Child’s Religiosity,Ethnic Origin,and Gender: A Randomized Experimental Examination of Risk Assessment and Placement Decisions in Cases of Ambiguous Risk to Children From Low SES Families

Abstract

Social workers are called upon as expert witnesses to assess risk to children and recommend interventions. It is hypothesized that risk assessments and decisions are influenced by the risk-to-child as well as personal attitudes and biases. We examined the role of the ethnic origin, religiosity, and child gender on assessments and recommendation.

Methods:

The current study used a survey-based balanced crossover experimental design. One hundred and twenty Israeli social workers responded to case descriptions of ambiguous risk cases of children from low socioeconomic status families. Design was based on eight vignettes. Manipulated factors included child’s gender, family religiosity, and ethnic origin. Participants assessed the child risk level and reported placement recommendation.

Results:

Male, ultraorthodox-Jewish child of Mizrahi origin had higher likelihood for risk assessment. Risk assessment predicted placement recommendation.

Implications:

Social workers tend to perceive child’s male gender, ultrareligiosity, and Mizrahi origin as risk factors. Such cultural biases indirectly affected the placement recommendation. Awareness of social workers regarding these potential biases may reduce the influence of heuristics on the professional decision-making process.

Keywords

child placement risk assessment psychometric study biases

The family is one of the domains in which children’s well-being and the protection of their rights are expected to be most prominent. As the family home is the natural setting for childrearing and serves as an existential anchor, it is generally run according to the wishes of the family, without intervention by the society via the law courts or other means (Zafran, 2010). However, with the creation of a significant risk to the child or circumstances of suspected risk, such as deficiencies in their environment that prevent their intact development, neglect, or situations that put their well-being, or even their lives, at risk (Zeira, Attar-Schwartz, & Benbenishty, 2012), the state is required to intervene (Legal Capacity and Guardianship Law, 1962). The challenge of intervention, according to the accepted approach in most countries (Benbenishty et al., 2015; Gilbert, Parton, & Skivenes, 2011; Spratt et al., 2014), is assigned to welfare service social workers, who must cope with making crucial decisions that have a significant impact on children and their families (Enosh & Bayer-Topilsky, 2014; Taylor, 2010, 2016).

The decision of whether to recommend out-of-home placement or to leave children at home with the family, despite the concern for their well-being, draws on the principle of “best interests of the child” as a primary consideration, which establishes their overall rights and needs (Silman Committee, 2014). Nonetheless, one criticism undermines the use of this interpretive framework because of the problems involved in defining the child’s best interests. According to this view, the “best interests of the child” principle is subject to manipulation, exploitation, and ethical bias according to the decision maker’s worldview and does not faithfully serve the needs of the child (Rotlevy Committee, 2003). Another criticism raises a doubt regarding the possibility of forming a professional opinion based on this principle under the complex conditions in which decisions are made. These decisions are often made under time pressure because of the potential risk to vulnerable children (Shlonsky & Wagner, 2005) without all the necessary information available (Benbenishty, Osmo, & Gold, 2003; Munro, 2008). They are also subject to the constant demand to balance conflicts of interest between the parents’ right to privacy and autonomy and the responsibility to protect their children (Gambrill, 2005).

Moreover, as has been proved by previous studies, a substantial number of these decisions fall into the “gray area” category (Davidson-Arad, 2010), in which the children’s physical well-being is not at immediate risk, but a potential risk is posed to their psychological well-being, which is more difficult to define and establish (Ben-Arieh & Haj-Yahia, 2006). In such situations, the most appropriate decision might elude even the most experienced social workers (Benbenishty et al., 2015). In fact, this decision, which must be based on rational consideration, without prejudice, and taking into account the implications of all its outcomes for the children (Musen & van Bemmel, 1997), is likely to be influenced by the values, attitudes, and personal characteristics of the practitioner, child in question, and overall social structure.

The Dual Process of Judgment

Judgment processes under ambiguous conditions is a well-known issue in the welfare and child protection fields (Banach, 1998; Fluke, Chabot, Fallon, MacLaurin, & Blackstock, 2010). Dual process theory offers an explanation of how professional decisions are made in ambiguous situations (Croskerry, 2009; Evans, 2008). According to dual process theory, judgment is a dual process involving two cognitive systems: heuristic intuitive (an automatic process based on cognitive rules of thumb) and analytic rational (a process that relies on clear rules based on professional knowledge). Judgment occurs simultaneously in both these cognitive systems. In cases of clear-cut criteria, for example, in the presence of unequivocal evidence of a parent’s physical abuse, the social worker will draw on mainly analytical methods (analytical rational) to adjust the criteria for the case in question. However, in cases of great ambiguity regarding the facts or applicable criteria, the social worker will be inclined to use heuristic (intuitive) methods to reach a decision. As a result, decisions in cases of great uncertainty might be sensitive to heuristics and personal biases such as cultural and professional differences or organizational culture and might lead to bias in professional considerations (e.g., Enosh & Bayer-Topilsky, 2014; Enosh, Nouman, & Anabtawi, 2016; Nouman, Enosh, & Niselbaum-Atzur, 2016).

Although rational decision-making is considered to be the ultimate standard (Tversky & Kahneman, 1975), it was criticized by other researchers in the field, claiming that human rationality is always bounded and that natural decision-making by professionals is based on fast and frugal heuristics that have been proved efficient through accumulated experience (Gigerenzer, Hertwig, & Pachur, 2011). On the other hand, given the lack of absolute and unequivocal outcome criteria on which to base their decision and given their accumulated experience, professionals might develop biased, rather than efficient, heuristics (e.g., Dhami, 2003; Dhami & Ayton, 2001).

