Criminal law

http://cmx.sagepub.com

Child Maltreatment

DOI: 10.1177/1077559508318400 2009; 14; 38 originally published online Sep 15, 2008; Child Maltreat

Theodore P. Cross and Cecilia Casanueva Caseworker Judgments and Substantiation

http://cmx.sagepub.com/cgi/content/abstract/14/1/38 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

On behalf of: American Professional Society on the Abuse of Children

can be found at:Child Maltreatment Additional services and information for

http://cmx.sagepub.com/cgi/alerts Email Alerts:

http://cmx.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://cmx.sagepub.com/cgi/content/refs/14/1/38 Citations

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

38

Caseworker Judgments and Substantiation

Theodore P. Cross Cecilia Casanueva RTI International

Substantiation can have an important effect on what interventions are pursued for children investigated for maltreatment, but researchers lack knowledge about how the decision to substantiate is made. Using information from 4,515 children from a national probability study of children investigated for maltreatment, this study examined how caseworker judgments of harm, risk, and evidence predicted substantiation. The substantiation rate was 29.9%, but the majority of cases were substantiated when caseworkers reported at least moderate harm, at least moderate risk, and/or probably to clearly sufficient evidence. Each judgment variable significantly predicted substantiation in a multivariable model, with evidence the strongest predictor. Child gender and age were significant predictors beyond harm, risk, and evidence, suggesting that other judgments also influence substantiation. In 9 of 100 cases, reports were not substantiated despite moderate to severe harm. Thus, substantiation is generally based on judgments of harm, risk, and evidence but not exclusively. The findings underline previous researchers’ conclusions that substantiation is a flawed measure of child maltreatment and suggest that policy and practice related to substantiation are due for a fresh appraisal by state child welfare service agencies.

Keywords: caseworker judgment; substantiation; investigation; risk assessment; child protective services; NSCAW

Caseworker Judgments and Substantiation

Substantiation is the statement from child protective services about the validity of a report of child mal-

treatment, and most states have a statutory requirement for agencies in the child welfare system (CWS) to make a decision to substantiate or not substantiate (Walter R. McDonald & Associates, 2003). Substantiation can have an important effect on what interventions are pursued and how professionals and potentially the public react to child maltreatment allegations. CWS agencies and child abuse professionals have articulated criteria for substan- tiation and developed models of how the decision is made (Drake, 1996). Several studies in specific states or for specific populations have examined caseworkers’ thinking about substantiation in cases that they investi- gated (DePanfilis & Girvin, 2005; English, Marshall, Coghlan, Brummel, & Orme, 2002; Fluke, Hollinshead, & Walter R. McDonald & Associates, 2003). However, we are aware of no studies using national data sampled from the entire population of CWS investigations that describe what judgments caseworkers make and how they relate to substantiation. This is a noticeable gap given how consequential this decision is. This knowledge is

even more important if CWS eventually develops statis- tically based decision models to assign investigation priority scores to cases on the basis of the likelihood of substantiation (Zuravin, Orme, & Hegar, 1995). The stakes are high in child maltreatment investigations. Underestimation of risk leaves children unprotected; overestimation can lead to excessive intervention in families’ lives (Gambrill & Shlonsky, 2000). This study explores how substantiation is related to caseworker judgments of three key variables: the harm the child has suffered, the risk of harm to the child, and the sufficiency of the evidence for maltreatment. It also reports national estimates on the proportions of children in substantiated and unsubstantiated cases at different levels of harm, risk, and evidence of maltreatment.

Implications of Substantiation

Delivery of services may depend on whether a case is substantiated. The National Study of Child Protective

Child Maltreatment Volume 14 Number 1 February 2009 38-52

© 2009 Sage Publications 10.1177/1077559508318400

http://cm.sagepub.com hosted at

http://online.sagepub.com

Authors’ Note: Theodore Cross is now a visiting research specialist at the University of Illinois at Urbana-Champaign. Correspondence concerning this article should be addressed to Theodore Cross, School of Social Work, 1203 W Oregon St., Urbana, IL 61801; email: [email protected].

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Cross, Casanueva / Caseworker Judgments and Substantiation 39

Services Systems and Reform Efforts (Walter R. McDonald & Associates, 2003) found that 11 states limited service referrals to those cases in which maltreatment was substantiated. Recent analysis of the National Survey of Child and Adolescent Well-Being (NSCAW) found that children in substantiated cases were significantly more likely to be receiving CWS services 12 months after the investigation than children in unsubstantiated cases, even though fairly large percentages of both groups had cogni- tive, emotional, and behavioral difficulties and the groups did not differ significantly in well-being (U.S. Department of Health and Human Services, Administration on Children and Families, n.d.). It should be noted, however, that the causal link between substantiation and services may go in the opposite direction; some evidence suggests that availability of services is one factor that caseworkers consider in deciding whether to substantiate (see below; Walter R. McDonald & Associates, 2003).

Substantiation could also be used to inform court deci- sions about custody, although we do not know of any empirical studies of this effect. Substantiation clearly has emotional valence for those accused of child maltreatment. The Federal Child Abuse Prevention and Treatment Act (CAPTA) directs states to have an appeal process whereby reported perpetrators can appeal the substantiation decision (U.S. Department of Health and Human Services, Office on Child Abuse and Neglect, 2003). The fact that 45 states have policies for expunging unsubstantiated reports (Walter R. McDonald & Associates, 2003) adds to the sense that when a case is not substantiated, the reported perpetrator avoids a black mark against him or her.

Finally, substantiation is used in calculating state and federal statistics on the incidence of child maltreatment (U.S. Department of Health and Human Services, Administration for Children and Families, 2007). Both policy analysts and researchers using administrative databases have often treated substantiation as tantamount to a measurement of maltreatment itself, assuming that cases with substantiation represent children who have been maltreated and those without substantiation repre- sent children who have not been maltreated (Besharov, 1988; see Drake, 1996).

Basis for the Substantiation Decision

According to DePanfilis and Salus’s (2003) Child Protective Services: A Guide for Caseworkers, a manual from the Federal Office of Child Abuse and Neglect, and Drake’s (2000) chapter in the Handbook for Child Protection Practice, the substantiation decision is based on the answers to two questions:

Is the harm to the child severe enough to constitute child maltreatment?

Is there sufficient evidence to support this being a case of child maltreatment?

Drake (1996) presented a theoretical model of sub- stantiation encompassing these questions. In his model, critical thresholds of both harm and evidence must be met for a case to be substantiated. In some cases, risk of future harm can take the place of current harm in this model (as discussed later).

These authors’ model of substantiation makes sense, but we are aware of no studies that look comprehen- sively at whether caseworkers use harm, risk, and evi- dence in making substantiation decisions. Moreover, research discussed later suggests that other variables may influence the substantiation decision as well, and it is unclear to what extent these variables are related to substantiation because they affect judgments of harm, risk, and evidence or because they have an effect on substantiation that is independent of harm, risk, and evidence.

The Process Leading to Substantiation

The investigation process begins with a report of sus- pected child maltreatment by a professional or layperson that is screened in for investigation if, on an initial tele- phone screen, the allegation seems credible. State statutes direct CWS to conduct an investigation with standardized components (e.g., child interview when the youngster’s developmental level permits, interview with caregivers, home visit, interview with alleged perpetra- tors) and complete it within a specified period, ranging from 10 to 90 days depending on the state (Walter R. McDonald & Associates, 2003).

The investigation typically includes an assessment of children’s current safety, their future risk of harm, and the evidence available for the case. Many states incor- porate structured safety and risk assessment protocols in their investigation procedures (Walter R. McDonald & Associates, 2003). The investigating caseworker, typ- ically in consultation with their supervisor, then decides whether to substantiate the abuse (DePanfilis & Salus, 2003). Some states have developed so-called dual track or differential response systems, in which only cases involving greater risk receive investigations and substantiation decisions, whereas cases involving lower risk receive family assessments without a sub- stantiation determination (Schene, Oppenheimer, & Senderling, 2005).

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Level of Harm and Substantiation

Common sense suggests that substantiation is more likely given more severe levels of harm, but this rela- tionship has not always been found. Some studies report that severity of harm is a predictor of substantiation (Trocmé, Tam, & McPhee, 1995; Winefield & Bradley, 1992; Zuravin et al., 1995), but the only study in the United States based on a nationally representative sam- ple (the National Incidence Study; N = 7,623) found no significant association between severity of harm and substantiation (King, Trocmé, & Thatte, 2003).

Level of Risk and Substantiation

Risk has a complicated and somewhat ambiguous relationship to substantiation. CWS investigations must not only determine whether maltreatment occurred but also assess the risk of future maltreatment. This helps guide decision making and planning of CWS interven- tions such as child placement and family services. Many states have formalized the evaluation of risk by develop- ing risk assessment protocols and risk management sys- tems (Gambrill & Shlonsky, 2000; Ryan, Wiles, Cash, & Siebert, 2005; Wald & Woolverton, 1990). In conducting investigations, caseworkers typically collect data inform- ing both risk assessment and the substantiation decision. English et al. (2002, p. 819) stated that “the relationship between risk assessment and substantiation is far from clear in actual CPS practice” and that “the research evi- dence suggests some mixing of these concepts in prac- tice.” Thus, information from the former may influence the latter. Most states use formal risk assessment models after the substantiation decision has been made (English et al., 2002), although caseworkers, of course, encounter information relevant to risk assessment from their first contact with a family.

Despite their reported relationship in practice, some experts warn against using risk assessment to help make substantiation decisions, because the former concerns future behavior and the latter concerns current behavior (Pecora, 1991; Wald & Woolverton, 1990). In almost all states, caregivers’ actions must meet the legal definition of abuse and neglect for a case to be substantiated, and substantiation cannot be based on a prediction of future risk to the child.

The caregiver’s risky behavior can be considered abuse or neglect, even if children have not been harmed, if the behavior has “created conditions that have a high likelihood of harming the child” (Wald & Woolverton, 1990, p. 497), such that “it was just a matter of fortuity that harm did not occur” (p. 509). Thus, for example, a case meets state requirements for child abuse and neglect

if children are in physical danger because of lack of supervision or living in an environment in which they are likely to be injured. The 2003 National Study of Child Protective Services Systems and Reform Efforts (Walter R. McDonald & Associates, 2003) found that seven states explicitly stated in their policies that substantiation could be based on either current evidence of maltreat- ment or the risk of maltreatment happening in the future. Thus, one would expect that the risk should be a predic- tor of substantiation, because of risky caregiver behavior that places children in danger or because of the potential influence of risk assessment on substantiation.

