in New Jersey were significant numbers of children who were reported to be sexually abused referred to alternative response. This mix of results suggests that States are using alternative response differently from one another.
Other Circumstances of the Reported Maltreatment. In the States that provided data on the reported child’s living arrangement, children living at home with their families were more likely to be referred to alternative response than were children in foster care or institutional settings. In all States but Minnesota, a higher percentage of children were referred to alternative response when other children were included in the same maltreatment report than when only one child was included in the report. In Wyoming, this difference was large, suggesting that the inclusion of more than one child in a maltreatment report may be a factor for referring a child to alternative response, or that the inclusion of only one child may be a reason for conducting an investigation. However, in Minnesota, the reverse was true: Children who were alone in the maltreatment report were more likely to be referred to alternative response. In the family risk assessment of abuse and neglect that is used in Minnesota, families with more children are given a higher-risk score, resulting in a lower likelihood that they would be referred to alternative response.
In New Jersey only the presence of family violence and caretaker substance abuse were associated with an increase in the proportion of children referred to alternative response, which is consistent with the focus of New Jersey’s alternative response system on families with these issues. In all other States, the presence of family violence was not associated with an increase in the likelihood that a child would be referred to a response or a particular type. Also, children whose caretakers had a history of drug abuse had a lower rate of referral to alternative response than those children with no history of caretaker drug abuse.
How do outcomes differ between children who receive an alternative response and children who receive an investigation response?
This aspect of the research compares children who received alternative response with those who received investigations on two variables pertaining to the course of action following a report — whether children or their family members received services, if children were placed, and whether they experienced an additional report of maltreatment within 6 months and, if so, how the CPS system responded. Findings illustrate that these two groups differed on these outcome variables.
Services. In-home services were provided more often to children and families in the alternative response track. These findings may support the notion that families who are engaged using a less adversarial approach may be more inclined to utilize services that are offered. Moreover, if cases assigned to alternative response are those that present less serious needs, this trend may also reflect that more services are available to address the needs of these families compared to the families presenting more serious needs. Children were more likely to be placed in foster care if they received investigations. This study’ finding reinforces the idea that more serious cases are likely to be referred for investigation.
Reentry or Reresponse. The findings from this research demonstrate that the rate of recurrence within 6 months was comparable for children who received an alternative response and those who received an investigation, or, in the case of Oklahoma, the rate of reentry was lower. If children received an investigation as their initial response from CPS, the likelihood of their receiving an alternative response was lower for a second report, but not at all out of the question. Among children who were found to be victims, fewer than 5 percent were referred to alternative response for a subsequent report in most States; although in Missouri and New Jersey — with their high overall rate of alternative response — 9 and 11 percent of victims were subsequently referred to alternative response. Among nonvictims in New Jersey, 13 percent were subsequently referred to alternative response while in other States this rate was less than 10 percent.
In general, these findings demonstrate that implementation of an alternative response system reflects its intention — to serve children and families who appear to be at lower risk or who present less immediate safety concerns. The findings are consistent with the expectation that these families’ circumstances may not warrant a traditional CPS response, but can benefit from some intervention to prevent future maltreatment.
These analyses of child, report, and maltreatment characteristics suggest that States are implementing their alternative response systems somewhat differently. Some of this may be due to the stage and scope of implementation in each State. Other explanations may include the degree to which policies clearly specify how the response assignment is made. Some discretion by individual caseworkers is likely responsible for much of the variation between alternative response and investigations, as much as client and report characteristics. State demographics and availability of resources may also factor in the decisions made and outcomes observed.
It appears that services are being provided to a greater proportion of families who receive an alternative response. It also appears from this data that even though children who had been previously referred to alternative response do experience subsequent reports and responses by CPS, in general they are not at any greater risk for subsequent reports than those who received an investigation. Furthermore, they are not at any greater risk for subsequent victimization. With this knowledge, at the system level, agencies that refer children and families to the alternative response or investigation track may be confident that, if specified guidelines guide the decision, the child’s future safety is no more likely to be compromised.
Clearly, many factors influence the processes and outcomes of alternative response systems, and it may be helpful to more closely examine the interaction between these factors. This study provides a more textured understanding of alternative response systems across States and the outcomes associated with families and children who benefit from such systems.
|This Research Summary presents findings from an analysis of child abuse and neglect reports for six states that use both traditional child maltreatment investigations and some other defined action that does not require a specific finding about whether the maltreatment occurred. Several states have recently begun using these systems, referred to in this study as “alternative responses,” in an effort to differentiate among cases in which the often confrontational nature of investigations is helpful and those for which a more assessment oriented approach may be more constructive. The analysis described here was conducted by staff of Walter R. McDonald and Associates under contract to ASPE and in cooperation with the Administration for Children and Families.
Office of the Assistant Secretary for Planning and Evaluation
Michael J. O’Grady, Ph.D.
Barbara B. Broman
Contractor: Walter R. McDonald and Associates