Many heuristic factors are mentioned in the professional literature as biasing the social worker’s professional consideration when making decisions in highly ambiguous situations. Among these factors are the personal characteristics of the child in question (Drake et al., 2011; Fallon et al., 2013) as well as the social worker’s personal (Brunnberg & Pe’cnik, 2007; Regehr, LeBlanc, Shlonsky, & Bogo, 2010), professional (Lulu, 2000; Mattison, 2000), and sociocultural characteristics (Enosh et al., 2016; Hacker & Frenkel, 2005; D. J. Hall & Paradice, 2007). Nevertheless, it is doubtful as to whether these factors alone present the breadth of the issue. The aim of this article is to focus on three additional factors, which have received inadequate attention to date in the discussions of this topic, even though they are important components of the judgment processes in ambiguous conditions in the case at hand. They are the degree of religiosity of the child’s family, the family’s ethnic origin, and the child’s gender. These factors were examined, in the present study, in ambiguous cases regarding facts or criteria that need to be applied among low socioeconomic status (SES) families. The importance of this empirical examination is reinforced in light of study findings that indicate that risk assessments and placement decisions are biased against minority and disadvantaged populations (Benbenishty & Chen, 2003; Miller, Cahn, & Orellana, 2012; Morton, Ocasio, & Simme, 2011).

The Heuristics of Religiosity, Ethnic Origin, and Gender

The use of various heuristic factors in the decision-making process under complex and ambiguous conditions emphasizes the essential diversity among social workers. It cannot be ignored that social workers are not working in a vacuum but carry a personal, cultural, religious, and professional load, which might directly and indirectly influence the discussion about the decision to be made about the fate of the children and their families (Slonim-Nevo & Lander, 2004) under the guise of implementing professional principles (Slonim-Nevo & Lander, 2004). Is it possible that the degree of religiosity and ethnic origin of the family of the child in question influences the child’s risk assessment? Moreover, does the fact that the family has chosen to lead a certain type of religious lifestyle or comes from a particular ethnic origin influences the placement decision? These questions illustrate the complexity of heuristic bias or what can be termed as “stereotypical attitudes” involved in decision-making regarding children at risk.

In the case under discussion, stereotypical attitudes refer to attitudes held by social workers toward certain population groups (distinguished, e.g., by gender, race, profession, or ethnic origin); in other words, they believe that people belonging to a specific group share certain characteristics (Tur-Kaspa Shimoni & Schwarzwald, 2003). The stereotypical discourse appears to be widespread in Israel in light of the society’s division into national, ethnic, religious, and socioeconomic groups, where each group is struggling, separately, to improve its well-being and to promote its status in Israeli society (Doron, 2004). One prominent factor that is the focus of division within Jewish–Israeli society is religiosity, especially regarding the division between secular and ultraorthodox (Jarbi & Levi, 2000). To be more precise, one social factor dividing the Jewish population in Israel is the issue of religious adherence. The Jewish population may be depicted as existing on a range of adherence to religious observances, ranging from total secularization and even antireligiousness, on the one hand, to extreme ultrareligious observance, on the other. A common way of dividing the population consists of four major groups: secular (Hiloni), traditional (Masortyi), national orthodox (Dati-Leumi), and ultraorthodox (Haredi). The four groups may be considered as existing on a rank-ordered continuum in terms of their adherence to the observance of the religious commandments.

Unlike the other three groups, the ultraorthodox Jewish population is a distinct cultural and economic sector in the state of Israel, especially due to their tendency for separatism (Rechnitzer, 2013). This distinction gives rise to many social and political conflicts. For example, one of the prominent cultural and social divides is their avoidance of military service. Based on their religious beliefs, most of ultraorthodox are exempt from the mandatory military service that all other Jewish groups in Israel are subject to. This exemption is a source of disdain many secular and traditional Jews feel toward the ultraorthodox. Furthermore, most ultraorthodox belong to low SES and segregate themselves from the rest of Israeli society in economic behavior and life patterns (Doron, 2004). Stereotypical perceptions of the ultraorthodox population, such as “greedy” and “primitive,” are conspicuous in local media channels (Zarembski, 2002). Repeated reference to such images contributes to people’s development of a self-identity with stereotypical attitudes, perpetuating the perception of “us and them” (First, 1997; S. Hall, 1997).

On the other hand, it should also be noted that although most national orthodox are not isolated as the ultraorthodox from the general social milieu in Israel, still many national orthodox tend to minimize the differences between themselves and the ultraorthodox, claiming that the differences are minor and are related to politics of isolation rather than to religious differences. Thus, national orthodox Jews tend to clump the two groups together as one general group of “adherents” (Datyim) or “orthodox.”

For the current research, we had two purposes. The first was to determine the effect that belonging to a separatist group as the ultraorthodox had on the way their children were perceived by social workers. Thus, in order to achieve this goal and highlight the possible inherent biases against ultraorthodox, we described the children in question as either secular or ultraorthodox, depicting each time one pole of the social–religious range. The second goal was to determine the group belongingness of the respondents per their religious observance. When asking the respondents regarding their social–religious identity, we asked whether they were identifying as belonging to any of the four major groups: secular (Hiloni), traditional (Masorti), national orthodox (Dati), or ultraorthodox (Haredi). We also gave a blank option of “other” for those who preferred to identify otherwise (none has chosen this blank option).