Level of Evidence and Substantiation

Tangible evidence to support substantiation comes from documentable facts based on child disclosure, per- petrator’s admission, eyewitness, testimony, medical evi- dence, and physical evidence (English et al., 2002). Each state’s legislature sets the standard of proof for deter- mining when sufficient evidence has been reached. As a consequence, standards differ, with some states requiring some credible evidence and other states using the stricter criterion of preponderance of evidence. Because of the need for child protection, the beyond a reasonable doubt standard used for criminal convictions is considered too rigorous for substantiation. The National Child Abuse and Neglect Data System has shown remarkable varia- tion in substantiation rates across states for many years (U.S. Department of Health and Human Services, Administration for Children and Families, 2007), and a plausible hypothesis is that at least some of this variation is due to state differences in evidentiary standards on substantiation. However, one study using 1994 national data found no statistically significant difference in the percentage of substantiated cases between states that used some credible evidence (a lower standard; 34.5% substantiated) and those states that used preponderance of the evidence (a higher standard; 30.9% substantiated), concluding that “the verbal formula hardly affects the investigator’s decision to substantiate a case” (Levine, 1998, p.345).

Whatever the evidence standard is, empirical research suggests that substantiation does indeed depend on evi- dence of maltreatment, even when caseworkers judge that there is harm or risk. One study of substantiation in three states found that caseworkers recorded harm or risk of harm in 46% cases and suspected maltreatment in 58% of cases but only substantiated 27% (Giovannoni, 1989). Other studies have found that sizable proportions of cases in some communities or states are classified as suspected or possible abuse but lack evidence (Jason,

40 Child Maltreatment

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Andereck, Marks, & Tyler, 1982; Zuravin et al., 1995). However, none of these studies included a quantitative measure of sufficiency of evidence.

Other Variables Related to Substantiation

Rates of substantiation vary because of the specific circumstances of the investigation, including referral source, county or region, number of decision categories, size of the CWS office, agency resource deficits, case- load, and quality of supervision (see, e.g., Child Welfare Information Gateway, 2003; English et al., 2002). Moreover, studies have also found that child and family characteristics such as child age, race/ethnicity, family income, living in urban areas, and prior reports or inves- tigation with CWS have been associated with substanti- ation, even when harm was statistically controlled, as several studies have done (Eckenrode, Powers, Doris, Munsch, & Bolger, 1988; English et al., 2002; King et al., 2003; Trocmé, Knoke, Fallon, & MacLaurin, 2006; Zuravin et al., 1995). The substantiation decision may even depend on the family’s response during the investi- gation: In a 1985 survey, workers reported that they were less likely to substantiate cases when caregivers admitted maltreatment (Alter, 1985).

In addition, characteristics of the investigators and the systems they work in relate to substantiation (Child Welfare Information Gateway, 2003). Caseworker experience, self-reported skills, supportive relation- ships with co-workers, and stated adherence to state policy were all related to substantiation, as were reports by CWS supervisors of the cohesiveness of their work units and their supportiveness to workers. DePanfilis and Girvin (2005) suggested that substantiation can depend on caseworkers’ facility with information pro- cessing. In their review of investigations of allegations in out-of-home care, the authors found that caseworkers sometimes failed to consider the complexity of case situations or to examine case facts in reference to defi- nitions of child abuse and neglect.

Whether appropriate services are available for a family is likely to influence the substantiation decision in many agencies. DePanfilis and Girvin (2005) and English et al. (2002) reported this relationship and cited a number of studies in reference to it, although we did not find that any of the studies empirically tested this relationship using case-based data. Nevertheless, in a national survey of a random sample of 383 local CWS agencies, 50% of agencies reported that they sometimes or always considered availability of services in making the substantiation decision (Walter R. McDonald Associates, 2003).

All this variation due to apparently extraneous factors raises questions about the degree to which substantiation follows from the degree of harm or risk and evidence. It is not clear to what extent these other factors influence substantiation because they affect judgments of risk, harm, or evidence or because they have an independent effect unrelated to these caseworker judgments.

Some studies have found that different variables predict substantiation depending on the type of maltreat- ment suffered by the child (Drake, 1995; Eckenrode et al., 1988; Zuravin et al., 1995). Thus, Zuravin et al. (1995) found that physical abuse cases with older children and those with African American children were more likely to be substantiated, and Eckenrode et al. (1988) found that physical abuse cases with Hispanic children were more likely to be substantiated. Eckenrode and colleagues (1988) also found that sexual abuse cases with older children were more likely to be substantiated, whereas neglect cases with younger children were more likely to be substantiated.

The Present Analysis

Despite the importance of substantiation, the need to understand what factors caseworkers consider, and the existence of Drake’s theoretical model, the relationship of harm, risk, and sufficiency of evidence to substantia- tion has never been fully examined empirically; nor do we understand whether other variables relate to substan- tiation because they affect levels of harm, risk, and evi- dence or have an effect independent of these caseworker judgments. This study measures caseworkers’ judgments of harm, risk, and sufficiency of evidence in a large national probability study of child welfare and examines each as a predictor of substantiation along with other predictors drawn from previous research. One goal is to understand better the decision-making processes under- lying substantiation for child maltreatment in general and for specific types of maltreatment. A related objec- tive is testing Drake’s (1996) model of substantiation. A second goal is to understand better the situation of children in both substantiated and unsubstantiated cases. A third goal is to contribute to the knowledge base needed to develop computerized prediction tools for investigation. The research questions are as follows:

• How do substantiation rates vary on the basis of caseworker judgments of harm, risk, and evidence?

• How do substantiation rates vary on the basis of demo- graphic and maltreatment characteristics, and are these effects independent of judgments of harm, risk, and evidence?

Cross, Casanueva / Caseworker Judgments and Substantiation 41

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

42 Child Maltreatment

• Do predictors of substantiation differ by type of maltreatment?

• What proportion of children involved in child protective investigation experience different combinations of harm (or risk), evidence, and substantiation (e.g., substantial harm with little evidence and no substantiation)?

Method

National Survey of Child and Adolescent Well-Being (NSCAW)

This national probability study provides ample data for studying substantiation. The NSCAW is a nationally representative study of children and youths who were reported to child welfare services in 1999 and 2000 (NSCAW Research Group, 2002; U.S. Department of Health and Human Services, Administration for Children and Families, 2005). The NSCAW used two-stage random sampling. It randomly sampled 92 primary sampling units (mostly counties) across the country and 5,501 children and youths within those units, permitting accu- rate national estimates on child maltreatment investiga- tions. Because of statistical power requirements for key categories of cases, the sample design required over- sampling of cases that were open in CWS, cases with infants, and cases with alleged sexual abuse cases. Thus, all NSCAW data analysis needs to apply weights for the results to represent accurately the national CWS popula- tion, and all percentages reported here represent weighted percentages. Data collection about the index child included child assessments and instruments com- pleted by children and families, as well as caseworkers and teachers. At baseline, all participants related to a given child were interviewed during the same window period. Children and case information were followed longitudinally with repeated measurements over time, although this report used baseline data.

The analysis in this study concerns the 4,514 children who were involved in investigations in which a substantiation decision was made at the time of the index investigation. Data on substantiation were missing for 14.7% (weighted) of the original sample of 5,501. Missing data on substantiation occurred because (a) a substantiation decision was not made; (b) the original caseworker was interviewed but could neither remem- ber the substantiation decision nor locate it in the case record; or (c) another caseworker was interviewed, did not know the substantiation decision, and could not locate it in the case record (procedures for caseworker interviews are discussed later).

A study of substantiation using the NSCAW has sev- eral advantages. It (a) provides national estimates, (b) includes data from both caregiver interviews and child assessments and interviews, and (c) provides more infor- mation from caseworkers on decision-making factors than most other studies.

Procedure

Field representatives contacted caregivers and asked permission to interview them about selected child, family, and caregiver characteristics. Baseline interviews of caregivers and caseworkers and child assessments were all conducted an average of 4 months after the CWS investigation for maltreatment. Interviews with the children’s caregivers were conducted in English (96%) or Spanish at the children’s homes, by means of com- puter-assisted personal interviewing. Caregivers received an honorarium of approximately $40 for their participa- tion in each interview. The NSCAW also conducted 1-hr computer-assisted personal interviewing sessions with the CWS caseworkers, who were instructed to consult the case record as needed during the interview. At base- line, the investigating caseworker was interviewed. When the investigating caseworker was not available, an ongoing caseworker completed the interview using the case record to provide information from the investiga- tion. The data analyzed here are drawn from caregiver and caseworker interviews.

Variables

Caseworkers used a 4-point scale to rate the degree of harm the child suffered and the degree of risk the child faced. They also used a 5-point scale to rate the suffi- ciency of evidence for substantiation. The text of the prompts and questions was as follows:

For the next set of questions, please do not be con- cerned with whether or not the report was substan- tiated when offering your responses. Regardless of the outcome of the investigation, how would you describe the level of harm to [CHILD’S NAME]? Would you say:

1 = None 2 = Mild 3 = Moderate 4 = Severe

Regardless of the outcome of the investigation, how would you describe the level of severity of risk? Would you say:

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

1 = None 2 = Mild 3 = Moderate 4 = Severe

Regardless of the outcome of the investigation, how sufficient was the evidence to substantiate the case? Would you say:

1 = There was no evidence of maltreatment 2 = Evidence was clearly not sufficient 3 = Evidence was probably not sufficient 4 = Evidence was probably sufficient 5 = Evidence was clearly sufficient

Caseworkers also completed a 29-item risk assess- ment instrument (U.S. Department of Health and Human Services, Administration for Children and Families, 2005). A total score on the risk assessment instrument was computed and correlated with the rating from the overall risk item listed earlier. Given the fairly high correlation between the risk assessment instru- ments and the overall risk rating (r = .55), as well as previous research reporting higher predictive value of an overall risk rating (English et al., 2002), results are only presented for the latter.

Caseworkers judged that 44.1% of children suffered no harm, 28.3% suffered mild harm, 20.0% suffered moderate harm, and 7.6% suffered severe harm. Caseworkers also judged that 31.1% of children were not at risk, 36.8% were at mild risk, 22.2% were at mod- erate risk, and 9.9% were at severe risk. Regarding suf- ficiency of evidence, 34.3% of cases were rated as having no evidence of maltreatment; 16.8%, as having clearly insufficient evidence; 11.4%, as having probably insufficient evidence; 8.8%, as probably sufficient evi- dence; and 28.7%, as clearly sufficient evidence. Not surprisingly, ratings of harm, risk, and evidence were significantly correlated. The Pearson r for harm and risk was .73; for harm and evidence, it was .61; and for risk and evidence, it was .64. These correlations, although high, still meant that harm, risk, and evidence ratings were dissimilar on many cases.

