Maltreatment type and behaviors: Does listening matter?

Child Abuse & Neglect 38 (2014) 2062–2071

Contents lists available at ScienceDirect

Child Abuse & Neglect

Maltreatment type and behaviors: Does listening matter? Darcey H. Merritt a,∗ , Susan M. Snyder b a b

Silver School of Social Work, New York University, 1 Washington Square North, Office # 316, New York, NY 10003, USA School of Social Work, University of North Carolina at Chapel Hill, 325 Pittsboro St. CB#3550, Chapel Hill, NC 27599, USA

a r t i c l e

i n f o

Article history: Received 1 September 2014 Accepted 20 October 2014 Available online 18 November 2014 Keywords: Child maltreatment Behaviors Caseworker support Listening Physical abuse Neglect

a b s t r a c t This article presents an exploratory assessment of whether children’s perceptions of caseworker support (e.g. feeling listened to) moderates the relationship between the type of maltreatment and problematic behaviors. Relying on data collected for the National Survey of Child and Adolescent Well-Being (NSCAW I), this research measures how often children felt listened to by their caseworkers and the effect on the relationship between two types of maltreatment (e.g. physical abuse and neglect) and problematic child behaviors. Results indicate that whereas children reported feeling listened to most of the time, there are significant differences in the probabilities of problematic behavior scores between physically abused and neglected children according to how often they felt listened to. With the exception of those children who felt listened to all of the time, physically abused children have a higher probability of problematic behaviors than neglected children. Comparisons between the two maltreatment types indicate a greater impact of listening on physically abused children across the continuum of feeling listened to (e.g. never to all of the time), than for neglected children, except for at the highest level of listening, as results indicate a small, but significant difference indicating neglected children are more positively impacted by listening than physically abused kids. Implications for practice are that children’s perceptions of support from caseworkers may influence behavioral outcomes differently according to maltreatment type. Additionally, these findings encourage the inclusion of children’s perspectives regarding the relationships they have with caseworkers. © 2014 Published by Elsevier Ltd.

Introduction An essential tenet of social work is beginning treatment from the client’s perspective. Gaining an understanding of children’s perceptions regarding the quality of relationships with their caseworkers is important to inform program developers about whether service delivery and treatment planning adequately account for the child’s perspective. Acknowledging the opinions and perceptions of children served by the child welfare system enhances the relationship between children and the adults who care for them (Kufeldt, 1984; Merritt, 2008; Merritt and Franke, 2010). Services provided by the child welfare system would be bolstered if evaluation protocol integrated an assessment of children’s perceptions of the caseworker/child relationship in relation to positive outcomes. An examination of ways in which children’s perceived levels of support are related to behavioral and developmental outcomes can contribute vital information for treatment planning. Understanding children’s perceptions regarding their experiences while interfacing with the child welfare system is essential to evaluating the relational aspect of service delivery, although consistently absent from the empirical literature (Fox, Frasch, & Berrick, 2000). Incorporating an assessment regarding the relationship between children and their

∗ Corresponding author. http://dx.doi.org/10.1016/j.chiabu.2014.10.019 0145-2134/© 2014 Published by Elsevier Ltd.