Another focus of division is ethnic origin, differentiating persons and families of Ashkenazi and Mizrahi origin (roughly, those originating from European countries and those who came from Middle Eastern countries; Doron, 2004). Even though no completely homogenous collective of either Ashkenazim or Mizrahim exists, this distinction has turned into a significant factor in the social structure in Israel and has become very important in political discourse. The differences are mainly in the definition of identity but can be observed also in the sociocultural sphere, in differences in status between Ashkenazim—who have largely achieved high income, advanced education, and prestige—and Mizrahim, who are associated more with manual labor and lower SES (Haberfeld, Semyonov, & Cohen, 2000). In the context of social workers, previous studies have shown that their decisions pertaining to at-risk children might be biased by stereotypical attitudes regarding race or ethnic origin (Benbenishty & Chen, 2003; Miller et al., 2012; Morton et al., 2011), and in Israel, toward children of Mizrahi origin. In a previous study (Enosh & Bayer-Topilsky, 2014), using hypothetical case descriptions, twice as many Mizrahi than Ashkenazi children were recommended by social workers for out-of-home placement. These findings imply that these differences are derived from the society’s overt or covert bias toward minority groups. Social workers might be influenced by such bias and might even use heuristics in their decision-making regarding children at risk.

Regarding the child’s gender, the question arises of whether the gender context is relevant to examining the best interests of the child. Is boys’ level of risk assessed as higher than that of girls, leading to higher rates of out-of-home placement recommendation for boys? The research literature indicates a gender difference, in that boys are at greater risk than girls of being included in welfare service settings, mainly child protection (Shimoni & Benbenishty, 2011). To the best of our knowledge, the influence of the child’s gender on decisions regarding the level of risk and out-of-home placement has not been empirically examined to date.

According to the review thus far, social workers’ considerations in the decision-making process regarding children at risk require theoretical and practical considerations in order to protect the child’s best interests. However, these considerations, as elaborated above, might be biased due to the influence of cultural and social biases factors. While previous studies have shown, again and again, that the family’s income is the best predictor of whether the child will be removed from home or will remain at home with supporting services from the community (e.g., Britner & Mossler, 2002; Enosh & Bayer-Topilsky, 2014; Harris & Hackett, 2008; Miller et al., 2012), our goal in the current study was to neutralize this factor, while examining the role of other factors. In order to achieve that, we use case description, in which all families are described as coming from a low SES. Therefore, the focus of the present study was the examination of three potentially biasing factors in the process of risk assessment and out-of-home placement recommendation: degree of religiosity of the child’s family, the family’s ethnic origin, and the child’s gender. The study hypotheses were as follows:

Hypothesis 1: Social workers’ risk assessments of children will be influenced by the characteristics of the child’s family and the child’s gender: In cases where the family is ultraorthodox and/or of Mizrahi origin and the child is male, the risk will be assessed as higher, as opposed to cases in which the family is secular and/or of Ashkenazi origin and the child is female.

Hypothesis 2: Social workers’ recommendations of out-of-home placement will be influenced by the characteristics of the child and the child’s family: When the family is ultraorthodox and/or of Mizrahi origin and the child is male, recommendation of out-of-home placement will be more likely than when the family is secular and/or of Ashkenazi origin and the child is female.

Hypothesis 3: Social workers’ recommendations of out-of-home placement will be influenced by the child’s risk assessment: In cases of higher risk assessment, social workers will recommend out-of-home placement more frequently than in cases of lower risk assessment.

Method

Sample and Sampling Procedures

A purposive convenience sample was used (Patton, 2011) of social workers from social service departments in Israel, working with families and with children at risk, who were involved in processes of decisions regarding children’s level of risk. The questionnaires were administered to the social workers personally by one of the researchers, alongside an introductory letter explaining the essence and aims of the study. Overall, 200 social workers were approached, of whom 120 responded (60% response rate). Of those, most were female (82%) and married (84%). Average age was 39 (SD = 10.15) and average professional experience was 11 years (SD = 8.58). In terms of education, 16% were at the BSW level, and 84% had achieved their MSW degree or higher. Regarding social workers’ level of religiosity, 33% defined themselves as secular and 67% defined themselves as religious. This distribution of personal and professional characteristics is similar to the distribution of personal and professional characteristics of social workers in other surveys of social workers working in the field of children at risk, regardless of religious affinity or ethnicity (e.g., Enosh et al., 2016; Nouman, Enosh, & Jarjoura, 2018; Nouman et al., 2016). Table 1 presents the descriptive statistics; Table 2 presents the correlation matrix among those variables. The questionnaire did not include questions regarding the respondent’s within-Jewish ethnic affiliation (Ashkenazi/Mizrahi), as such questions are uncommon in public surveys in Israel, and could have alerted the respondents to the inherent manipulation of the presented vignettes. On the other hand, data such as family status, level of religiosity, and number of children are quite common and were deemed relevant to the current research questions, so were included in the questionnaire.

Table 1.

Sample Characteristics.

Characteristic	%/Mean (±SD)
Female	82%
Married	84%
Age	39 (±10.15)
Children	3.02 (±2.19)
Years of experience	10.7 (±8.58)
Education
BSW degree	16%
MSW degree	84%
Religiosity of the social work
Secular	33%
Religious	67%

Note. N = 120.

Table 2.