Substantiation. Children were classified as having a substantiated maltreatment case on the basis of case- workers’ response at baseline to the question, “Was the outcome of the investigation (a) substantiated, (b) indicated, or (c) neither substantiated or indicated?” In so-called three-tier states, indicated is a classification option that means some evidence of maltreatment exists but not enough for substantiation.(DePanfilis & Salus, 2003). Only when caseworkers chose substantiated were

children classified as such. All other cases, including indicated cases, were classified as unsubstantiated. Developing models predicting substantiated, unsubstan- tiated, and indicated in three-tier states was beyond our scope because of our focus on the national population of CWS investigations.

Maltreatment characteristics. The most serious type of maltreatment was assessed by asking the caseworkers the following question in the baseline interview: “Of the types of abuse or neglect that were reported, please tell me the type that you felt was the most serious (physical maltreatment, sexual maltreatment, emotional maltreat- ment, physical neglect [failure to provide], neglect [lack of supervision], abandonment, moral/legal maltreatment, educational maltreatment, exploitation, and other).” Because of the small number of children in several cate- gories of maltreatment, children who were reported for emotional maltreatment (6.9%), abandonment (1.4%), moral/legal maltreatment (0.5%), educational maltreat- ment (1.5%), exploitation (0.1%), and other (3.8%), were classified as “all others.”

Prior reports to CWS. Prior reports were assessed by asking the caseworkers the following question in the baseline interview: “Was there any prior reports of mal- treatment to the agency?”

Demographics. Caregivers were asked about their child’s gender, age, race/ethnicity, and about the family’s income. Information about urbanicity was determined on the basis of the classification of the agency (urban or rural) where the report of maltreatment was received.

Analyses

All analyses were conducted with weighted data, using the SUDAAN statistical package, version 9.0.1(RTI International, 2002), to take into account the NSCAW’s complex sampling design. Thus, all percent- ages are adjusted (weighted) for sampling probabilities; sample sizes listed in this article, however, are unweighted. Following descriptive analyses, contin- gency tables with Pearson χ2 tests adapted for complex samples were computed to examine the bivariate rela- tionship of harm, risk, evidence, demographic, and mal- treatment variables to substantiation. The demographic and maltreatment variables were chosen on the basis of previous research (Eckenrode et al., 1988; English et al., 2002; King et al., 2003; Trocmé et al., 1995; Winefield & Bradley, 1992); Pearson correlations with appropriate weights were used to describe the association between harm, risk, and sufficiency of evidence.

Cross, Casanueva / Caseworker Judgments and Substantiation 43

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

44 Child Maltreatment

Binomial logistic regression analysis was used to model substantiation and several models were employed. The first model examined the same demographic and maltreatment variables used in the bivariate analyses. Subsequently, substantiation was analyzed as a function of both demographic and maltreatment variables and each of the main predictor variables (harm, risk, and sufficiency of evidence). The final model examined sub- stantiation as a function simultaneously of the demo- graphic and maltreatment variables and of the three main predictors. This last model was then repeated for four separate groups of children depending on the primary type of abuse that was reported (physical abuse, sexual abuse, physical neglect, and neglect). All two-way inter- actions among the three main predictors were tested, but none was statistically significant; thus, those results are not presented. The maximum variance inflation factor for the variables in the logistic regression analyses was 2.2, well within acceptable limits, indicating that multi- collinearity was not a problem in these analyses.

Population Characteristics

Table 1 shows estimated population characteristics with standard errors. Girls and boys were about evenly represented. The most frequent age group was children aged 6–10, although every age group was well repre- sented. Children who were Hispanic and African American were overrepresented relative to the overall population, as is true in most studies of the child welfare population. Nearly half of children were living in families with very low incomes (<$15,000 a year), and over three quarters were living in urban communities. As many child maltreatment studies have found, neglect (failure to provide and lack of supervision) was the most common form of reported maltreatment, followed by physical abuse.

Results

Substantiation Rate

The overall substantiation rate was 29.7% (SE = 1.9), similar to the estimate based on the National Child Abuse and Neglect Data System (NCANDS) at a com- parable time period (31.2%, calculated as number sub- stantiated / [number substantiated + number indicated + number unsubstantiated]; U.S. Department of Health and Human Services, Administration for Children and Families, 2002).

Bivariate Predictors of Substantiation

Harm, risk, and evidence had a linear relationship to substantiation in bivariate analyses (see Table 2). At lower levels of harm, risk and evidence, only a few alle- gations were substantiated. At moderate or higher levels of harm, risk, and evidence, a higher percentage of cases were substantiated. The difference in substantiation rates was particularly striking between cases with probably insufficient evidence, which had a substantiation rate of 7.8%, and cases with probably sufficient evidence, which had a substantiation rate of 56.6%.

Three demographic characteristics were also signifi- cantly related to substantiation. Allegations involving girls were significantly more likely to be substantiated compared with allegations involving boys. Substantiation was significantly related to child’s age group but not in a linear way. The substantiation rate was over 30% for children aged 0–2 and 6–10 but under 30% for children and youths aged 3–5 and 11–14. Substantiation was sig- nificantly more likely among families with incomes ≥30,000, compared with lower income families. Substantiation rates did not vary significantly by child

Table 1 Sample Composition and Estimated Population

Characteristics (N == 4,514)

Weighted Demographic n percentage SE

Gender Male 2,225 49.3 2.0 Female 2,289 50.7 2.0

Age 0–2 years 1,630 18.3 1.0 3–5 years 686 20.3 1.5 6–10 years 1,235 36.7 1.8 11+ years 962 24.8 1.3

Race Black 1,404 27.9 2.9 White 1,954 46.0 3.8 Hispanic 808 19.1 2.7 Other 342 7.0 0.9

Income <$15,000 1,530 43.7 2.0 $15,000–$29,999 1,230 29.3 1.5 $ ≥30,000 1,346 27.0 1.8

Urbanicity Nonurban 988 24.0 5.9 Urban 3,526 76.0 5.9

Most serious type of abuse Physical abuse 1,023 26.5 1.4 Sexual abuse 537 11.2 1.4 Physical neglect 1,018 19.7 1.5 Supervisory neglect 1,164 26.3 1.9 All others 772 16.3 1.6

Prior reports of maltreatment Yes 2,585 50.8 1.5 No 2,412 49.2 1.5

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Cross, Casanueva / Caseworker Judgments and Substantiation 45

race/ethnicity or urbanicity of the CWS agency. Substantiation was slightly but significantly more likely when there were prior reports of maltreatment, but the most serious type of abuse in a case was not significantly related to substantiation.

Multivariable Model Predicting Substantiation

Table 3 shows the logistic regression analysis assessing the relative effect of demographic variables; maltreatment characteristics; and harm, risk, and evi- dence on the substantiation decision. In all models in which they were included, all three caseworker judgment variables were significant predictors of substantiation at the .001 level. The relative size of the odds ratios, how- ever, was striking. In the final model, every 1-point dif- ference on the evidence scale was associated with 3.1 greater odds of substantiation. For the harm scale, every 1-point difference was associated with 1.6 greater odds of substantiation, about half the size of the odds ratio for the evidence scale. For the risk scale, every 1-point dif- ference was associated with 1.9 greater odds of substan- tiation, a little less than two thirds the size of the evidence odds ratio. This larger odds ratio for evidence is even more striking when we consider that the evidence scale had 5 points and the harm and risk scales had 4 points; a priori, we would expect the odds ratios to be smaller for a scale with more points, because we would expect the increments between points to be smaller. Although some of this may be explained by the specific wording of the questions (as discussed later in Limitations), these results nevertheless suggest that evi- dence has the most important effect on substantiation.

All five models were also analyzed in terms of their ability to predict correctly which cases were substanti- ated. The model with only demographic variables cor- rectly predicted only 6.5% of the cases that were substantiated. In contrast, models with harm and risk (Models 2 and 3) correctly predicted around 60% of sub- stantiated cases, whereas models with evidence and with all three main predictors (Models 4 and 5) correctly pre- dicted substantiation in more than 80% of the cases.

Gender and age were also associated with substantia- tion, independent of harm, risk, and evidence. Boys were less likely to be substantiated in all models, and reports with children aged 6–10 were more likely to be substantiated than cases with children aged 0–2 in three of the models. Income and prior report were significant in the initial model but no longer significant when caseworker judgment variables were entered into the logistic regression equation.

Table 2 Substantiation Status by Caseworker

Judgments, Demographics, and Maltreatment Characteristics

Substantiated weighted

Variable percentage SE χ2 p

Overall caseworker judgments 29.7 1.9 — — Harm 82.80 <.0001 None 7.4 1.2 Mild 25.4 3.2 Moderate 67.5 3.7 Severe 74.9 5.0

Risk 102.43 <.0001 None 5.0 1.2 Mild 19.5 2.5 Moderate 58.8 4.1 Severe 79.7 3.6

Evidence 133.40 <.0001 None 1.6 0.6 Clearly not sufficient 7.5 3.5 Probably not sufficient 7.8 3.3 Probably sufficient 56.6 5.4 Clearly sufficient 76.9 3.2

Demographic characteristics Gender 9.74 .003 Male 25.3 2.1 Female 33.9 2.5

Age 15.46 .003 0–2 years 31.8 3.0 3–5 years 20.8 2.7 6–10 years 35.1 3.5 11+ years 27.2 3.0

Race/ethnicity 2.54 .47 Black 26.4 3.0 White 31.8 2.6 Hispanic 30.7 3.4 Other 25.6 4.5

Income 10.03 .009 <$15,000 $15,000–$29,999 ≥$30,000

PSU urbanicity 0.24 .624 Urban 29.1 2.1 Nonurban 31.6 4.6

Maltreatment characteristics Most serious type 7.68 .115

of maltreatment Physical abuse 26.8 2.8 Sexual abuse 29.4 4.8 Physical neglect 24.4 2.8 Neglect 30.9 3.2 All others 38.8 5.0