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

2063

caseworkers has not been previously utilized in quantitative child welfare program analysis, nor has it typically influenced program or policy development (Chapman, Wall, & Barth, 2004). Given children are the primary clients receiving services, effective treatment delivery should incorporate an evaluation from the perspective of the children. There have been crosssectional, retrospective assessments regarding the perceptions of previous foster children, now adults (Wilson, 1996), but few inquiring with the client while in care (Berrick, Frasch, & Fox, 2000), such as the research presented herein. Scholars have noted that children want to be queried about their feelings in placement (Festinger, 1983; Johnson, Yoken, & Voss, 1990; Wilson, 1994, 1996). Moreover, children’s relational experiences influence their commitment to continuing with the decided upon treatment plan (Fox et al., 2000). Hence, it is essential to assess the extremely important relationship with the practitioners, as influential regarding desired treatment outcomes. An inquiry such as this can provide a meaningful evaluation component in assessing the efficacy of child welfare intervention. The research presented here focuses on assessing behavioral outcome differences between children who have been physically abused and those neglected based on how often they felt listened to by their caseworkers. As noted, studies have documented a dearth of research that includes children’s perceptions while under the care of the child welfare system (Merritt, 2008; Berrick et al., 2000; Chapman & Christ, 2008; Colton, 1989; Fox et al., 2000; Johnson et al., 1990; Needell, Webster, Curraco-Alamin, & Armijo, 1998). The retrospective nature of previous studies should be considered with caution due to memory loss, social desirability biases, and maturation. Other limitations with previous research have been small sample sizes, lack of control groups, indistinguishing between types of fostering (i.e., kinship, long term foster care) and a lack of studies incorporating mixed methods (quantitative and qualitative). Perceptions are subject to many externalized factors, which further necessitate an inquiry inclusive of possible moderating variables. For the most part, research has been designed in ways that seek to capture children’s perceptions of love, comfortability, strength of bond, feeling secure and closeness to their caregivers (Merritt, 2008; Merritt and Franke, 2010; Fanshel, Finch, & Grundy, 1990; Johnson et al., 1990; Triseliotis, 1984, 2002; Wilson, 1996; Zimmerman, 1982), as opposed to perceived relatedness to caseworkers as a factor in positive outcomes. While assessments from the perspective of caregivers, teachers, and caseworkers involved with the child welfare system in their efforts to support children in care have been widely documented, feedback from the children has been scant (Goerge, Wulcyzn & Harden, 1996; Whiting & Lee, 2003). Children are the primary consumers of the child welfare system, and as such, must have input in any thorough, client centered evaluation effort (Fox et al., 2000). Researchers have noted the seminal work of Trudy Festinger (1983), which, while focusing on children’s reflections of their experiences in foster care, identified the gap in research concerning the perspectives of the children receiving child welfare services (Festinger, 1983). Of those studies that address client feedback from the perspective of the child, the focus has been understanding what children feel about their experiences living away from home, being separated from siblings, school transitions and explanations for removal rather than client satisfaction with caseworker contribution (Fox et al., 2000; Johnson et al., 1990; Wilson, 1996). Why Does It Matter If Children Feel Listened to by Their Caseworkers? This study contributes a rarely studied inquiry regarding components of relatedness between the child welfare system population and their caseworkers. In particular, it presents much needed information about the supportive nature of the relationship with their caseworkers from the perspective of children while they are experiencing child welfare services using a large nationally representative sample. Despite the scant efforts to include children’s perceptions in the literature, there is a skeletal framework with which to begin further inquiry (Merritt and Franke, 2010). An examination of factors that characterize the relational quality between children and their caseworkers can contribute rich information to scholars studying perceived relatedness as associated with child well-being outcomes and successful adjustment in transitional environments. Caseworkers should remain cognizant of the value of listening well and querying children about their feelings regarding the provided treatment and services. How well and how often children feel listened to by caseworkers is related to whether they feel understood and supported during traumatic and difficult times while in care. Practitioners charged with implementing treatment plans are remiss in not assessing the child’s perceived quality of the service provided. An inquiry such as this provides insight about the level of trust children have with their caseworkers, how they feel about the treatment plan, and ways in which they can be empowered through active listening. Additionally, listening well as a caseworker is equally important to children’s successful adjustment and development. Maltreated children have reported puzzling patterns of relatedness (e.g., expressing high quality relationships while simultaneously displaying a strong desire for increased closeness), while children who have not been maltreated have more consistent patterns (e.g., satisfaction with the quality of the relationship and level of closeness) (Lynch and Cichetti, 1991). Patterns of relatedness or levels of perceived support have not been studied with the intent to distinguish outcome differences according to maltreatment type. Moreover, in addition to contributing an understanding of outcome differences based on perceptions of caseworker support, this study is also poised to contribute to the growing emphasis on highlighting distinct and separate characteristics of physical abuse and neglect. Oftentimes, physical abuse and neglect (of which there are a few categories, e.g. provision and supervisory) are characterized under the overarching term ‘child maltreatment’, however, the risk factors and accompanying characteristics are distinctly different, thus warranting refined definitions and comparative analyses. Specifically, components of the relatedness concept where it concerns caseworkers, such as the whether children feel listened to, understood, satisfactorily informed, and invested in their treatment plan has not been widely addressed