Correlation Matrix Among Respondent’s Characteristics, Manipulated Variables, Risk Assessments, and Removal Recommendation.

Variables	1	2	3	4	5	6	7
1. SW religiosity	1
2. SW education	−.11***	1
3. SW immigrant	−.1***	.19***	1
4. Child female	.0	.0	.0	1
5. Child ultraorthodox	.0	.0	.0	.0	1
6. Child Mizrahi	.0	.0	.0	.0	.0	1
7. Estimated risk	.0	.07*	−.04	−.08**	.04	.08**	1
8. Removal recommendation	.07*	.01	.0	−.08*	.04	.04	.46***

*p < .05. **p < .01. ***p < .001.

Instruments

Two-part questionnaire was distributed to all the study participants a personal–professional background questionnaire and a set of vignettes to be assessed, manipulated within a factorial balanced crossover-experimental design.

Background questionnaire

The first part was a sociodemographic background questionnaire including issues such as family status (married = 1/not married = 0), age, gender (male = 0/female = 1], number of children, religiosity (secular = 0, religious = 1], education (BSW = 0/MSW = 1], and years of experience.

Factorial design using vignettes (case descriptions)

Each participant received eight detailed vignettes (case descriptions). Since the study was meant to assess the two-stage process inherent in risk assessment—first the assessment of risk to the child by the social worker and second constructing a recommendation regarding placement issues—therefore, each of the vignettes was followed by two separate questions regarding the specific case: (i) “How would you estimate the level of risk to the child in the case presented? (Likert-like scale from 0 to 5; 0 = no risk at all and 5 = very high risk level)” and (ii) “Would you recommend out of home placement for this child?” (yes = 1/no = 0).

The vignettes were derived from real-life case descriptions obtained from social services records. The manipulated variables in the 2 ×2 × 2 experimental design included (1) religiosity of parents (secular = 0/ultraorthodox = 1], ethnic origin of parents (Ashkenazi = 0/Mizrahi = 1), and child’s gender (male = 0/female = 1). To neutralize the possible effects of the order in which the cases appeared, we systematically shuffled all eight cases by constructing four variously arranged versions. Those four versions were distributed randomly to the respondents. Based on previous research that indicated that ambiguous cases are most common in the field of children at risk, and judgments taken in the context of ambiguity as to level of risk are prone to be affected by personal attitudes and biases, all cases were ambiguous as to level of risk to the child (Enosh & Bayer-Topilsky, 2014). As previous research indicated a significant difference in risk assessment when the vignettes presented middle-class families as compared to poor families, in order to maintain invariance, all families in the current study’s vignettes were presented as poor families. We validated the case descriptions using interrater reliability in a previous study that used the same instrument (Enosh & Bayer-Topilsky, 2014): A preceding questionnaire was presented to six experts from the field of child welfare. Interrater agreement among the four expert judges was .9 (for procedures, see Ebel, 1951; McLennan, 1993). The validation was carried out by having the professional experts judge the cases without the manipulated variables and verifying that without mention of those issues, the responses received were in high agreement (interrater reliability) among the judges. Thus, we conclude that the variation in the final results can be attributed to the experimental manipulation.

Analysis

The social workers’ decisions regarding risk assessment and placement decisions were examined in the study, as mentioned above, through the eight cases presented to them. The results were analyzed at the case level rather than respondent level. Overall, 120 social workers evaluated 960 cases. Since all the analyses were conducted at the case level, to control for possible clustering at the respondent level, we used robust generalized estimating equations (GEEs) logistic regressions (Rogers, 1993; Williams, 2000) in STATA-14 (StataCorp, 2015). GEE is similar to hierarchical linear modeling in certain aspects, but rather than allowing for random nesting effects, it constrains for population averaged fixed effects. In that it allows for control over clustering effects in nested designs as this one is (cases nested within respondents). All models controlled for the respondents’ personal and professional characteristics, including family status, gender, number of children, religiosity, education, and years of experience (age was omitted from analysis for multicolinearity with years of experience).

Ethical Considerations

The participants were informed about the goals of the study, which were defined as examining attitudes of social workers faced with decisions regarding risk assessment and placement decisions in disputed cases. They were also informed that participation was anonymous, and that by filling out the questionnaire, they were consenting to participate in the study. The University of Haifa committee for ethical research with human beings approved the study.

Findings

The influence of the following variables (1) family’s religiosity, (2) ethnic origin of the family, (3) child’s gender, and (4) social workers’ personal and professional characteristics, on the social workers' decison making, as experessed in their risk assessments and out-of-home placement recommendations, was examined usign regresson analyses. Specifically, to examine the influence of the manipulated factors on predicting the risk assessment and out-of-home placement recommendation, we performed two regression models. Since the study was meant to assess the two-stage process inherent in risk assessment—first, the assessment of risk to the child by the social worker and second, constructing a recommendation regarding placement issues—we estimated two separate models, one examining risk assessment and the second examining placement recommendation. Thus, the first model was a linear regression for predicting risk assessment. The second model examined a logistic regression predicting out-of-home placement recommendation. The findings are presented in Table 3. The models controlled for possible effects of personal and professional characteristics of the social worker.

Table 3.

Predicting Risk Assessments and Removal Recommendations—Linear Regression and Logistic Robust Regression Models.