Prior reports of maltreatment 4.77 .031 Yes 32.3 2.3 No 26.9 2.5

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

46 Child Maltreatment

Logistic regressions by primary type of abuse (see Table 4) showed that harm, risk, and evidence were significantly associated with substantiation for physical abuse and neglect. For sexual abuse, only harm and evi- dence were associated with substantiation, whereas only evidence was associated with substantiation for

physical neglect. Being a boy was associated with a lower likelihood of substantiation among children reported for physical abuse and neglect. Children aged 3–5 were less likely than children aged 0–2 to have substantiated cases for sexual abuse, whereas children aged 6–10 were more likely to have substantiated cases

Table 3 Logistic Regression Predicting Substantiation

Odds ratios (and 95% confidence intervals) for:

Model 2: Model 3: Model 4: Model 5: Model 1: Harm and Risk and control Evidence and Harm, risk, evidence,

Control variablesa control variablesb variablesc control variablesd and control variablese

Variable (n = 4,029) (n = 4,004) (n = 4,005) (n = 3,995) (n = 3,977)

Harm 4.39 (3.48–5.56)*** 1.60 (1.20–2.14)*** Risk 5.21 (4.03–6.75) 1.92 (1.51–2.43)*** Evidence 3.94 (3.11–5.00) 3.10 (2.43–3.95)*** Gender ** ** ** ** ***

Male 0.65 (0.49–0.87) 0.68 (0.51–0.90) 0.53 (0.39–0.71) 0.59 (0.41–0.85) 0.53 (0.38–0.74) Female 1.0 1.0 1.0 1.0 1.0

Age ** * * * 0–2 years 1.0 1.0 1.0 1.0 1.0 3–5 years 0.53 (0.35–0.83) 0.80 (0.49–1.31) 0.89 (0.54–1.46) 0.64 (0.35–1.11) 0.89 (0.52–1.54) 6–10 years 1.08 (0.73–1.59) 1.46 (1.00–2.14) 2.18 (1.41–3.36) 1.50 (0.91–2.47) 2.15 (1.27–3.63) 11+ years 0.66 (0.47–0.94) 0.82 (0.57–1.19) 1.21 (0.83–1.76) 0.97 (0.64–1.45) 1.24 (0.81–1.88)

Race/ethnicity White 1.0 1.0 1.0 1.0 1.0 Black 0.84 (0.58–1.22) 0.78 (0.46–1.32) 0.80 (0.50–1.28) 1.03 (0.61–1.73) 0.90 (0.51–1.61) Hispanic 1.04 (0.76–1.42) 1.00 (0.66–1.50) 1.05 (0.70–1.58) 1.29 (0.56–2.96) 1.32 (0.66–2.66) Other 0.73 (0.42–1.26) 0.86 (0.51–1.46) 0.61 (0.33–1.12) 1.21 (0.73–2.00) 1.07 (0.64–1.80)

Income * <15,000 1.0 1.0 1.0 1.0 1.0 $15,000–$29,999 1.10 (0.77–1.57) 1.12 (0.76–1.64) 1.23 (0.81–1.85) 1.06 (0.59–1.91) 1.16 (0.68–1.98) ≥30,000 or more 1.54 (1.12–2.12) 1.33 (0.86–2.05) 1.24 (0.85–1.80) 1.55 (0.96–2.50) 1.33 (0.84–2.09)

PSU urbanicity Urban 1.0 1.0 1.0 1.0 1.0 Nonurban 1.19 (0.71–1.99) 1.13 (0.57–2.21) 1.05 (0.54–2.05) 0.90 (0.49–1.64) 0.87 (0.44–1.70)

Most serious type of abuse Neglect 1.0 1.0 1.0 1.0 1.0 Physical abuse 0.80 (0.54–1.18) 0.71 (0.48–1.09) 0.93 (0.62–1.40) 1.23 (0.66–2.29) 0.94 (0.59–1.48) Sexual abuse 0.96 (0.63–1.48) 0.48 (0.24–0.95) 0.93 (0.50–1.75) 1.23 (0.66–2.29) 0.93 (0.46–1.89) Physical neglect 0.73 (0.52–1.04) 0.79 (0.55–1.18) 0.92 (0.63–1.37) 0.88 (0.57–1.35) 0.97 (0.63–1.49) All others 1.49 (0.89–2.47) 1.37 (0.80–2.34) 1.91 (1.03–3.55) 1.99 (1.03–3.85) 1.96 (1.05–3.66)

Prior reports of maltreatment ** Yes 1.42 (1.10-1.84) 1.05 (0.80-1.37) 0.81 (0.61-1.07) 1.09 (0.78-1.53) 0.84 (0.61-1.17) No 1.0 1.0 1.0 1.0 1.0

Note: Statistically significant results are printed in bold. aCox & Snell pseudo R2 = .05; Wald F = 4.67, p < .0001. Predictive accuracy of substantiated cases, 6.5%; predictive accuracy of unsubstanti- ated cases, 96.3%. bCox & Snell pseudo R2 = .30; Wald F = 13.44, p < .0001. Predictive accuracy of substantiated cases, 59.7%; predictive accuracy of unsub- stantiated cases, 78.1%. cCox & Snell pseudo R2 = .31; Wald F = 13.22, p < .0001. Predictive accuracy of substantiated cases, 61.3%; predictive accuracy of unsubstan- tiated cases, 77.0%. dCox & Snell pseudo R2 = .44; Wald F = 17.99; p < .0001. Predictive accuracy of substantiated cases, 85.5%; predictive accuracy of unsub- stantiated cases, 73.0%. eCox & Snell pseudo R2 = .47; Wald F = 12.37; p < .0001. Predictive accuracy of substantiated cases, 83.9%; predictive accuracy of unsub- stantiated cases, 72.5%. * p < .05. ** p < .01. *** p < .001.

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Cross, Casanueva / Caseworker Judgments and Substantiation 47

for physical neglect than children aged 0–2. Sexual abuse cases with Hispanic children were less likely to be substantiated than sexual abuse cases with White children. Sexual abuse cases in which family income was <$15,000 were less likely to be substantiated than cases with families earning >$15,000. All mod- els predicted correctly around 80% of substantiated cases.

Figure 1 provides a more intuitive picture of the importance of evidence across all cases. The percentage substantiated increased as harm increased from none to severe, but very much depended on the level of evi- dence. When evidence was judged to be probably insuf- ficient or less, substantiation rates were small, even when harm was rated as severe. A parallel figure depict- ing risk instead of harm showed similar results.

Table 4 Logistic Regression Predicting Substantiation by Type of Main Maltreatment

Odds ratios (and 95% confidence intervals) for:

Model 4: Model 4: Model 1: Model 2: Physical neglect Neglect (failure

Physical abuse Sexual abuse (failure to provide; to supervise; Variable (n = 896)a (n = 597)b n = 918)c n = 1,032)d

Harm * ** * 1.80 (1.00–3.24) 2.36 (1.37–4.07) 1.06 (0.68–1.65) 1.64 (1.03–2.62)

Risk ** ** 2.54 (1.44–4.47) 1.75 (0.85–3.60) 1.74 (0.93–3.25) 1.87 (1.24–2.81)

Evidence *** *** *** *** 3.23 (2.30–4.53) 5.02 (3.04–8.27) 4.38 (2.70–7.13) 3.52 (2.43–5.12)

Gender * ** Male 0.46 (0.22–0.98) 0.59 (0.22–1.63) 0.59 (0.26–1.31) 0.43 (0.25–0.73) Female 1.0 1.0 1.0 1.0

Age ** * 0–2 years 1.0 1.0 1.0 1.0 3–5 years 2.69 (0.82–8.86) 0.08 (0.02–0.36) 0.78 (0.24–2.58) 0.82 (0.32–2.10) 6–10 years 4.02 (1.18–13.69) 0.98 (0.28–3.41) 3.17 (1.21–8.29) 0.92 (0.45–1.91) 11+ years 2.45 (0.74–8.07) 1.12 (0.41–3.08) 1.11 (0.37–3.37) 0.62 (0.23–1.66)

Race/ethnicity (ref. White) ** White 1.0 1.0 1.0 1.0 Black 0.49 (0.21-1.12) 0.51 (0.15-1.75) 2.39 (0.87-6.57) 0.72 (0.32-1.60) Hispanic 2.07 (0.71-6.07) 0.21 (0.05-0.83) 1.20 (0.44-3.29) 0.63 (0.26-1.53) Other 1.02 (0.46-2.29) 4.17 (0.94-18.50) 2.02 (0.49-8.28) 0.62 (0.26-1.47)

Income ** <$15,000 1.0 1.0 1.0 1.0 $15,000–$29,999 0.64 (0.26-1.58) 0.14 (0.03-0.62) 1.32 (0.59-2.94) 2.33 (1.09-4.96) $30,000 or greater 0.66 (0.26-1.58) 0.14 (0.03-0.58) 1.82 (0.46-7.19) 1.75 (0.81-3.79)

PSU urbanicity Urban 1.0 1.0 1.0 1.0 Nonurban 0.41 (0.16-1.02) 0.63 (0.13-3.12) 1.25 (0.51-3.06) 0.87 (0.37-2.09)

Prior reports of maltreatment Yes 0.87 (0.45-1.71) 0.77 (0.21-2.85) 1.13 (0.50-2.54) 0.90 (0.48-1.69) No 1.0 1.0 1.0 1.0

Note: Statistically significant results are printed in bold. aCox & Snell pseudo R2 = .49; Wald F = 8.86, p < .0001. Predictive accuracy of substantiated cases, 77.5%; predictive accuracy of unsubstantiated cases, 75.8%. bCox & Snell pseudo R2 = .57; Wald F = 6.07, p < .0001. Predictive accuracy of substantiated cases, 81.9%; predictive accuracy of unsubstantiated cases, 79.4%. cCox & Snell pseudo R2 = .47; Wald F = 11.20, p < .0001. Predictive accuracy of substantiated cases, 87.2%; predictive accuracy of unsubstantiated cases, 73.5%. dCox & Snell pseudo R2 = .47; Wald F = 10.51, p < .0001. Predictive accuracy of substantiated cases, 82.0%; predictive accuracy of unsubstantiated cases, 74.9%. * p < .05. ** p < .01. *** p < .001.