2064

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

in the literature. Possible correlations between the types of maltreatment a child has experienced and outcomes should be assessed according to influential factors, such as the perceived quality of relationships with caseworkers. One must be cognizant of the subjective nature in assessing perceived relatedness with caseworkers, as well as the potential vulnerability to a variety of interpretations. Assessing children’s perceptions concerning their relationships with caseworkers has implications regarding the level of trust children afford non-relational authority figures. In doing so, such an inquiry may consequently alleviate children’s anxiety by allowing them to assess and evaluate the genuine efforts of those practitioners charged with protecting them from harm. Gaining understanding and supporting children’s relationships with social workers is a crucial part of the treatment process. Children’s perceptions of relatedness with their caseworkers are formed and affected by many factors at both individual and environmental levels, however, listening well is a necessary first step. As children navigate the child welfare system, their coping skills are heavily impacted and influenced by the characteristics of their environment, which includes those responsible for implementing child protective services. Child Perceived Caseworker Relationship Quality as a Measure of Social Capital This study employs the tenets of Social Capital theory (Coleman, 1988) in an effort to assess children’s feelings regarding the relationship with their caseworkers. Social capital in this context refers to the social networks and the relationships between adults and children, outside of the family that are of value for the child’s well-being (Coleman, 1990). Moreover, social capital refers to the quality and depth of relationships, characterized as the interpersonal networks of those with reciprocal trust (Coleman, 1990; Lochner, Kawachi, & Kennedy, 1999). Applying the tenets of social capital theory in this context will help researchers and practitioners to understand the relational dynamics between caseworkers and the children they serve. A key component of social capital refers to the quality of relationships considered as a resource and positively associated with well being. Patterns of relatedness have not been studied specifically between children and caseworkers; or with an emphasis on the perceived quality of listening. Levels of social capital available to children, especially maltreated children, impact their ability to cope with challenging personal experiences. Hence, this concept of relatedness (e.g. listening well) as supported by social capital theory, represents a fluid and trusting child/caseworker relationship. Although there are many interpretations of relatedness, we are positing the concept of relatedness in this context can be appropriately measured by child welfare involved children’s perceptions of feeling listened to consistently. It is the interplay between the child and non-relational external networks (e.g. caseworkers) that is likely associated with a child’s treatment outcomes, which may vary according to the specific type of maltreatment experienced. The network between the child and caseworker is considered a valuable resource, consisting of reciprocity and mutual trust, which enable children and caseworkers to work toward a common purpose. If children who have experienced maltreatment are able to express themselves and perceive their feelings have been properly acknowledged, perhaps they will be more willing to engage in behavioral treatment plans. Maltreated children are likely to have lower levels of social capital due to tenuous familial relations; multiple moves to temporary homes, communities and schools; and environmental stressors. Their caseworkers are oftentimes one of a few reliable people in their lives, charged with their care. Applying social capital to this line of inquiry allows one to understand how these resources (e.g. listening caseworkers) or components of social capital differentially influence behaviors according to specific maltreatment types. The characteristics of physically abused children versus those of neglected children may differ according to their perceptions of relatedness and feeling listened to, and allow for an assessment of the strength of external networks. Caseworkers are very important in the lives of children who have experienced maltreatment; hence an evaluation of their relational skills with these children is necessary to identify best practices. Children’s perceptions of support from child welfare caseworkers in the context of traumatic experiences can be considered a component of social capital because feeling heard (e.g. listened to well) will likely result in children feeling understood, and relates to the quality of the interpersonal bond between the child and caseworker. Further, from a service delivery perspective, children who perceive high-quality relational experiences with their caseworkers are more likely to feel empowered as a participant in their treatment plan. The coping mechanisms of maltreated children would be greatly enhanced if they were able to benefit from acknowledgment of their perceptions about how well they feel heard and listened to during such a difficult time in their lives as they navigate involvement with the child welfare system. Current Study According to U.S. Department of Health and Human Services (1997), conducting research within an ecological perspective is the best approach to address causes, consequences and treatment formulations for maltreated children. This research presents an assessment at the mezzo level (e.g. child welfare system) of a child’s ecological environment with respect to their perceptions of caseworker support and serves to incorporate the voices of children regarding these experiences with caseworkers in the child welfare literature. Further, this study aims to identify associations between children’s perceived levels of support via listening well and behavioral outcomes. Relying on data collected for the National Survey of Child and Adolescent Well-Being (NSCAW I), maltreatment type (e.g. physical abuse and neglect) are used to predict clinical levels of problematic behaviors differentiated by levels of feeling listened to by caseworkers. Overall, this study seeks to contribute an account of children’s opinions and ways in which levels of support predict behaviors as they vary according to maltreatment type. Children who have either been physically abused or neglected (combined failure to protect neglect and supervisory

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

2065

neglect) responded to a measure of caseworker relatedness (feeling listened to), which was tested as a moderator of the relationship between maltreatment type and behavior scores. Child Behavior Checklist (CBCL). The dependent variable is a standardized measure of behavior scores, Child Behavior Checklist (CBCL) (Achenbach, 1991). Based on parent report, children were rated on a host of emotional and behavioral problems. The CBCL was developed to capture accounts of behavior across eight categories (Withdrawn, Somatic Complaints, Anxious/Depressed, Social Problems, Thought Problems, Attention Problems, Delinquent Behavior, and Aggressive Behavior) of syndromes. The CBCL is widely used and solid psychometric properties are well documented (Cronbach’s alpha is .96 for total problems; Intraclass correlation coefficient (ICC) is .96 for problem items; Construct validity ranging from .59 to .88) (Achenbach, 1991). The scale is measured as externalizing, internalizing, and total problem behaviors. The cutoff for children classified as having clinical/borderline problem behaviors is a score of 60 and above (Achenbach, 1991). This recommended cutoff was implemented in this study to distinguish between children with scores below 60 (no problems) and those with scores 60 and above (problematic behaviors). Maltreatment Type. The NSCAW I dataset documents multiple types of maltreatment as recorded by the investigating social worker at the data collection points. There are limitations of approaching research concerning the broadly defined concept of child maltreatment as a sum of multiple types of maltreatment. Current research lacks a consensus on a refined definition of the meaning of specific types of child maltreatment (e.g. educational neglect). Generally speaking, child maltreatment refers to “child abuse and neglect.” However, the expression “child maltreatment” fails to distinguish between physical abuse and neglect. Oftentimes types of child neglect and physical abuse co-occur and present challenges for assessing correlates of specific types of maltreatment. Distinguishing the correlates of each type of maltreatment is a useful approach to inform service delivery efforts and tailor treatment for optimal outcomes. Neglect is a nebulous term characterized by many forms, such as emotional neglect and educational neglect. The two primary forms of neglect measured in this study are ‘failure to provide’ and ‘supervisory’ neglect. The former refers to a lack of provisions appropriate for the well being of children and the latter refers to negligence concerning supervision. For purposes of this research, these two most common forms were combined as one category in order to capture overall neglect. Further, assessing children’s satisfaction with having been listened to by caseworkers as it relates to behavior scores according to types of neglect is more easily interpretable as a combined measure. In essence, the distinction made herein is between the commission of an abusive act (physical abuse) and the omission of a certain level of care, provisional and supervisory (neglect). Children Feel Listened to by Caseworker. An item measuring how often children felt listened to was used as a proxy for child satisfaction and relatedness with caseworkers. Unfortunately, there have not been any studies specifically investigating how children feel about the support they’ve been given by caseworkers as they navigate the child welfare system and cope with the consequences of maltreatment (Fox et al., 2000). Optimal support serves to buffer the potential negative impact children face while interfacing with the child welfare system and is likely associated with positive developmental trajectories and behavioral outcomes. Moreover, ensuring children are empowered by supportive relationships may increase their resiliency. Methods Overall, the research question examined in this article focuses on whether the type of maltreatment is predictive of problem behaviors and how children’s cognitive structures regarding their satisfaction with their caseworkers moderates this relationship. The specific research questions are: Does maltreatment type (physical abuse and neglect) predict problem behaviors? Are children’s responses regarding how often they felt listened to by caseworkers moderate the relationship between maltreatment type and problem behaviors? Sampling The secondary NCSAW dataset includes longitudinal data collected in five waves, which facilitates an investigation tracking the effects of out-of-home care over time. At the time of sampling, the entire NSCAW cohort consisted of 6,228 children (ages birth to 14), who previously interfaced with the child welfare system during a 15-month time period, beginning in October 1999. These data were collected longitudinally over the course of 2000–2007. This article presents cross-sectional results from Wave 1 only because the sample size decreases exponentially with complete data on the measures of interest across waves. Wave 1 data collection took place 2–6 months after the close of the investigation. Waves 2 (children not queried), 3, and 4 took place 12, 18, and 36 months after the investigation, respectively, and the wave 5 follow-up occurred in five staged cohorts 59–97 months after the initial investigation. This research utilizes data collected from the Child Protective Services (CPS) cohort of the dataset, consisting of 5,501 children who were investigated by Child Protective Services for abuse and neglect during the reference period (October 1999–December 2000). Based on the research questions posed in this study, the CPS sample was reduced in order to conduct analyses on children over six years old (N = 2,671). The sample was reduced even further due to the relatedness measure utilized, which only queried children 11 years and older and the

2066

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

Table 1 Descriptive statistics for variables. Variable

% or M

SD

Min

Max

% Problem behavior scores Behavior scale total (M) Maltreatment type % Physically abused % Neglected Caseworker listens Never Not very often Some of the time All of the time % Female Race % White % Black % Other race Age

0.29 46.5

0.45 27.7

0 0

1.00 148

0.51 0.49

0.50 0.50

0 0

1.00 1.00

0.08 0.06 0.27 0.59 0.53

0.26 0.24 0.44 0.49 0.50

0 0 0 0 0

1.00 1.00 1.00 1.00 1.00

0.49 0.33 0.18 12.9

0.50 0.47 0.39 1.29

0 0 0 11

1.00 1.00 1.00 15

Note. Unweighted; % problem behavior scores = above cutoff score >60; Behavior Scale Total = full continuous scale; M = mean; SD = standard deviation; Max = maximum; Min = minimum.

CBCL measure, resulting in a sample size of 669. With complete data for all pertinent items assessed at wave one including information about maltreatment type, the final sample size is 335 (due to missing values and appropriate skips). Measures The computer-assisted (CAI) method of interviewing was used, allowing for complex routings in questionnaires and increasing the level of accuracy in the administration of the surveys. As noted above, the instrument used to query children was administered in waves 1, 3, 4 and 5. The moderator variable used for this research was selected from the Satisfaction with Caseworker Services module of the child instrument, designed to measure degrees of caseworker satisfaction. Children who had regular contact with caseworkers between the ages of 11–17 were queried about whether they felt their caseworkers listened to them. This is the primary predictor variable assessed in this study in an effort to be parsimonious. Descriptive and Exploratory Analysis Descriptive statistics were assessed, inclusive of final sample means, frequencies, and appropriate assumption adherence. Analysis weights were applied due to the variation in selection probabilities and in order to arrive at population representative estimates. The analyses conducted in this study utilized the two CPS weights, one based on a national level and a second used for the stratum-level. Applying these weights allowed for unbiased estimates of characteristics such as means, proportions, and regression coefficients. All of the included socio-demographic variables (age, gender, and race) and the predictor variable (maltreatment type) were assessed for significant differences. Significance was determined using chi-square and t tests. The following model was tested in this research: behavior scores dichotomized as the dependent variable at wave 1 = ˛ + ␤1maltreatment type + ␤2cwlistens + ␤3 maltreatment typeX cwlistens + ␤4gender + ␤5ethnicity + ␤6age. This model was analyzed with each type of maltreatment at each of the four response categories for the moderator variable assessing how well caseworker listened. Binary logistic regressions were conducted using maltreatment type to predict whether children scored above or below the cutoff indicating problematic behaviors. Additionally, interaction terms were defined as each type of maltreatment with each response category of how often children felt listened to by caseworkers. Binary logistic regression was implemented as a method of investigating relationships between variables because the dependent variable was dichotomized as a 0/1 variable, with a score of one indicating clinical levels of problematic behaviors. Using the logit/logistic syntax in STATA statistical software allowed for survey data estimation for a less restrictive and more easily interpreted logistic regression model (proportional odds), and allowed for an assessment of the probabilities of scoring above the cutoff for the dependent variable according to each type of maltreatment and each response category of the moderator variable (how often the caseworker listens). Behavior Score (Dependent Variable) The Child Behavior Checklist (CBCL) described above is assessed as the dependent variable. At wave one, the mean score on the total behavior scale is 46.5 with a range of one to 148 (Table 1). The total combined (internalizing and externalizing) behavior scale was not assessed as a continuous variable in order to identify children who would be designated as exhibiting problem behaviors above the clinical cutoff level.

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

2067

Maltreatment Type (Predictor Variable) The primary predictor variable was recoded to indicate two categories of maltreatment: physical abuse and neglect (combined provisional and supervisory). As noted above, these two categories of neglect were combined because of ease of interpretation. Theoretically assessing the value of listening between supervisory and physical neglect does not lend well to meaningful interpretation. Caseworker Listens (Moderator Variable) The response categories for item below is ‘never’, ‘not very often’, ‘some of the time’, ‘and all of the time.’ 1. How much do you feel your caseworker listened to you? At this wave of data collection, the relatedness item measuring whether children felt listened to with complete cases generated 669 cases. Covariates The covariate variables are three demographic variables: gender (male/female), age (children over 11), and racial background (White, Black, Other). Results Binary logistic (logit) regression analyses for the adjusted full model generated model significance in each assessment of maltreatment type as a predictor with all four responses of caseworker listens as a moderator (eight models not presented in tabular form) (p < 0.001). Prior to the moderator analyses, each type of maltreatment and the variable indicating satisfaction with caseworker listening was investigated to determine separate predictability of behavior scores. The analysis of maltreatment type as the sole predictor of behavior scores indicated neither maltreatment type predicts behavior scores alone. However, the caseworker listens variable is marginally predictive of behavior scores (aOR = .77, p = 0.05) separately. Descriptive Results The sample demographic differences are shown in Table 1. The final sample consisted of 335 cases with complete data on all measures. The final un-weighted sample results indicate 29% of the children had problem behaviors (above cutoff) and 71% had no problem behaviors; the mean score on the caseworker listen variable (1–4) was 3.36; 49% of the final sample were physically abused and 51% were neglected. Only 8% of the children indicated they were never listened to by caseworkers, with the majority of respondents stating they were listened to some of the time (27%) or all of the time (59%). Fifty-three percent of the sample consisted of girls, nearly half identify as White and the average age of the respondents was 12.9. Based on pairwise correlation, there is a significant negative correlation between caseworker listens and problem behaviors (−0.15, p = .00); significant positive correlation between girls and problem behaviors (0.10, p = .00); a negative correlation between age and listening (−0.13, p = .00); and a positive correlation between age and behavioral scores (0.09, p = .00). Black children have a significant positive correlation with neglect (0.04, p = .00). How Much Do You Feel the Caseworkers Listened to You? As noted above, the item questioning children about being listened to by their caseworkers (never; not very often; some of the time; and all of the time) consisted of 669 cases with complete data. Given this variable is assessed at each response level, correlates of each response category are presented in Table 2. Forty-six percent of the children who responded that they were never listened to scored above the cutoff for problematic behaviors (p < 0.05) compared to 24% of those children who felt listened to all of the time. Although not a significant difference, more neglected children (52%) responded that they never felt listened and fewer neglected children felt listened to all of the time (48%). Significant differences exist with age, as slightly younger children (12.7, p < 0.05) felt listened to all of the time compared to the other response categories. Problem Behaviors Correlates of the dependent variable, dichotomized behavior scores generated significant differences among the moderator, caseworker listens, gender, and age (Table 3). Among children who did not exhibit problem behaviors, listening had a larger effect, with 66% expressing they felt listened to all of the time compared to 5.8% who never felt listened to. Among those who were above the cutoff for problem behaviors, there are conflicting results in that the impact of listening well does not demonstrate an expected pattern across response categories. Perhaps the positive benefits of listening well past the cutoff for problem behaviors are diminished due to other endogenous reasons specific to the child. Sixty-six percent of

2068

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

Table 2 Correlates of caseworker listens response categories (N = 669). How much do you think your caseworker listened to you?

Variables

Never (n = 50) % or M

Behavior score* % Problematic >60 Maltreatment type % Physical abuse % Neglect % Female Race % White % Black % Other Age*

Not very often (n = 39) % or M

Some of the time (n = 174) % or M

All of the time (n = 406) % or M

46

31

40

24

48 52 62

57 43 62

40 60 63

52 48 56

50 30 20 13.1

49 28 23 13.1

47 35 18 13.1

51 30 19 12.7

Note. Percentages and means refer to those above the problematic behavior score for each response category. * p < .05. Table 3 Correlates of problematic behaviors. Analysis group (n = 335)

Variables

Non-problematic behaviors (n = 230) % or M

Problematic behaviors (n = 96) % or M

44.4 55.6

48.1 51.9

5.8 5.8 22.5 66.0 54.9

11.4 5.9 34.2 48.5 65.7

51.1 32.5 16.4 12.7

54.4 28.5 17.1 12.9

Maltreatment type Physical abuse Neglect Caseworker listens* Never Not very often Some of the time All of the time % Female* Race % White % Black % Other race Age*

Note. Unweighted; non-problematic behaviors = score <61; problem behaviors = score >60 * p < .05.

the children with problematic behaviors were girls, and slight significant differences were detected among age with older children classifying above the cutoff (12.7 vs 12.9). Physical Abuse versus Neglect According to Caseworker Listening As noted above, wave one data were collected approximately two to six months following the close of the child abuse and neglect investigation. At this stage in their experiences with caseworkers, the majority of children (86%) responded that they felt listened to some of the time or all of the time. Probabilities of Listening Response Categories Based on the analyses from the adjusted full model with caseworker listens as a moderator predicting behavior scores according to maltreatment type, the probabilities of being classified in each response category of the caseworker listens variable are presented comparing physical abuse to neglect. Based on an assessment of the logistic output computing odds ratios, Table 4 presents the odds coefficients of behavior problems (intercept) converted to probabilities. All of the behavior score probability differences between maltreatment types were significant. The probability of physically abused children scoring above the cutoff for problematic behaviors is .66 if they never felt listened to compared to .19 for those children neglected and reporting never feeling listened to. For children who have been physically abused, feeling listened to by caseworkers has a buffering effect on behavior scores, as indicated by the large difference in probabilities of higher behavior scores at the lowest level of listening compared to children who have been neglected. Larger differences between maltreatment types occurred at the lowest level of listening and evened out across listening satisfaction responses, indicating more listening results in fewer differences between maltreatment types. In all response categories, except the highest level of listening (.21 vs .24), physically abused children have a higher probability of problematic behaviors than neglected children. For those

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

2069

Table 4 Probabilities of listening as a moderator predicting behavior scores according to maltreatment type. Caseworker listens Physical abuse Neglect

Never Prob (odds)

Not very often Prob (odds)

Some of the time Prob (odds)

All of the time Prob (odds)

.66 (1.97)* .19 (.24)*

.50 (1.01)* .21 (.26)*

.34 (.52)* .22 (.28)*

.21 (.27)* .24 (.31)*

Note. Unweighted; probabilities: odds/odds + 1 (probability of a score of 1 (least listened to) in each cell). * p < .05.

Table 5 Binary logistic regression (adjusted) full model with interaction. Variables

Odds ratio

Physical abuse (ref. neglect) Never Not very often Some of the time All of the time Male Black Other Age

8.07** 3.83* 1.82 .87 .32* .54 1.12 1.38*

95% confidence interval 1.96 1.47 .92 .39 .14 .25 .40 1.03

33.19 9.99 3.59 1.92 .73 1.15 3.15 1.86

Note. Reference groups for feeling listened to = 0 at each response category. * p < .05. ** p < .01.

children physically abused, more listening resulted in lower probabilities of having more problematic behaviors. However, the opposite occurred for neglected children compared to those physically abused. Counter intuitively, the probability of neglected children scoring above the cutoff for high behaviors increases slightly the more they feel more listened to. Binary Logistic Regression (Adjusted) Full Model with Interaction The interpretation of odds ratios with categorical by categorical interactions can present challenges in interpretation. The results annotated in Table 5 indicate the effect of maltreatment type on behavior scores according to the reference value of the other interaction term (i.e. never = 0). In other words, these results indicate the effect of being physically abused according to listening satisfaction responses on behavior scores. Physically abused children (aOR = 8.07, p = 0.00) are eight times more likely to exhibit problem behaviors when they responded they were never listened to. These effects decrease as children feel more listened to, however feeling listened to some or all of the time does not significantly moderate the relationship between physically abused children and higher behavior scores. Prior to the analyses, each of the covariates was centered at the mean based on the final sample (n = 335). Results indicate covariate significance among gender and age. Boys are significantly (aOR = 0.32, p = 0.01) less likely to distribute problematic behaviors than girls and older children (aOR =1.38, p = 0.03) are more likely to score above the cutoff for problem behaviors. Discussion There is an abundant amount of research focusing on the experiences of children served by the child welfare system, but very few incorporate an assessment from the perspective of the children served. Children’s perceptions regarding the care they receive are crucial to understanding treatment outcomes and evaluating the services provided. There is utility in assessing the differences in experiences with caseworkers according to the type of maltreatment children have endured. Increasingly, researchers are investigating the differences between types of maltreatment as it relates to child, parental, and environmental factors. This research contributes an account of differences between physical abuse and neglect (both failure to provide and supervise) in behavioral outcomes according to levels of satisfaction with caseworker listening. These results provide insight into correlates associated with children’s responses to the treatment process (i.e. relationship with caseworker) based on maltreatment type. Moreover, an inquiry such as this addresses ways in which children feel supported and how levels of perceived support impact outcomes. Findings in this research indicate physically abused children who responded that they have never been listened to are much more likely to score above the cutoff for problematic behaviors than neglected children. Thus, caseworkers should be cognizant of how well physically abused children feel heard as it relates to their behavioral outcomes. However, the more children feel listened to, the differences between maltreatment types decrease and reverse at the highest level of listening satisfaction. Intuitively, one would expect children to display fewer problem behaviors the more supported or listened to they feel. These findings suggest the perceived quality of relationships

2070

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

with caseworkers can have an impact on outcomes in different ways according to the type of maltreatment children have experienced. The finding that girls are positively correlated with problematic behaviors as opposed to boys warrants further inquiry. Gender is a standard demographic factor assessed descriptively, however, painting a picture about gender differences does not allow one to meaningfully interpret differences. An account of the kinds of problematic behaviors displayed by girls compared to boys would lend to a clearer understanding of these findings. Based on other significant correlations, older children seem to feel less supported in that they indicated they felt listened to less often by caseworkers. This has implications for differences in how caseworkers develop relationships with children according to their developmental stage. More research is needed to understand the relational dynamics between children and their caseworkers according to their developmental age. These analyses were conducted with children over the age of eleven, which is the beginning of a volatile and confusing time in a child’s life. Perhaps, the caseworker/child relationship should be built with consideration of the child’s age and their developmental capacity. These analyses were limited in that the research question and applicable measures were confined to the parameters of the dataset. The items chosen required subjective responses, which limits reliability, however, the NSCAW dataset consists of extremely high-quality data measurements. Some items in the dataset were excluded because of data incompletion, which hindered the sample size. Many constructs and possible indicators from NSCAW were omitted from this research to focus the scope of this research on the one item measuring how often children felt listened to by caseworkers. The measure was not given to all age groups because of cognitive and developmental capacities to respond. This study focused on those children who were able to answer the questions regarding their relationship with their caseworkers. Given this research is crosssectional and conducted only at the first wave of data collection, an assessment of caseworker changes over time was not possible, thus preventing one to comment on a true developing relationship over time. Further, there is debate about how to measure the concept of relatedness, however, we posit assessing maltreated children’s perceptions regarding being listened to, is suited to measure relatedness between children and their caseworkers. Notwithstanding these study limitations, this research proves of great import to our understanding as scholars and practitioners regarding the significance of a supportive caseworker/child relationship as an influential factor in behavioral outcomes as they differ according to maltreatment type. The strength and nature of the network between children and caseworkers is considered a valuable resource, requiring mutual trust. Listening well is the first step in developing a trusting relationship. Further, listening well to children influences behaviors differently according to maltreatment type. In an effort to support children who have experienced different forms of maltreatment, caseworkers should remain aware of how supported the children feel while receiving services. Moreover, caseworkers should be cognizant of the importance of listening well while interacting with these children. Conducting this kind of assessment is poised to contribute usefully to many working within the child welfare research arena, inclusive of policymakers and parents, but especially for child welfare practitioners. Perhaps, future research and service evaluation tools should focus more attention on the relationship between the caseworker/child relationship and behavioral and developmental outcomes. Rarely do we include as a mandate an assessment from the child’s perspective, yet we are required to do due diligence in evaluating our services. It is very important to assess these relational experiences and any associations with behavior in an effort to inform treatment planning. Acknowledgement We would like to thank Jim Jaccard for his statistical consultation. References Achenbach, T. (1991). Manual for the child behavior checklist/4–18 and 1991 profile. Burlington, VT: University of Vermont Department of Psychiatry. Berrick, J. D., Frasch, K., & Fox, A. (2000). Assessing children’s experiences of out-of-home care: Methodology challenges and opportunities. Social Work Research, 24(2), 119–133. Chapman, M., & Christ, S. (2008). Attitudes toward out-of-home care over 18 months: Changing perceptions of youths in foster care. Social Work Research, 32(3), 135–145. Chapman, M., Wall, A., & Barth, R. P. (2004). Children’s voices: The perceptions of children in foster care. American Journal of Orthopsychiatry, 74(3), 293–304. Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, 94–120. Coleman, J. (1990). Foundations of social theory. Cambridge, MA: Harvard University Press. Colton, M. (1989). Foster and residential children’s perceptions of their social environments. British Journal of Social Work, 19, 217–233. Fanshel, D., Finch, S. J., & Grundy, J. F. (1990). Foster children in a life course perspective. New York: Columbia University Press. Festinger, T. (1983). No one ever asked us: A postscript to foster care. New York: Columbia University Press. Fox, A., Frasch, K., & Berrick, J. D. (2000). Listening to children in foster care: An empirically based curriculum. Berkeley, CA: Child Welfare Research Center. Goerge, R., Wulczyn, F., & Harden, A. (1996). New comparative insights into states and. their foster children. Public Welfare, 54(3), 12–25. Johnson, P., Yoken, C., & Voss, R. (1990). Family foster care placement: The child’s perspective. Chicago: The Chapin Hall Center for Children at the University of Chicago. Kufeldt, K. (1984). Listening to children—Who cares? British Journal of Social Work, 14(3), 257–264. Lochner, K., Kawachi, I., & Kennedy, B. P. (1999). Social capital: A guide to its measurement. Health and Place, 5, 259–270. Lynch, M., & Cicchetti, D. (1991). Patterns of relatedness in maltreated and nonmaltreated children: Connections among multiple representation models. Development and Psychopathology, 3(2), 207–226. Merritt, D. H. (2008). Placement preferences among children living in foster or kinship care: A cluster analysis. Children and Youth Services Review, 30, 1336–1344. Merritt, D. H., & Franke, T. M. (2010). Should I stay or should I go? Children’s placement preferences: Longitudinally. Journal of Social Service Research, 36(1), 46–67.

D.H. Merritt, S.M. Snyder / Child Abuse & Neglect 38 (2014) 2062–2071

2071

Needell, B., Webster, D., Curraco-Alamin, S., & Armijo, M. (1998). Performance indicators for child welfare services in California: 1997. Berkeley, CA: Child Welfare Research Center, U.C. Berkeley. Triseliotis, J. (1984). Identity and security in adoption and long term fostering. Early Childhood Development and Care, 15, 149–170. Triseliotis, J. (2002). Long-term foster care or adoption? The evidence examined. Child & Family Social Work, 7(1), 20–23. Whiting, J. B., & Lee, R. E. (2003). Voices from the system: A qualitative study of foster children’s stories. Family Relations, 52, 288–295. Wilson, L. (1994). The 1994 annual client evaluation (unpublished report). Tulsa, OK: Wilson Resources Inc. Wilson, L. (1996). The 1996 annual client evaluation (unpublished report). Tulsa, OK: Wilson Resources Inc. Zimmerman, R. (1982). Foster care in retrospect. New Orleans: Tulane University.