Independent Variables	Risk Assessments	Removal Recommendation
Independent Variables	Robust Linear Regression (b)	Robust Logistic Regression (Odds Ratio)
Risk assessments	—	14.83***
Child’s factors
Ultraorthodox	0.09*	1.22
Ethnic origin (Mizrahi)	0.19***	1.06
Child’s gender (male)	0.18***	0.70
Wald-χ²(df)	32.92(6)***	91.94(7)***

Note. All models controlled for the worker’s education level, gender, religiosity, years of experience, family status, and age.

*p < .005. **p < .001.

The first hypothesis was that social workers’ risk assessments of children will be influenced by the characteristics of the child’s family and the child’s gender: In cases where the family is ultraorthodox and/or of Mizrahi origin and the child is male, the risk will be assessed as higher, as opposed to cases in which the family is secular and/or of Ashkenazi origin and the child is female.

As can be seen in Table 3, the regression model examining this hypothesis was found to be statistically significant, χ²(6) = 32.92, p < .01. The findings indicate that risk assessment is influenced by the child’s characteristics. More specifically, being of ultraorthodox family increased the assessed risk (b = .09, p < .005), being of Mizrahi origin increased the assessed risk (b = .19, p < .001), and being a male increased the assessed risk (b = .18, p < .001). Thus, in accordance with the hypothesis, a male, ultraorthodox child of Mizrahi origin has a greater chance of being assessed at a higher risk than a female child from a secular, Ashkenazi family.

The second hypothesis was that social workers’ recommendations of out-of-home placement will be influenced by the characteristics of the child and the child’s family: When the family is ultraorthodox and/or of Mizrahi origin and the child is male, recommendation of out-of-home placement will be more likely than when the family is secular and/or of Ashkenazi origin and the child is female. We examined this hypothesis using a GEE-based logistic regression with robust standard errors. The model was significant, χ²(7) = 91.94, p < .01). Accordingly, the findings show also that social workers’ recommendations for out-of-home placement are influenced by their estimation of risk to the child but are not directly affected by any of the manipulated variables (Child’s family being ultraorthodox, Mizrahi, or the child being a male). Hence, we may conclude that the impact of child’s characteristics on the decision to recommend out-of-home placement is mediated by the assessment of level of risk to the child. In other words, the child’s gender, ethnic origin, and religiosity level influence the child’s risk assessment, which in turn influences the social worker’s recommendation for out-of-home placement.

Discussion and Applications to Practice

Social workers have tremendous responsibility when making decisions pertaining to children at risk. They need to ascertain constantly that they have the appropriate conceptual tools and knowledge to perform risk assessment while protecting the child’s physical and psychological well-being. However, despite the demand for rational consideration, one cannot ignore the fact that decision-making is a complex, ambiguous reality that might be sensitive to the decision maker’s heuristics and personal biases. Few studies have investigated social workers’ attitudes toward disadvantaged and minority populations and the effects such attitudes may have on crucial decisions made by social workers. Most previous studies have focused mainly on the analysis of actual cases derived from welfare and social work files (Britner & Mossler, 2002; Fialkov, Cohen, & Grabe, 1990; Harris & Hackett, 2008; Miller et al., 2012). Indeed, previous research has demonstrated that minority children and children from disadvantaged families are overrepresented in out-of-home placements, including residential and foster care (Davidson-Arad, 2010; Harris & Hackett, 2008; Miller et al., 2012).

The aim of the present study was to examine heuristic biases that play a role in the decision-making process of social workers when performing children’s risk assessment and recommending out-of-home placement. We focused on three factors that might influence the decision: the level of religiosity and ethnic origin of the family of the child in question, and the child’s gender.

The study findings reveal that characteristics of the children and their families play an important role in the social workers’ decisions regarding the child’s risk assessment. The fact that a child is from a family of a minority ethnic origin—Mizrahi, in the case under study—will bias the assessment toward a higher level of risk than in the case of a child from a family of Ashkenazi origin. These findings are in keeping with the study hypothesis that stereotypical attitudes lead to biased decisions against families of Mizrahi origin and are in line with studies that indicate this bias in issues related to decision-making regarding children at risk. Thus, in cases of ambiguous risk, social workers tend to attribute a higher level of risk to Mizrahi children than to Ashkenazi children (Enosh & Bayer-Topilsky, 2014).

The findings present a similar picture regarding bias in relation to children from ultraorthodox families. As hypothesized, children from ultraorthodox families will have a higher risk assessment than children from secular families. Stereotypical attitudes appear to lead to biased decisions against ultraorthodox families. Even though we have found no evidence of previous studies that dealt with this issue, this finding is not unexpected. In Israel, the relationship between the secular and ultraorthodox sectors is complex. It is characterized by mutual severing and ostracism, accusations and reciprocal unreceptiveness to all ideas, and stances belonging to the other side (Kahaner, Yozgof-Orbakh, & Sofer, 2012). Moreover, the groups’ cultural identity does not develop through dialogue with the environment but through negation and withdrawal. This negating stance has become even more extreme in recent years since the cultural groups, ultraorthodox versus secular, are emphasizing their unique characteristics and their conflict is generating processes of mutual rejection, segregation, and alienation (Leshem, 2003). Social workers, who are exposed to this discourse, identify the child by means of latent, automatic stereotypes. In this context, it is interesting that social workers’ personal characteristics were found to have no impact on this bias, implying that even social workers who define themselves as either religious or ultraorthodox have a stereotypical perception of their own sector of belonging.