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

48 Child Maltreatment

Distribution of Children by Harm, Evidence, and Substantiation

To summarize the results on harm, evidence, and sub- stantiation, we imagined a hypothetical group of 100 children and portrayed them in Figure 2 where they stood at the end of the investigation in terms of these variables. In effect, Figure 2 portrays a picture of Drake’s harm–evidence model as it applies to actual children. To parallel the harm–evidence model, we left risk out of this analysis, but our analysis of risk instead of harm yielded similar results. Using the number 100, of course, means that the numbers of children in this section can be thought of as percentages of children as well. In 56 cases (Figure 2, lower left), children were judged to have suf- fered little or no harm and the caseworker found little or no evidence of maltreatment; 55 of the 56 were not sub- stantiated. In 22 cases, there was probably to clearly suf- ficient evidence and children were judged to have suffered moderate to severe harm (Figure 2, upper right). Of these cases, 18 were substantiated, although 4 of them were not, despite high levels of evidence and moderate to severe harm.

Those cases in which judged harm and evidence were at different levels deserve special scrutiny. In 16 cases, evidence of maltreatment was probably to clearly suffi- cient, but there was little or no harm (Figure 2, upper left). Nine of the 16 were substantiated. In 7 cases, the caseworkers judged that there was moderate to severe

harm, but evidence was probably to clearly insufficient (Figure 2, lower right), and only 2 of these cases were substantiated. Altogether, 9 of the 29 cases in which children were judged to have suffered moderate to severe harm were not substantiated.

Discussion

The first research question asked whether substantia- tion rates varied on the basis of caseworker judgments of harm, risk, and evidence. Harm, risk, and evidence were all substantial predictors of substantiation, even when demographic and maltreatment characteristics were taken into account. Clearly, caseworkers consider harm to the child, future risk to the child, and evidence of mal- treatment when they make substantiation decisions. This does not mean, however, that harm, risk, and evidence predict the substantiation decision for every child, as we discuss here. Note that the significance of harm contrasts with a previous report from the 1993 Third National Incidence Study, in which severity of harm was not sig- nificantly associated with substantiation (King et al., 2003). Evidence appeared to play the largest role. Even when caseworkers believe that children have been harmed or are at risk, substantiation is unlikely unless evidence of maltreatment is sufficient.

The second research question asked how substantiation rates varied on the basis of demographic and maltreatment

Figure 1 Percent Substantiated by Levels of Harm and Evidence

0

10

20

30

40

50

60

70

80

90

100

None Mild Moderate Severe

Caseworker ratings of harm

% s

u b

st an

ti at

ed

Clearly sufficient evidence Probably sufficient evidence Probably insufficient to no evidence

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Cross, Casanueva / Caseworker Judgments and Substantiation 49

characteristics and whether these effects were indepen- dent of judgments of harm, risk, and evidence. Income and prior report were initially significant but not so when caseworker judgment variables were included. This sug- gests that these two variables have an indirect effect on substantiation through their effect on appraisals of risk, harm, and evidence. On the other hand, child gender and age were each significantly related to substantiation beyond harm, risk, and evidence. Boys’ cases were less likely to lead to substantiation than girls. Cases with children aged 6–10 were more likely to be substantiated than cases of younger or older children. Because of the NSCAW’s weighting scheme, these effects cannot be due to state or agency differences between cases.

Because caseworkers make the substantiation deci- sion, the independent gender and age effects suggest that some other judgment must be influencing the substanti- ation decision in addition to harm, risk, and evidence— something that differs by child gender and age. At a given level of harm or risk, caseworkers may see the sub- stantiation decision as more or less necessary depending on gender and age. They may also tend to treat evidence of maltreatment as more or less consequential for different

groups of youths. Caseworkers may also be more likely to see mitigating factors for children of different genders and ages that would lead them not to substantiate certain groups of cases. Investigation, caseworker, and system characteristics related to substantiation may also vary by child gender and age, and so may caseworkers’ facility with processing information about the case. Clearly, more research is needed that examines caseworkers’ thinking in more depth and considers a wider range of judgment constructs.

Other demographic and maltreatment variables were not significant predictors. Previous studies have reported an association between race/ethnicity and substantiation (King et al., 2003; Trocmé et al., 2006), but a race/ethnicity effect was only found for sexual abuse.

Regarding the third research question on specific types of maltreatment, harm, risk, and evidence were dif- ferentially important for predicting substantiation depending on the type of maltreatment. Although evi- dence was a significant predictor of substantiation for all types of abuse, harm was not significantly associated with physical neglect and risk was associated with nei- ther sexual abuse nor physical neglect. It makes sense

Figure 2 Distribution of 100 Cases on Harm, Evidence, and Substantiation

Note: Each figure represents 1 out of 100 children. Shaded figures = substantiated. Total = 101 because of rounding.

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

50 Child Maltreatment

that risk is not an important predictor of substantiation for sexual abuse, because few circumstances exist in which sexual abuse would be considered risky but not harmful. There may be no additional independent role for risk to play beyond harm and evidence. The lack of effect of harm and risk on substantiation of physical neglect may relate to lack of variation on harm and risk. Such lack of variation may be substantial among physical neglect cases, related in part to economic and social conditions. Substantiation might then depend on evidence that caregivers themselves are responsible for the physical neglect of the child. Eckenrode et al. (1998) found that substantiation of sexual abuse was more likely in older children, but the present study did not find this effect. We have no ready explanation for the finding that being Hispanic and having an income over $15,000 were associated with a lower likelihood of substantiation of sexual abuse, independent of the significant effects of harm and evidence. More research is needed on the effects of race/ethnicity and social class on substantia- tion of sexual abuse. The lack of effect of child age on substantiation of physical neglect differed from Eckenrode et al.’s (1988) finding that neglect cases with younger children were more likely to be substantiated. In contrast to previous studies on child physical abuse that reported effects of child race and age on substantiation (Eckenrode et al., 1988; Zuravin et al., 1995), only harm, risk, evidence, and gender were associated with substan- tiation of physical abuse in the present study. Two caveats of comparison with previous studies are that (a) variables in the predictive models differ and that (b) most previous studies do not incorporate harm, risk, and evi- dence in their models.

The fourth research question concerns the percentages of children in different combinations of harm, evidence, and substantiation. More than half (55%) were rated with little or no harm and no evidence of maltreatment and were not substantiated. The next largest percentage was the 18% of cases with moderate to severe harm, probably to clearly sufficient evidence, and a decision to substanti- ate. Thus, in 73% of cases, there was a straightforward match among harm, evidence, and substantiation.

What is most concerning is the group of children who were judged to be at moderate to severe risk and harm and did not have cases that were substantiated. We found 9 of 100 children in this category. This outcome in 5 of these 9 cases could be attributed to insufficient evidence, as Drake’s (1996) model predicted. In 4 of these 9, however, cases seemed to meet all the criteria for sub- stantiation, but caseworkers reported that they were unsubstantiated. Nine of 100 is not a trivial percentage

of children. Children who have been harmed and/or remain at risk without substantiation should be a group of high concern for CWS. These cases should receive an evaluation of need for child and family services to pro- tect their safety and promote their well-being and devel- opment. They also need regular monitoring. Because of groups like this, most states, but not all, make CWS ser- vices available in at least some cases that are not sub- stantiated (Walter R. McDonald & Associates, 2003), although children in substantiated cases are nevertheless more likely to receive CWS services (see U.S. Department of Health and Human Services, Administration for Children and Families, n.d.).

For a different 9 of 100 children, cases were substan- tiated, even though they were judged to have suffered little or no harm or risk, because there was probably to clearly sufficient evidence of abuse. States considering an alternative response or dual track system may want to consider this result in their decision making; these are children who would probably be served by the alterna- tive response pathway.

In many ways, the results are consistent with Drake’s (1996) harm–evidence model, in that substantiation is likely only when cases meet both a threshold level of harm and a threshold level of evidence. However, there is enough counterevidence to suggest that this model is imperfect in explaining substantiation. In particular, some cases are not substantiated even when both harm and evidence are substantial. In addition, as discussed earlier, the independent effects of child age and gender suggest that there are other caseworker judgments that influence substantiation but are not accounted for by the model. More research is needed to test the harm–evi- dence model and perhaps develop modifications to it.

The results here underline previous researchers’ con- clusions that substantiation is an imperfect measure of child maltreatment (Drake, 1996; Yuan, Schene, English, & Johnson, 2005). Points to consider are (a) the element of unpredictability of substantiation decisions; (b) the relationship of the substantiation decision to child age and gender, independent of harm, risk, and evidence; and (c) the dependence of substantiation on the availability of evidence even when harm and risk are judged to be moderate to high. These results are likely to be relevant to current work debating the wisdom of the current reliance of CWS on substantiation (Drake, 1996; English et al., 2002; Leiter, Myers, & Zingraff, 1994). The cited researchers and other experts have argued that substanti- ation does not rationally guide which families should receive services, nor does it provide an accurate basis for data on child maltreatment.

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

Child welfare policy should take into account these results and those from other studies on substantiation. First, as other researchers have suggested, CWS agencies should not use substantiation as a proxy variable for maltreatment in data analysis and decision making. Second, the policy of a number of states to reserve child and family services for substantiated cases should be re-examined, given that the substantiation decision often does not correspond to the level of risk and harm that children experience and, therefore, the need for ser- vices they have. Likewise caseworkers should not let the substantiation decision overly influence their decisions on service delivery for individual cases. Third, CWS agencies should assess carefully what judgments their investigators are making that influence the substantiation decision. If investigators have legitimate criteria that are not included in the definition of substantiation, the defi- nition may need to be changed. If investigators’ criteria are not always sound, more training and quality control is needed. Fourth, CWS agencies should consider the limitations of substantiation in their decision making about dual track systems. Assigning lower risk reports to an assessment track that does not include a substantia- tion decision is one way to mitigate the limitations of the substantiation process (see, e.g., Yuan et al., 2005).

Limitations

The data used here for substantiation decision, level of harm, risk assessment, and evidence were provided ret- rospectively by the caseworker. This may introduce a degree of subjectivity and error to the analysis. On the other hand, caseworkers’ knowledge and perceptions are exactly what would be operating to lead them to make decisions about substantiation.

In addition, the time interval between the substantia- tion decision and the caseworker interview may have introduced some measurement error. It is possible that the substantiation decision affected caseworkers’ retro- spective judgments, despite the instruction to respond regardless of the outcome of the investigation. Also, new information learned after the substantiation decision may have influenced caseworkers’ memories. However, given that the different sources of possible measurement error might lead either to higher or lower scores on harm, risk, and evidence, we see no reason to expect a particular bias.

The caseworker judgment items have limitations as well. The categories of mild, moderate, and severe harm and risk were not operationalized, nor were the cate- gories of probably or clearly sufficient or insufficient evidence. Note, however, that it would often be a state, an agency, or an individual worker’s own definition of

these categories that would influence the substantiation decision. Another limitation is that the text of the evi- dence question specifically mentions substantiation, which may be one reason it was the most influential caseworker judgment variable in this analysis.

Future Research

Additional research is needed on the processes under- lying substantiation. Studies could gather a wider range of judgments that might influence substantiation, such as the likelihood of service delivery or other pragmatic consider- ations. Research could also examine further how state and agency policy interacts with worker decision making. A study of states that use a three-category substantiation process could examine how harm, risk, and evidence dif- fer for substantiated, unsubstantiated, and indicated cases. New studies could also compare determinants of substan- tiation in states with a traditional, one-track investigation system versus states with a dual-track, alternative response system. More on substantiation could be done with both the current NSCAW data set and data from a new NSCAW study that begins data collection in 2008. Data on substantiation from the 50 states are available from the NCANDS (see http://ndacan.cornell.edu). State and local agencies could also conduct their own studies of substantiation using agency data systems.

Conclusion

According to caseworker reports on harm, risk, and evidence, substantiation works in most cases as it should, but there are many cases in which the relationship between these caseworker judgments and substantiation raises question about appropriate policy and practice. Clearly, substantiation along with policy and practice related to it are due for a fresh appraisal by state CWS agencies, as well as researchers and policymakers.

References

Alter, C. (1985). Decision-making factors in cases of child neglect. Child Welfare, 74, 99–111.

Besharov, D. J. (1988). Child abuse and neglect reporting and inves- tigation: Policy guidelines for decision making. Family Law Quarterly, 22, 1–15.

Child Welfare Information Gateway. (2003, June). Decision-making in unsubstantiated child protective services cases: Synthesis of recent research. Retrieved January 11, 2007, from http://www.childwel- fare.gov/pubs/focus/ decisionmaking/decisionmaking.pdf

DePanfilis, D., & Girvin, H. (2005). Investigating child maltreatment in out-of-home care: Barriers to effective decision-making. Children and Youth Services Review, 27, 353–374.

Cross, Casanueva / Caseworker Judgments and Substantiation 51

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

52 Child Maltreatment

DePanfilis, D., & Salus, M. K. (2003). Child protective services: A guide for caseworkers. Washington, DC: U.S. Department of Health and Human Services. Available online at http://www.ssw.umaryland .edu/faculty_and_research/bios/depanfilis/cps.pdf

Drake, B. (1995). Associations between reporter type and assessment outcomes in child protective services referrals. Children and Youth Services Review, 17, 503–522.

Drake, B. (1996). Unraveling “unsubstantiated.” Child Maltreatment, 1, 261–271.

Drake, B. (2000). How do I decide whether to substantiate a report? In H. Dubowitz & D. DePanfilis (Eds.), Handbook for child pro- tection practice (pp. 113–117). Thousand Oaks, CA: Sage.

Eckenrode, J., Powers, J., Doris, J., Munsch, J., & Bolger, N. (1988). Substantiation of child abuse and neglect reports. Journal of Consulting and Clinical Psychology, 56, 9–16.

English, D. J., Marshall, D. B., Coghlan, L., Brummel, S., & Orme, M. (2002). Causes and consequences of the substantiation deci- sion in Washington State child protective services. Children and Youth Services Review, 24, 817–851.

Fluke, J. D., Hollinshead, D. M., & Walter R. McDonald & Associates. (2003, January). Child maltreatment recurrence. Duluth, GA: National Resource Center on Child Maltreatment. Retrieved August 8, 2004, from http://www.nrccps.org/PDF/Maltreatment Recurrence.pdf

Gambrill, E., & Shlonsky, A. (2000). Risk assessment in context. Children and Youth Services Review, 22, 813–837.

Giovannoni, J. M. (1989). Substantiated and unsubstantiated reports of child maltreatment. Children and Youth Services Review, 11, 299–318.

Jason, J., Andereck, N., Marks, J., & Tyler, C. (1982). Child abuse in Georgia: A method to evaluate risk factors and reporting bias. Journal of Public Health, 72, 1353–1358.

King, G., Trocmé, N., & Thatte, N. (2003). Substantiation as a multi- tier process: The results of a NIS-3 analysis. Child Maltreatment, 8, 173–182.

Leiter, J., Myers, K., & Zingraff, M. (1994). Substantiated and unsub- stantiated cases of child maltreatment: Do their consequences dif- fer? Social Work Research, 18, 67–78.

Levine, M. (1998). Do standards of proof affect decision making in child protective investigations? Law and Human Behavior, 22, 341–347.

NSCAW Research Group. (2002). Methodological lessons from the National Survey of Child and Adolescent Well-Being: The first three years of the USA’s first national probability study of children and families investigated for abuse and neglect. Children and Youth Services Review, 24, 513–541.

Pecora, P. (1991). Investigating allegations of child maltreatment: The strengths and limitations of current risk assessment systems. Child & Youth Services, 15, 73–92.

RTI International. (2002). SUDAAN user’s manual, release 8.0 (Vol. 2). Research Triangle Park, NC: Author.

Ryan, S., Wiles, D., Cash, S., & Siebert, C. (2005). Risk assessments: Empirically supported or values driven? Children and Youth Services Review, 27, 213–225.

Schene, P., Oppenheimer, S., & Senderling, C. (2005, Summer). Choosing the path less traveled: Strengthening California families through differential response. Sacramento, CA: Foundation Consortium for California’s Children and Youth. Retrieved August, 2007, from http://www.cwda.org/downloads/ DifferentialResponsePolicyBrief_FINAL.pdf

Trocmé, N., Knoke, D., Fallon, B., & MacLaurin, B. (2006). Substantiating child maltreatment: CIS-2003. Toronto, Ontario,

Canada: Centre of Excellence for Child Welfare. Retrieved September 14, 2007, from http://www.cecw-cepb.ca/files/file/en/ SubChildMaltreatment40E.pdf

Trocmé, N., Tam, K. K., & McPhee, D. (1995). Correlates of sub- stantiation of maltreatment in child welfare investigation. In J. Hudson & B. Galaway (Eds.), Child welfare in Canada: Research and policy implications (pp. 20–39). Toronto, Ontario, Canada: Thompson Educational Press.

U.S. Department of Health and Human Services, Administration for Children and Families. (2002). Child maltreatment 2000. Washington, DC: U.S. Government Printing Office.

U.S. Department of Health and Human Services, Administration for Children and Families. (2005). National Survey of Child and Adolescent Well-Being: CPS sample component, Wave 1 data analysis report. Washington, DC: Author.

U.S. Department of Health and Human Services, Administration for Children and Families. (2007). Child maltreatment 2005. Washington, DC: U.S. Government Printing Office.

U.S. Department of Health and Human Services, Administration for Children and Families. (n.d.). Does the substantiation of child maltreatment relate to child well-being and service receipt? Findings from the NSCAW Study. [National Survey of Child and Adolescent Well-Being Research Brief No. 9.] Washington, DC: Author.

U.S. Department of Health and Human Services, Office on Child Abuse and Neglect. (2003). The Child Abuse Prevention and Treatment Act, as amended by the Keeping Children and Families Safe Act of 2003. Washington, DC: Author.

Wald, M. S., & Woolverton, M. (1990). Risk assessment: The emperor’s new clothes? Child Welfare, 69, 483–511.

Walter R. McDonald & Associates. (2003). National study of child protective services systems and reform efforts. Denver, CO.

Winefield, H. R., & Bradley, P. W. (1992). Substantiation of reported child abuse or neglect: Predictors and implications. Child Abuse & Neglect, 16, 661–671.

Yuan, Y. Y., Schene, P., English, D., & Johnson, C. (2005, April 18–23). Whither substantiation? The role of substantiation in future child protective service policies and practice. Paper pre- sented at the 15th National Conference on Child Abuse and Neglect, Boston, MA.

Zuravin, S. J., Orme, J. G., & Hegar, R. L. (1995). Disposition of child physical abuse reports: Review of the literature and test of a predictive model. Children and Youth Services Review, 17, 547–566.

Theodore P. Cross is a clinical psychologist and is now a visiting research professor at the Children and Family Research Center in the School of Social Work at the University of Illinois at Urbana–Champaign. He currently conducts research on place- ment outcomes, service delivery, and well-being for children in foster care. Over the past 23 years, his research has focused on children’s services and on the criminal justice response to child maltreatment.

Cecilia Casanueva is a research psychologist at RTI International. She currently conducts research on developmental problems and special health care needs among young children who have been abused or neglected and their need for early intervention ser- vices. In recent years, her research has focused on domestic vio- lence, parenting, adoption, placement outcomes, service delivery, and well-being for children investigated for maltreatment.

at NOVA SOUTHEASTERN UNIV on November 18, 2009 http://cmx.sagepub.comDownloaded from