Regarding the influence of the child’s gender on the decision, the study findings show that social workers tend to make a higher risk assessment for boys than for girls. In that, it is in line with research on actual out-of-home-placement settings in Israel characterized by higher rates of boys compared to girls (Shimoni & Benbenishty, 2011). While this finding supports the research hypotheses, it contrasts the findings of a previous study that the child’s gender had no effect on the social worker’s decision in complex and ambiguous situations such as cases of parental custody in divorce cases (Nouman et al., 2016). It is possible that this difference is derived from difference in the nature and context of decisions involved. While in custodial cases, the judicial tendency is toward gender neutrality, in the case of out-of-home placement, the perceptions of risk and best interest of the child are more attuned to gender differences.

While those three factors had an impact on the risk assessment, they had no direct impact on the placement decision beyond risk assessment. In other words, risk assessment served as a mediating factor that was influenced by the cultural biases of the social workers, but those biases did not affect directly the placement recommendation. This finding is in contrast to previous findings (Enosh & Bayer-Topilsky, 2014), which indicated that cultural biases of social workers affect both the risk assessment and the following placement recommendation. This previous research has examined the effects of SES and ethnic origin and compared low socioeconomic and midclass families. The current study has neutralized the effects of perception regarding SES by describing case vignettes that are all poor families (low socioeconomic). Thus, this finding indicates that SES may serve as a moderating factor vis-à-vis workers’ decision-making processes. This should be examined further in future research.

The essential finding of the present study, which is in line with previous studies of social workers’ decision-making regarding children in needy situations (Enosh & Bayer-Topilsky, 2014, Enosh et al., 2016; Enosh, Nouman, & Sharon, 2016; Nouman et al., 2016, 2018), is that stereotypical attitudes play a part in the decision-making processes regarding children at risk. Social workers are not working in a vacuum but have their own personal history made up of rules that they have formed about life, causal relationships, and stances. These rules, according to the “representative heuristic” (Tversky & Kahneman, 2005), are constructed through a mental process and help the social worker to process the large amount of information in the environment while anticipating possible events. The facts acquire meaning only if they adhere to a consistent and significant structure. Since social workers in ambiguous and complex situations have difficulty in evaluating the probability of events or the degree of precision without a clear structure, such as in the case of children’s risk-assessment decisions, they will project their internally created structure onto the decision, complete with internalized stereotypes.

The main limitation of the study was methodological, in that we examined specific variables, which might influence the measurement’s credibility. Additional variables, such as the child’s living environment, or other familial variables might have an influence on this complexity. Therefore, this limitation can serve as a basis for the recommendation for further research that will examine other variables, as well as those that were examined in this study. This comprehensive examination might negate the present claim that decision-making pertaining to children at risk is still influenced by long-standing social and cultural stereotypes. Another limitation is that the study population was exclusively in Israel. We recommend examining the implications of the role of ethnicity in biasing social workers’ decision-making processes in other sociocultural contexts. Finally, decision-making regarding children at risk in Israel takes place in the framework of treatment planning committees, who examine the case in question in a group forum. This study examined the personal perceptions and attitudes of individual social workers, and not within a group framework. It is possible that group biases exist, which influence the social worker’s decision-making and warrant further research.

Limitations notwithstanding, child protection decision-making carries high levels of personal and social responsibility on the one hand and may be prone to misjudgments and biases due to lack of sufficient information on the other. The findings of the present study indicate the need to raise awareness and sensitivity to the role played by sociocultural biases of social workers in the child protection field, regarding of ethnic minority families as well as regarding the child’s gender. There is an urgent need for clear and unambiguous policies and criteria in order to reduce the heuristic use of personal and cultural biases in such cases. This is an opportunity for learning, growth, and development of professional work methods, which include assimilation of organizational mechanisms and professional training in ways that will help to overcome these failures and to maximize the decision-making processes to benefit children at risk.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Banach

(1998). The best interests of the child: Decision-making factors. Families in Society, 71, 331–340.

Ben-Arieh

Haj-Yahia

M. M.

(2006). The “geography” of child maltreatment in Israel: Findings from a national data set of cases reported to the social services. Child Abuse & Neglect, 30, 991–1003.

Benbenishty

Chen

(2003). Decision making by the child protection team of a medical center. Health & Social Work, 28, 284–292. doi:10.1093/hsw/28.4.284

Benbenishty

Davidson-Arad

López

Devaney

Spratt

Koopmans

… Hayes

(2015). Decision making in child protection: An international comparative study on maltreatment substantiation, risk assessment and interventions recommendations, and the role of professionals’ child welfare attitudes. Child Abuse & Neglect, 49, 63–75. doi:10.1016/j.chiabu.2015.03.015

Benbenishty

Osmo

Gold

(2003). Rationales provided for risk assessments and for recommended interventions in child protection: A comparison between Canadian and Israeli professionals. British Journal of Social Work, 33, 137–155. doi:10.1093/bjsw/33.2.137

Britner

P. A.

Mossler

D. G.

(2002). Professionals’ decision-making about out-of-home placements following instances of child abuse. Child Abuse & Neglect, 26, 317–332. doi:10.1016/S0145-2134(02)00311-3

Brunnberg

Pe’cnik

. (2007). Assessment processes in social work with children at risk in Sweden and Croatia. International Journal of Social Welfare, 16, 231–241.

Croskerry

(2009). Clinical cognition and diagnostic error: Applications of a dual process model of reasoning. Advances in Health Sciences Education, 14, 27–35. doi:10.1007/s10459-009-9182-2

Davidson-Arad

(2010). Comparison of quality of life of children at risk remaining at home and children removed from home. Hevra Urevaha (Society and Welfare), 30, 397–421. (In Hebrew)

10.

Dhami

M. K

. (2003). Psychological models of professional decision making. Psychological Science, 14, 175–180.

11.

Dhami

M. K.

Ayton

. (2001). Bailing and jailing the fast and frugal way. Journal of Behavioral Decision Making, 14, 141–168.

12.

Doron

(2004). Multiculturalism in Israel and the erosion of support for the “welfare state.” Iyunim Bitkumat Israel (Studies in Israeli and Modern Jewish Society), 14, 55–72. (In Hebrew)

13.

Drake

Jolley

J. M.

Lanier

Fluke

Barth

R. P.

Jonson-Reid

(2011). Racial bias in child protection? A comparison of competing explanations using national data. Pediatrics, 127, 471–478.

14.

Ebel

R. L.

(1951). Estimation of the reliability of ratings. Psychometrika, 16, 407–424.

15.

Enosh

Bayer-Topilsky

. (2014). Reasoning and bias: Heuristics in safety assessment and placement decisions for children at risk. The British Journal of Social Work, 45, 1771–1787.

16.

Enosh

Nouman

Anabtawi

. (2016). Evaluating child-custody recommendations of Israeli Arab social workers: Between traditionalism and professional decision-making. Research on Social Work Practice. doi:1049731516666382

17.

Enosh

Nouman

Sharon

(2016). Between professionalism and traditional social norms: Social workers’ parental custody recommendations. British Journal of Social Work, 47, 2032–2048.

18.

Evans

J. S. B. T.

(2008). Dual-processing accounts of reasoning, judgment, and social cognition. Annual Review of Psychology, 59, 255–278.

19.

Fallon

Chabot

Fluke

Blackstock

MacLaurin

Ronmyr

(2013). Placement decisions and disparities among Aboriginal children. Further analysis of the Canadian incidence study of reported child abuse and neglect, part A: Comparisons of the 1998 and 2003 surveys. Child Abuse and Neglect, 37, 47–60.

20.

Fialkov

M. J.

Cohen

(1990). The mental health professional, the legal process, and the child in out-of-home care. In Grabe

P. V.

(Ed.), Adoption Resources for Mental Health Professionals (pp. 189–202). New Brunswick, NJ: Transaction Publishers.

21.

First

(1997). Television and the construction of social reality: An Israeli case study. In McCombs

Shaw

D. L.

Weaver

(Eds.), Communication and Democracy (pp. 41–50). Hillsdale, NJ: Lawrence Erlbaum.

22.

Fluke

J. D.

Chabot

Fallon

MacLaurin

Blackstock

(2010). Placement decisions and disparities among aboriginal groups: An application of the decision making ecology through multi-level analysis. Child Abuse & Neglect, 34, 57–69. doi:10.1016/j.chiabu.2009.08.009

23.

Gambrill

E. D.

(2005). Decision making in child welfare: Errors and their context. Children and Youth Services Review, 27, 347–452.

24.

Gigerenzer

Hertwig

Pachur

(2011). Heuristics: The foundations of adaptive behavior. New York, NY: Oxford University Press.

25.

Gilbert

Parton

Skivenes

. (Eds.). (2011). Child protection systems: International trends and orientations. New York, NY: OUP.

26.

Haberfeld

Semyonov

Cohen

. (2000). Ethnicity and labour market performance among recent immigrants from the former Soviet Union to Israel. European Sociological Review, 16, 287–299. doi:10.1093/esr/16.3.287

27.

Hacker

Frenkel

(2005). Active parenthood and employment equal opportunities: The need to change the characteristics of the labor force. Labor, Society and Law, 11, 275–303. (In Hebrew)

28.

Hall

D. J.

Paradice

(2007). Investigating value-based decision bias and mediation: Do you do as you think? Communications of the ACM, 50, 81–85. doi:10.1145/1232743.1232752

29.

Hall

(1997). Representation: Cultural representations and signifying practices. London, England: Sage.

30.

Harris

M. S.

Hackett

. (2008). Decision points in child welfare: An action research model to address disproportionality. Children and Youth Services Review, 30, 199–215. doi:10.1016/j.childyouth.2007.09.006

31.

Jarbi

Levy

(2000). The socio-economic rift in Israel. Jerusalem, Israel: The Israel Democracy Institute. (In Hebrew)

32.

Kahaner

Yozgof-Orbakh

Sofer

(2012). The Haredim (Ultra-Orthodox) in Israel: Breadth, society, community. Haifa, Israel: University of Haifa. (In Hebrew)

33.

Legal Capacity and Guardianship Law. (1962). Retrieved from http://index.justice.gov.il

34.

Leshem

(2003). Israel as a multicultural state at the turn of the 21st century. In Leshem

Roer-Strier

(Eds.), Cultural diversity: A challenge to human services (pp. 112–113). Jerusalem, Israel: Magnes Press. (In Hebrew)

35.

Lulu

. (2000). Attitudes of social workers and child protection officers for legal proceedings in a social services department to out-of-home placement of children (Unpublished master’s thesis). Bar-Ilan University, Ramat Gan, Israel. (In Hebrew)

36.

Mattison

(2000). Ethical decision making: The person in the process. Social Work, 45, 201–212.

37.

McLennan

R. N.

(1993). Interrater reliability with SPSS for Windows 5.0. The American Statistician, 47, 292–297.

38.

Miller

K. M.

Cahn

Orellana

E. R

. (2012). Dynamics that contribute to racial disproportionality and disparity: Perspectives from child welfare professionals, community partners, and families. Children and Youth Services Review, 34, 2201–2207. doi:10.1016/j.childyouth.2012.07.022

39.

Morton

C. M.

Ocasio

Simmel

. (2011). A critique of methods used to describe the overrepresentation of African Americans in the child welfare system. Children and Youth Services Review, 33, 1538–1542. doi:10.1016/j.childyouth.2011.03.018

40.

Munro

(2008). Lessons from research on decision-making. In Lindsey

Shlonsky

(Eds.), Child welfare research: Advances for practice and policy (pp. 194–200). New York, NY: Oxford University Press.

41.

Musen

M. A.

van Bemmel

J. H.

(Eds.). (1997). Handbook of medical informatics. Houten, the Netherlands: Bohn Stafleu Van Loghum.

42.

Nouman

Enosh

Jarjura

. (2018). Between professional norms and professionalism: Risk assessment and decision-making of Arab-social-workers regarding children-at-risk. Research on Social Work Practice. doi:10.1177/1049731518766677

43.

Nouman

Enosh

Niselbaum-Atzur

. (2016). The role of parental communication, child’s wishes and child’s gender in social workers’ custody recommendations. Children and Youth Services Review, 70, 302–308. doi:10.1016/j.childyouth.2016.09.034

44.

Patton

M. Q.

(2011). Essentials of utilization-focused evaluation. Thousand Oaks, CA: Sage.

45.

Rechnitzer

(2013). The experiences of Haredi therapists treating child sexual abuse cases in the Haredi sector (Unpublished MA thesis). Haifa, Israel: University of Haifa.

46.

Regehr

LeBlanc

Shlonsky

Bogo

(2010). The influence of clinicians’ previous trauma exposure on their assessment of child abuse risk. Journal of Nervous and Mental Disease, 198, 64–68.

47.

Rogers

W. H.

(1993). Regression standard errors in clustered samples. Stata Technical Bulletin, 13, 19–23.

48.

Rotlevy Committee. (2003). Conclusions of the committee on the rights of the child. Jerusalem, Israel: Ministry of Justice. (In Hebrew)

49.

Shimoni

Benbenishty

(2011). Child victims of maltreatment—A comparison between children placed in out-of-home frameworks and those treated in the community: Background, current situation, and results. Jerusalem, Israel: The Haruv Institute. (In Hebrew)

50.

Shlonsky

Wagner

. (2005). The next step: Integrating actuarial risk assessment and clinical judgment into an evidence-based practice framework in CPS case management. Children and Youth Services Review, 27, 409–427. doi:10.1016/j.childyouth.2004.11.007

51.

Silman Committee. (2014). Report examining the Ministry’s policy on out-of-home placement of children and the issue of visiting rights. (Hebrew). Retrieved from http://www.molsa.gov.il

52.

Slonim-Nevo

Lander

(2004). Can the child’s best interest exist separately from the good of the family? Thoughts and recommendations for change. Hevra Urevaha (Society and Welfare), 34, 401–433. (In Hebrew)

53.

Spratt

Nett

Bromfield

Hietamäki

Kindler

Ponnert

. (2014). Child protection in Europe: Development of an international cross-comparison model to inform national policies and practices. The British Journal of Social Work, 45, 1508–1525. doi:10.1093/bjsw/bcu109

54.

StataCorp. (2015). Stata statistical software: Release 14. College Station, TX: StataCorp LP.

55.

Taylor

(2010). Professional decision making in social work. Exeter, England: Learning Matters.

56.

Taylor

B. J.

(2016). Heuristics in professional judgement: A psycho-social rationality model. British Journal of Social Work, 47, 1043–1060.

57.

Tur-Kaspa Shimoni

Schwarzwald

(2003). Perceived threat and prejudice in three domains of intergroup tension in Israeli society. Megamot, 42, 549–584. (In Hebrew)

58.

Tversky

Kahneman

(1975). Judgment under uncertainty: Heuristics and biases. In: Wendt

Vlek

(Eds.). Utility, Probability, and Human Decision Making (pp. 141–162). Springer: Dordrecht.

59.

Tversky

Kahneman

(2005). Judgment under uncertainty: Heuristics and biases. In Bar-Hillel

(Ed.), Rationality, fairness, happiness: Selected articles by Daniel Kahneman (pp. 45–63). Haifa, Israel: University of Haifa and Keter. (In Hebrew)

60.

Williams

R. L

. (2000). A note on robust variance estimation for cluster correlated data. Biometrics, 56, 645–646. doi:10.1111/j.0006-341X.2000.00645.x

61.

Zafran

(2010). Children’s rights in the family as relational rights. In Morag

Rahamim

(Eds.), Children’s rights and Israeli law (pp.129–158). Tel Aviv, Israel: Ramot and Tel Aviv University. (In Hebrew)

62.

Zarembski

(2002). The Floersheimer Institute for policy studies the religious-secular divide in the eyes of Israel’s leaders and opinion makers. Retrieved from http://fips.huji.ac.il/sites/default/files/floersheimer/files/zarembsky_religious_secular_divide_202_en

63.

Zeira

Attar-Schwartz

Benbenishty

(2012). Children and youth in out-of-home placement in Israel—Issues and challenges. Mifgash: Journal of Social-Educational Work, 36, 9–16. Retrieved from http://www.jstor.org/stable/23693112