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.1000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 100 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true /ACaslon-Ornaments /AGaramond-BoldScaps /AGaramond-Italic /AGaramond-Regular /AGaramond-RomanScaps /AGaramond-Semibold /AGaramond-SemiboldItalic /AGar-Special /AkzidenzGroteskBE-Bold /AkzidenzGroteskBE-BoldIt /AkzidenzGroteskBE-It /AkzidenzGroteskBE-Light /AkzidenzGroteskBE-LightOsF /AkzidenzGroteskBE-Md /AkzidenzGroteskBE-MdIt /AkzidenzGroteskBE-Regular /AkzidenzGroteskBE-Super /AlbertusMT /AlbertusMT-Italic /AlbertusMT-Light /Aldine401BT-BoldA /Aldine401BT-BoldItalicA /Aldine401BT-ItalicA /Aldine401BT-RomanA /Aldine401BTSPL-RomanA /Aldine721BT-Bold /Aldine721BT-BoldItalic /Aldine721BT-Italic /Aldine721BT-Light /Aldine721BT-LightItalic /Aldine721BT-Roman /Aldus-Italic /Aldus-Roman /AlternateGothicNo2BT-Regular /Anna /AntiqueOlive-Bold /AntiqueOlive-Compact /AntiqueOlive-Italic /AntiqueOlive-Roman /Arcadia /Arcadia-A /Arkona-Medium /Arkona-Regular /AssemblyLightSSK /AvantGarde-Book /AvantGarde-BookOblique /AvantGarde-Demi /AvantGarde-DemiOblique /BakerSignetBT-Roman /BaskervilleBE-Italic /BaskervilleBE-Medium /BaskervilleBE-MediumItalic /BaskervilleBE-Regular /BaskervilleBook-Italic /BaskervilleBook-MedItalic /BaskervilleBook-Medium /BaskervilleBook-Regular /BaskervilleBT-Bold /BaskervilleBT-BoldItalic /BaskervilleBT-Italic /BaskervilleBT-Roman /BaskervilleMT /BaskervilleMT-Bold /BaskervilleMT-BoldItalic /BaskervilleMT-Italic /BaskervilleMT-SemiBold /BaskervilleMT-SemiBoldItalic /BaskervilleNo2BT-Bold /BaskervilleNo2BT-BoldItalic /BaskervilleNo2BT-Italic /BaskervilleNo2BT-Roman /Bauhaus-Bold /Bauhaus-Demi /Bauhaus-Heavy /BauhausITCbyBT-Bold /BauhausITCbyBT-Medium /Bauhaus-Light /Bauhaus-Medium /BellCentennial-Address /BellGothic-Black /BellGothic-Bold /Bell-GothicBoldItalicBT /BellGothicBT-Bold /BellGothicBT-Roman /BellGothic-Light /Bembo /Bembo-Bold /Bembo-BoldExpert /Bembo-BoldItalic /Bembo-BoldItalicExpert /Bembo-Expert /Bembo-ExtraBoldItalic /Bembo-Italic /Bembo-ItalicExpert /Bembo-Semibold /Bembo-SemiboldItalic /Berkeley-Black /Berkeley-BlackItalic /Berkeley-Bold /Berkeley-BoldItalic /Berkeley-Book /Berkeley-BookItalic /Berkeley-Italic /Berkeley-Medium /Berling-Bold /Berling-BoldItalic /Berling-Italic /Berling-Roman /BernhardModernBT-Bold /BernhardModernBT-BoldItalic /BernhardModernBT-Italic /BernhardModernBT-Roman /Bodoni /Bodoni-Bold /Bodoni-BoldItalic /Bodoni-Italic /Bodoni-Poster /Bodoni-PosterCompressed /Bookman-Demi /Bookman-DemiItalic /Bookman-Light /Bookman-LightItalic /Boton-Italic /Boton-Medium /Boton-MediumItalic /Boton-Regular /Boulevard /BremenBT-Black /BremenBT-Bold /CaflischScript-Bold /CaflischScript-Regular /Carta /Caslon224ITCbyBT-Bold /Caslon224ITCbyBT-BoldItalic /Caslon224ITCbyBT-Book /Caslon224ITCbyBT-BookItalic /Caslon540BT-Italic /Caslon540BT-Roman /CaslonBT-Bold /CaslonBT-BoldItalic /CaslonTwoTwentyFour-Black /CaslonTwoTwentyFour-BlackIt /CaslonTwoTwentyFour-Bold /CaslonTwoTwentyFour-BoldIt /CaslonTwoTwentyFour-Book /CaslonTwoTwentyFour-BookIt /CaslonTwoTwentyFour-Medium /CaslonTwoTwentyFour-MediumIt /CastleT-Bold /CastleT-Book /Caxton-Bold /Caxton-BoldItalic /Caxton-Book /Caxton-BookItalic /Caxton-Light /Caxton-LightItalic /CelestiaAntiqua-Ornaments /Centennial-BlackItalicOsF /Centennial-BlackOsF /Centennial-BoldItalicOsF /Centennial-BoldOsF /Centennial-ItalicOsF /Centennial-LightItalicOsF /Centennial-LightSC /Centennial-RomanSC /CenturyOldStyle-Bold /CenturyOldStyle-Italic /CenturyOldStyle-Regular /CheltenhamBT-Bold /CheltenhamBT-BoldItalic /CheltenhamBT-Italic /CheltenhamBT-Roman /Christiana-Bold /Christiana-BoldItalic /Christiana-Italic /Christiana-Medium /Christiana-MediumItalic /Christiana-Regular /Christiana-RegularExpert /Christiana-RegularSC /Clarendon /Clarendon-Bold /Clarendon-Light /ClassicalGaramondBT-Bold /ClassicalGaramondBT-BoldItalic /ClassicalGaramondBT-Italic /ClassicalGaramondBT-Roman /CMTI10 /CommonBullets /ConduitITC-Bold /ConduitITC-BoldItalic /ConduitITC-Light /ConduitITC-LightItalic /ConduitITC-Medium /ConduitITC-MediumItalic /CooperBlack /CooperBlack-Italic /CopperplateGothicBT-Bold /CopperplateGothicBT-BoldCond /CopperplateGothicBT-Heavy /CopperplateGothicBT-Roman /CopperplateGothicBT-RomanCond /Copperplate-ThirtyThreeBC /Copperplate-ThirtyTwoBC /Coronet-Regular /Courier /Courier-Bold /Courier-BoldOblique /Courier-Oblique /Critter /CS-Special-font /DextorD /DextorOutD /DidotLH-OrnamentsOne /DidotLH-OrnamentsTwo /DINEngschrift /DINEngschrift-Alternate /DINMittelschrift /DINMittelschrift-Alternate /DINNeuzeitGrotesk-BoldCond /DINNeuzeitGrotesk-Light /Dom-CasItalic /Dom-CasualBT /Ehrhard-Italic /Ehrhard-Regular /EhrhardSemi-Italic /EhrhardtMT /EhrhardtMT-Italic /EhrhardtMT-SemiBold /EhrhardtMT-SemiBoldItalic /EhrharSemi /ElectraLH-Bold /ElectraLH-BoldCursive /ElectraLH-Cursive /ElectraLH-Regular /EnglischeSchT-Bold /EnglischeSchT-Regu /ErasContour /ErasITCbyBT-Bold /ErasITCbyBT-Book /ErasITCbyBT-Demi /ErasITCbyBT-Light /ErasITCbyBT-Medium /ErasITCbyBT-Ultra /EUEX10 /EUFB10 /EUFB5 /EUFB7 /EUFM10 /EUFM5 /EUFM7 /EURB10 /EURB5 /EURB7 /EURM10 /EURM5 /EURM7 /EuropeanPi-Four /EuropeanPi-One /EuropeanPi-Three /EuropeanPi-Two /Eurostile /Eurostile-Bold /Eurostile-BoldExtendedTwo /Eurostile-ExtendedTwo /EUSB10 /EUSB5 /EUSB7 /EUSM10 /EUSM5 /EUSM7 /ExPonto-Regular /Fenice-Bold /Fenice-BoldOblique /FeniceITCbyBT-Bold /FeniceITCbyBT-BoldItalic /FeniceITCbyBT-Regular /FeniceITCbyBT-RegularItalic /Fenice-Light /Fenice-LightOblique /Fenice-Regular /Fenice-RegularOblique /Fenice-Ultra /Fenice-UltraOblique /FlashD-Ligh /Folio-Bold /Folio-BoldCondensed /Folio-ExtraBold /Folio-Light /Folio-Medium /FontanaNDEeOsF /FontanaNDEeOsF-Semibold /FormalScript421BT-Regular /Formata-Bold /Formata-MediumCondensed /FournierMT-Ornaments /FrakturBT-Regular /FranklinGothic-Book /FranklinGothic-BookItal /FranklinGothic-BookOblique /FranklinGothic-Condensed /FranklinGothic-Demi /FranklinGothic-DemiItal /FranklinGothic-DemiOblique /FranklinGothic-Heavy /FranklinGothic-HeavyItal /FranklinGothic-HeavyOblique /FranklinGothic-Medium /FranklinGothic-MediumItal /FranklinGothic-Roman /FrizQuadrataITCbyBT-Bold /FrizQuadrataITCbyBT-Roman /Frutiger-Black /Frutiger-BlackCn /Frutiger-BlackItalic /Frutiger-Bold /Frutiger-BoldCn /Frutiger-BoldItalic /Frutiger-Cn /Frutiger-ExtraBlackCn /Frutiger-Italic /Frutiger-Light /Frutiger-LightCn /Frutiger-LightItalic /Frutiger-Roman /Frutiger-UltraBlack /Futura /FuturaBlackBT-Regular /Futura-Bold /Futura-BoldOblique /Futura-Book /Futura-BookOblique /FuturaBT-Bold /FuturaBT-BoldCondensed /FuturaBT-BoldCondensedItalic /FuturaBT-BoldItalic /FuturaBT-Book /FuturaBT-BookItalic /FuturaBT-ExtraBlack /FuturaBT-ExtraBlackCondensed /FuturaBT-ExtraBlackCondItalic /FuturaBT-ExtraBlackItalic /FuturaBT-Heavy /FuturaBT-HeavyItalic /FuturaBT-Light /FuturaBT-LightCondensed /FuturaBT-LightItalic /FuturaBT-Medium /FuturaBT-MediumCondensed /FuturaBT-MediumItalic /Futura-ExtraBold /Futura-ExtraBoldOblique /Futura-Heavy /Futura-HeavyOblique /Futura-Light /Futura-LightOblique /Futura-Oblique /GalliardITCbyBT-Italic /GalliardITCbyBT-Roman /Garamond-Antiqua /Garamond-BoldCondensed /Garamond-BoldCondensedItalic /Garamond-BookCondensed /Garamond-BookCondensedItalic /Garamond-Halbfett /GaramondITCbyBT-Bold /GaramondITCbyBT-BoldCondensed /GaramondITCbyBT-BoldCondItalic /GaramondITCbyBT-BoldItalic /GaramondITCbyBT-BoldNarrow /GaramondITCbyBT-BoldNarrowItal /GaramondITCbyBT-Book /GaramondITCbyBT-BookCondensed /GaramondITCbyBT-BookCondItalic /GaramondITCbyBT-BookItalic /GaramondITCbyBT-Light /GaramondITCbyBT-LightCondensed /GaramondITCbyBT-LightCondItalic /GaramondITCbyBT-LightItalic /GaramondITCbyBT-LightNarrow /GaramondITCbyBT-LightNarrowItal /GaramondITCbyBT-Ultra /GaramondITCbyBT-UltraCondensed /GaramondITCbyBT-UltraCondItalic /GaramondITCbyBT-UltraItalic /Garamond-Kursiv /Garamond-KursivHalbfett /Garamond-LightCondensed /Garamond-LightCondensedItalic /GaramondThree /GaramondThree-Bold /GaramondThree-BoldItalic /GaramondThree-Italic /GaramondThreeSMSspl /GaramondThreespl /GaramondThreeSpl-Bold /GaramondThreeSpl-Italic /GarthGraphic /GarthGraphic-Black /GarthGraphic-Bold /GarthGraphic-BoldCondensed /GarthGraphic-BoldItalic /GarthGraphic-Condensed /GarthGraphic-ExtraBold /GarthGraphic-Italic /Geometric231BT-HeavyC /GeometricSlab712BT-BoldA /GeometricSlab712BT-ExtraBoldA /GeometricSlab712BT-LightA /GeometricSlab712BT-LightItalicA /GeometricSlab712BT-MediumA /GeometricSlab712BT-MediumItalA /Giddyup /Giddyup-Thangs /GillSans /GillSans-Bold /GillSans-BoldCondensed /GillSans-BoldItalic /GillSans-Condensed /GillSans-ExtraBold /GillSans-Italic /GillSans-Light /GillSans-LightItalic /GillSans-UltraBold /GillSans-UltraBoldCondensed /Gill-Special /Giovanni-Bold /Giovanni-BoldItalic /Giovanni-Book /Giovanni-BookItalic /Glypha /Glypha-Bold /Glypha-BoldOblique /Glypha-Oblique /Goudy /Goudy-Bold /Goudy-BoldItalic /Goudy-ExtraBold /Goudy-Italic /GoudyOldStyleBT-Bold /GoudyOldStyleBT-BoldItalic /GoudyOldStyleBT-ExtraBold /GoudyOldStyleBT-Italic /GoudyOldStyleBT-Roman /GoudySans-Bold /GoudySans-BoldItalic /GoudySansITCbyBT-Bold /GoudySansITCbyBT-BoldItalic /GoudySansITCbyBT-Medium /GoudySansITCbyBT-MediumItalic /GoudySans-Medium /GoudySans-MediumItalic /Granjon /Granjon-Bold /Granjon-BoldOsF /Granjon-Italic /Granjon-ItalicOsF /Granjon-SC /GreymantleMVB-Ornaments /Helvetica /Helvetica-Black /Helvetica-BlackOblique /Helvetica-Black-SemiBold /Helvetica-Bold /Helvetica-BoldOblique /Helvetica-Condensed /Helvetica-Condensed-Black /Helvetica-Condensed-BlackObl /Helvetica-Condensed-Bold /Helvetica-Condensed-BoldObl /Helvetica-Condensed-Light /Helvetica-Condensed-LightObl /Helvetica-Condensed-Oblique /Helvetica-Light /Helvetica-LightOblique /Helvetica-Narrow /Helvetica-Narrow-Bold /Helvetica-Narrow-BoldOblique /Helvetica-Narrow-Oblique /HelveticaNeue-BlackCond /HelveticaNeue-BlackCondObl /HelveticaNeue-Bold /HelveticaNeue-BoldCond /HelveticaNeue-BoldCondObl /HelveticaNeue-BoldExt /HelveticaNeue-BoldExtObl /HelveticaNeue-BoldItalic /HelveticaNeue-Condensed /HelveticaNeue-CondensedObl /HelveticaNeue-ExtBlackCond /HelveticaNeue-ExtBlackCondObl /HelveticaNeue-Extended /HelveticaNeue-ExtendedObl /HelveticaNeue-Heavy /HelveticaNeue-HeavyCond /HelveticaNeue-HeavyCondObl /HelveticaNeue-HeavyExt /HelveticaNeue-HeavyExtObl /HelveticaNeue-HeavyItalic /HelveticaNeue-Italic /HelveticaNeue-Light /HelveticaNeue-LightCond /HelveticaNeue-LightCondObl /HelveticaNeue-LightItalic /HelveticaNeueLTStd-Md /HelveticaNeueLTStd-MdIt /HelveticaNeue-Medium /HelveticaNeue-MediumCond /HelveticaNeue-MediumCondObl /HelveticaNeue-MediumExt /HelveticaNeue-MediumExtObl /HelveticaNeue-MediumItalic /HelveticaNeue-Roman /HelveticaNeue-ThinCond /HelveticaNeue-ThinCondObl /HelveticaNeue-UltraLigCond /HelveticaNeue-UltraLigCondObl /Helvetica-Oblique /HelvLight /Humanist521BT-Bold /Humanist521BT-BoldCondensed /Humanist521BT-BoldItalic /Humanist521BT-ExtraBold /Humanist521BT-Italic /Humanist521BT-Light /Humanist521BT-LightItalic /Humanist521BT-Roman /Humanist521BT-RomanCondensed /Humanist521BT-UltraBold /Humanist521BT-XtraBoldCondensed /Humanist777BT-BlackB /Humanist777BT-BlackItalicB /Humanist777BT-BoldB /Humanist777BT-BoldItalicB /Humanist777BT-ItalicB /Humanist777BT-LightB /Humanist777BT-LightItalicB /Humanist777BT-RomanB /ICMEX10 /ICMMI8 /ICMSY8 /ICMTT8 /ILASY8 /ILCMSS8 /ILCMSSB8 /ILCMSSI8 /Imago-Book /Imago-BookItalic /Imago-ExtraBold /Imago-ExtraBoldItalic /Imago-Medium /Imago-MediumItalic /Industria-Inline /Industria-InlineA /Industria-Solid /Industria-SolidA /Insignia /Insignia-A /IPAExtras /IPAHighLow /IPAKiel /IPAKielSeven /IPAsans /JoannaMT /JoannaMT-Bold /JoannaMT-BoldItalic /JoannaMT-Italic /KlangMT /Kuenstler480BT-Black /Kuenstler480BT-Bold /Kuenstler480BT-BoldItalic /Kuenstler480BT-Italic /Kuenstler480BT-Roman /KunstlerschreibschD-Bold /KunstlerschreibschD-Medi /Lapidary333BT-Black /Lapidary333BT-Bold /Lapidary333BT-BoldItalic /Lapidary333BT-Italic /Lapidary333BT-Roman /LASY10 /LASY5 /LASY6 /LASY7 /LASY8 /LASY9 /LASYB10 /LatinMT-Condensed /LCIRCLE10 /LCIRCLEW10 /LCMSS8 /LCMSSB8 /LCMSSI8 /LDecorationPi-One /LDecorationPi-Two /Leawood-Black /Leawood-BlackItalic /Leawood-Bold /Leawood-BoldItalic /Leawood-Book /Leawood-BookItalic /Leawood-Medium /Leawood-MediumItalic /LegacySans-Bold /LegacySans-BoldItalic /LegacySans-Book /LegacySans-BookItalic /LegacySans-Medium /LegacySans-MediumItalic /LegacySans-Ultra /LegacySerif-Bold /LegacySerif-BoldItalic /LegacySerif-Book /LegacySerif-BookItalic /LegacySerif-Medium /LegacySerif-MediumItalic /LegacySerif-Ultra /LetterGothic /LetterGothic-Bold /LetterGothic-BoldSlanted /LetterGothic-Slanted /Life-Bold /Life-Italic /Life-Roman /LINE10 /LINEW10 /Lithos-Black /Lithos-Regular /LOGO10 /LOGO8 /LOGO9 /LOGOBF10 /LOGOSL10 /LOMD-Normal /LubalinGraph-Book /LubalinGraph-BookOblique /LubalinGraph-Demi /LubalinGraph-DemiOblique /LucidaMath-Symbol /LydianBT-Bold /LydianBT-BoldItalic /LydianBT-Italic /LydianBT-Roman /LydianCursiveBT-Regular /Marigold /MathematicalPi-Five /MathematicalPi-Four /MathematicalPi-One /MathematicalPi-Six /MathematicalPi-Three /MathematicalPi-Two /Melior /Melior-Bold /Melior-BoldItalic /Melior-Italic /MercuriusCT-Black /MercuriusCT-BlackItalic /MercuriusCT-Light /MercuriusCT-LightItalic /MercuriusCT-Medium /MercuriusCT-MediumItalic /MercuriusMT-BoldScript /Meridien-Medium /Meridien-MediumItalic /Meridien-Roman /Minion-Black /Minion-Bold /Minion-BoldCondensed /Minion-BoldCondensedItalic /Minion-BoldItalic /Minion-Condensed /Minion-CondensedItalic /MinionExp-Italic /MinionExp-Semibold /MinionExp-SemiboldItalic /Minion-Italic /Minion-Ornaments /Minion-Regular /Minion-Semibold /Minion-SemiboldItalic /MonaLisa-Recut /MSAM10 /MSAM10A /MSAM5 /MSAM6 /MSAM7 /MSAM8 /MSAM9 /MSBM10 /MSBM10A /MSBM5 /MSBM6 /MSBM7 /MSBM8 /MSBM9 /MTEX /MTEXB /MTEXH /MTGU /MTGUB /MTMI /MTMIB /MTMIH /MTMS /MTMSB /MTMUB /MTMUH /MTSY /MTSYB /MTSYH /MTSYN /MusicalSymbols-Normal /Myriad-Bold /Myriad-BoldItalic /Myriad-CnBold /Myriad-CnBoldItalic /Myriad-CnItalic /Myriad-CnSemibold /Myriad-CnSemiboldItalic /Myriad-Condensed /Myriad-Italic /Myriad-Roman /Myriad-Sketch /Myriad-Tilt /NeuzeitS-Book ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox false /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (U.S. Web Coated \050SWOP\051 v2) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown /CreateJDFFile false /SyntheticBoldness 1.000000 /Description << /FRA <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> /JPN <FEFF3053306e8a2d5b9a306f30019ad889e350cf5ea6753b50cf3092542b308000200050004400460020658766f830924f5c62103059308b3068304d306b4f7f75283057307e30593002537052376642306e753b8cea3092670059279650306b4fdd306430533068304c3067304d307e305930023053306e8a2d5b9a30674f5c62103057305f00200050004400460020658766f8306f0020004100630072006f0062006100740020304a30883073002000520065006100640065007200200035002e003000204ee5964d30678868793a3067304d307e30593002> /DEU <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> /PTB <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> /DAN <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> /NLD <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> /ESP <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> /SUO <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> /ITA <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> /NOR <FEFF004200720075006b00200064006900730073006500200069006e006e007300740069006c006c0069006e00670065006e0065002000740069006c002000e50020006f00700070007200650074007400650020005000440046002d0064006f006b0075006d0065006e0074006500720020006d006500640020006800f80079006500720065002000620069006c00640065006f00700070006c00f80073006e0069006e006700200066006f00720020006200650064007200650020007500740073006b00720069006600740073006b00760061006c0069007400650074002e0020005000440046002d0064006f006b0075006d0065006e00740065006e00650020006b0061006e002000e50070006e006500730020006d006500640020004100630072006f0062006100740020006f0067002000520065006100640065007200200035002e00300020006f0067002000730065006e006500720065002e> /SVE <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> /ENU <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> >> >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice