Introduction

The relationship between religious experiences and health outcomes has been well established. Greater levels of religion and spirituality (RS) are generally associated with more positive mental and physical health outcomes among adults and children (Koenig 2012; Paloutzian and Park 2005; Pargament et al. 2013). Aspects of RS such as positive religious coping (Pargament et al. 2013), secure attachment to God (Homan and Boyatzis 2010a, b), spiritual transcendence (Piedmont 2001) and a greater sense of identity (King et al. 2013), are associated with better physical and psychological health. With regard to overall psychological well-being, these variables are associated with lower levels of depression and anxiety and higher levels of self-esteem (Homan and Boyatzis 2010a, b; Koenig 2012; Latzer et al. 2007; Maltby and Day 2000). Conversely, certain aspects of religious experience may also have a negative impact. For example, when religious experience is accompanied by guilt, it may increase or exacerbate underlying psychological issues (Exline and Rose 2005; Silton and Fogel 2010). Furthermore, negative religious coping is associated with more negative mental and physical health outcomes (Latzer et al. 2015; Rosmarin et al. 2009). In addition, the role of religious orientation or underlying attitudes toward religious practice has consistently been linked to mental and physical health outcomes. An intrinsic religious orientation (an internally focused form of motivation toward religious practice and belief) (Allport and Ross 1967) has been found to have a protective effect with regard to mental and physical health, whereas an extrinsic orientation (an externally and socially focused form of motivation toward religious practice and belief) is linked to more negative outcomes. Specifically, it has been consistently found that an extrinsic religious orientation is associated with higher levels of depression, anxiety and lower levels of psychological well-being relative to an intrinsic religious orientation (Lewis et al. 2005; Maltby and Day 2000; Smith et al. 2003).

It has increasingly been recognized that religious variables and attitudes play an important role in the understanding of eating disorder-related psychopathology (Berrett et al. 2010; Latzer et al. 2015; Richards et al. 2009, 2013). The last several decades have seen a dramatic increase in the prevalence of such psychopathology such as preoccupation with dieting behaviors, weight, body shape and a cultural obsession with thinness. This is most pronounced among adolescent and young adult females in Western-oriented cultures (Smink et al. 2012; Treasure et al. 2010). These behaviors and attitudes are risk factors associated with a continuum of disturbed eating behavior ranging from mild disordered eating, to moderate levels of disordered eating pathology (DEP), through full-blown clinical levels of eating disorders. Throughout this paper, we refer primarily to disordered eating pathology (DEP) which relates to a moderate level of eating disturbance which in some cases may predispose one to the development of a full-blown eating disorder.

Religious variables such as coping, attachment and religious orientation have been found to be associated with disordered eating pathology (DEP) and related risk factors (Homan and Boyatzis 2010a, b; Latzer et al. 2015; Weinberger-Litman et al. 2008). Specifically, an extrinsic religious orientation (ERO) has been found to be associated with greater risk of DEP among non-clinical samples and worse prognosis among clinical samples of individuals with eating disorders (Forthun et al. 2003; Smith et al. 2004; Weinberger-Litman et al. 2008). This relationship was found among different religious groups including various Christian denominations and Jews.

One of the most salient risk factors strongly associated with DEP is high levels of body dissatisfaction and low body-esteem. Body dissatisfaction is a core feature of DEP and full-blown eating disorders (Fairburn et al. 2008; Fairburn and Harrison 2003; Treasure et al. 2010). Body dissatisfaction is often related to unrealistic sociocultural messages related to appearance targeted at young women (Rohde et al. 2015). These often unrealistic appearence related expectations, coupled with the pressure to fulfill multiple and often conflicting roles such as traditional female roles (e.g., wife and mother) as well as more contemporary roles (e.g., those related to education and career) has been termed the “Superwoman” Ideal. Greater adherence to the perfectionistic Superwoman Ideal has been associated with greater levels of body dissatisfaction as well as DEP (Levine and Smolak 2013; Mensinger 2005; Mensinger et al. 2007; Steiner-Adair 1986).

In previous studies, greater adherence to the Superwoman Ideal has been found to mediate the relationship between extrinsic religious orientation and DEP among young adult Jewish women in the USA (Weinberger-Litman et al. 2016). The concept of the Superwoman Ideal is particularly relevant to Modern Orthodox women who are faced with fulfilling traditional roles as well as secular ones. For example, Modern Orthodox women are expected to marry young and have many children, dress modestly yet fashionably, be a good hostess and display exemplary character and selflessness. In addition, Modern Orthodox women are expected to pursue higher education and often prestigious careers. Some have noted that the Orthodox community engenders its own version of the Superwoman Ideal, labeled the “Eshet Chayil,” or woman of valor (Diamond and Lenik 2009). It is unknown whether the association with extrinsic orientation and DEP or the mediating effect of the Superwoman Ideal is relevant in other Jewish communities around the world that may have different cultural pressures or expectations. In particular, Israeli adolescents may face a different set of cultural expectations with regard to role fulfillment among different religious subgroups.

The religious milieu of Israeli society is highly complex and diverse. About 20% of the population identify as religious, whereas the majority identify as traditional. Of the religious population in Israel, a smaller percentage (approximately 10%) identify as “Dati Leumi,” or Modern Orthodox. Young women in this segment of the population face pressures from both traditional and contemporary societal expectations. Most Modern Orthodox young women serve in the Israeli Defense Force (Glazer 2013) are exposed to Israeli media and pursue higher education while also being expected to marry at a young age and fulfill traditional roles of wife and mother. This may lead to a higher level of role conflict than experienced by young women in other segments of society as the traditional roles are more emphasized in this subgroup. Therefore, the Superwoman Ideal may be particularly salient among this group and may be associated with religious attitudes such as religious orientation and psychological experiences such as body dissatisfaction.

Thus, this study has several goals. First we aim to establish the relationship between extrinsic orientation and DEP among this sample. Further, in order to elucidate how extrinsic religious orientation is associated with DEP, we explore whether the Superwoman Ideal mediates this relationship. Next, we aim to assess whether higher levels of adherence to the Superwoman Ideal translate into higher levels of body dissatisfaction whereby body dissatisfaction functions as a serial mediator of DEP.

These questions are examined in the context of a multiple serial mediation model, in which the Superwoman Ideal serially precedes body dissatisfaction. The main goals of this mediation model were to test for 1) whether extrinsic orientation directly effects DEP (independent of the mediators), (2) whether extrinsic orientation effects DEP through adherence to the Superwoman Ideal, (3) whether extrinsic orientation effects DEP through increases in body dissatisfaction and (4) whether the mediating effect of body dissatisfaction is mediated by the Superwoman Ideal (serial mediation).

Method

Procedure

A total of 180 Modern Orthodox (“Dati Leumi”) Israeli female adolescents were recruited and 120 completed the study. Participants were recruited from three schools and ranged from grades 7–12. The study was approved by the Israeli Ministry of Education as well as the Haifa University ethics committee. In addition, the school principals approved of the study. Parental consent was required for participation. Students were told they were participating in a survey of adolescent eating-related attitudes and behaviors. Participation was voluntary and anonymous.

Measures

Validated Hebrew versions of all measures were used or were translated into Hebrew for the purpose of the study. A validated Hebrew version of the EDI-2 was used. All other measures were translated into Hebrew for the purposes of this study. The BSQ, ROS and Superwoman Ideal were translated into Hebrew and back-translated into English by two dual-language (Hebrew–English) researchers.

Disordered Eating Pathology

The Eating Disorder Inventory (EDI) (Garner et al. 1983) is one of the most widely used self-report questionnaires for assessing psychological characteristics related to eating disorder pathology among Western populations. It is not intended to be used as a diagnostic instrument, but rather provides a profile of certain clusters of symptoms commonly found among individuals with eating disorders. It consists of eight subscales, and the following core subscales were used in the current study: Drive for Thinness, Bulimia and a total composite score. The Hebrew translation of this inventory was found to be valid and reliable (Niv et al. 1997).

The EDI-2 contains 91 items, which are scored on a three-point scale. There are six response options, which include “always,” “usually,” “often,” “sometimes,” “rarely” and “never.” Responses are scored from 0 to 3, with answers of “never,” “rarely” and “sometimes” assigned a 0 and “always,” “usually” and “often,” assigned scores of 3, 2 and 1, respectively. Total scores range from 0 to 273, with higher scores indicating more disordered eating pathology. Subscale scores range from 0 to 30 depending on the number of items in the subscale. Internal consistency as measured by Cronbach’s ranged from .91 to .65 (Garner et al. 1983). In the current study, the Drive for Thinness and Bulimia Subscale as well as the total score was used. Cronbach’s alpha was .83 for the Drive for Thinness Subscale, .62 for the Bulimia Subscale and .91 for the total score.

Body Dissatisfaction

The Body Shape Questionnaire (BSQ) (Cooper et al. 1987) is a 34-item scale that measures body dissatisfaction through concerns about body shape and weight. Responses are scored in a six-point Likert-type scale with responses ranging from “always” = 6 to “never” = 1. Scores on the BSQ can range from 34 to 204, with a higher score indicating more body dissatisfaction. The BSQ has demonstrated high internal consistency (Cooper et al. 1987). In the current study, Cronbach’s alpha was .97.

Religious Orientation

The ROS is a 20-item scale that measures intrinsic (I) and extrinsic (E) religiosity and is the most widely used measure of religious orientation (Allport and Ross 1967; Donahue 1985; Kirkpatrick and Hood 1990). The intrinsic subscale consists of nine items, and the extrinsic subscale consists of 11 items. An intrinsic orientation is considered a form of religious devotion in which individuals internalize their religious beliefs, are internally driven to be engaged in religious life and “live their religion” (Allport and Ross 1967; Donahue 1985; Kirkpatrick and Hood 1990). An extrinsic orientation refers to an external focus on religious beliefs whereby individuals engage in religious practice for social or personal reasons (Kirkpatrick and Hood 1990) and are thought to “use their religion” (Allport and Ross 1967). The extrinsic subscale was used in the current study. The ROS has demonstrated good Cronbach’s alpha reliability, ranging from .69 to .93 (Donahue 1985; Kirkpatrick and Hood 1990; Masters et al. 2004). In the current study, the extrinsic subscale was used. Cronbach’s alpha was .52.

Superwoman Ideal

The 20-item Superwoman Ideal scale was used to measure the Superwoman construct, which represents gender role conflict via perfectionist achievement in multiple roles. We used a modified version (Mensinger 2003, 2005; Mensinger et al. 2007) of the original Superwoman scale developed by Murnen et al. (1994). Participants rated the importance of a variety of items related to educational success, relationships, physical attractiveness and social obligations. Sample items include: “In order to feel fulfilled someday, I would need to ‘have it all’ (that is, a strong marriage, a prestigious career and great kids),” and “It means a great deal to me to be the very best at everything I do.” Response options were scored on a six-point Likert scale ranging from “strongly disagree” = 1 to “strongly agree” = 6. Higher scores indicate greater endorsement of the Superwoman Ideal. Cronbach’s alpha reliability for the revised scale was .88 (Mensinger 2003, 2005). Cronbach’s alpha reliability in the current study was .81.

Statistical Analyses

Descriptive statistics and zero-order correlations were conducted. Next, a serial mediation model was utilized to examine three hypotheses (see Fig. 1). First, we examined the hypothesis that extrinsic religious orientation is associated with an increase in adherence to the Superwoman Ideal, which in turn increases disordered eating pathology. This hypothesis was examined by evaluating the parameters for the indirect path from extrinsic religious orientation to disordered eating pathology through Superwoman Ideal as the single mediator (path a1 → d1 in Fig. 1) (Tables 1, 2).

Fig. 1
figure 1

Mediated meditation model. Mediated mediation model in which extrinsic religious orientation is the predictor, and Superwoman Ideal and body dissatisfaction are two serial mediators, for each of three disordered eating pathology outcomes. EDI Eating Disorder Inventory

Full size image
Table 1 Descriptive statistics
Full size table
Table 2 Zero-order correlations
Full size table

We then examined the hypothesis that that extrinsic religious orientation is associated with an increase in body dissatisfaction, which in turn increases disordered eating pathology. This hypothesis was examined by evaluating the parameters for the indirect path from extrinsic religious orientation to disordered eating pathology through body dissatisfaction as the single mediator (path b1 → a3 in Fig. 1).

Lastly, we examined the hypothesis that extrinsic orientation is associated with an in increase in adherence to the Superwoman Ideal, which subsequently increases body dissatisfaction, which in turn increases disordered eating pathology. This hypothesis was examined by evaluating whether the indirect path from extrinsic religious orientation to disordered eating pathology is serially mediated by, first, Superwoman Ideal and, second, body dissatisfaction (path a1 → a2 → a3 in Fig. 1).

The three hypotheses above were evaluated for three separate outcome measures of disordered eating pathology: (1) total EDI scale, (2) EDI Bulimia Subscale and (3) EDI Drive for Thinness Subscale. Thus, three separate serial mediations models were run. Each model used a different EDI subscale as an outcome measure (see Table 3).

Table 3 Serial mediation models examining the association between extrinsic religious orientation and disordered eating pathology
Full size table

Mediation models were computed utilizing PROCESS, which is a mediation and moderation SPSS add-on (Hayes 2012). To test mediation effects, we followed the bootstrapping method, utilizing 5000 iterations (Hayes 2013; Hayes and Scharkow 2013; Preacher et al. 2007). The bootstrapping procedure tests the null hypothesis that the indirect path from the interaction term to the dependent variable via the mediator does not significantly differ from zero. If zero is not contained within the confidence intervals (CI) computed by the bootstrapping procedure, one can conclude that the indirect effect is indeed significantly different from zero at p < .05. In addition to the confidence intervals, p values based on the Sobel method for evaluating the significance of an indirect effect were also utilized.

Results

Participant Characteristics

Participants ranged in age from 12 to 18 years with mean age of 14.88 (SD = 1.97). The mean body mass index (BMI), measured using self-reported height and weight (calculated by taking participants’ weight in kilograms divided by the square of their height in meters), was 19.81 (SD = 3.33). (See Table 1 for descriptive statistics).

As shown in Table 2, extrinsic religious orientation (ERO) was significantly correlated with EDI Drive for Thinness, EDI total score and the Superwoman Ideal. ERO was not significantly correlated with the EDI Bulimia Subscale or body dissatisfaction as measured by the BSQ .

Serial Mediation Models

Table 3 presents the three serial mediation models and includes unstandardized coefficients, standard errors, Sobel test p values and bootstrapped confidence intervals for all pathways in the models. The three models were consistent in showing that the 95% confidence interval for the direct effect between extrinsic religious orientation and DEP always included zero, (p (model 1) = .34, p (model 2) = .75, p (model 3) = .26), suggesting that the direct effect of extrinsic religious orientation is not significant and that the proposed mediators are driving the observed association.

When the indirect effects were examined using the Superwoman Ideal as a single mediator, the 95% confidence interval did not include zero, indicating that Superwoman Ideal is a significant mediator of the association between extrinsic religious orientations and eating attitudes (all p’s < .01). However, when the indirect effects of body dissatisfaction as a single mediator were examined, the three models were also consistent in showing that body dissatisfaction was not a significant mediator of extrinsic religious orientation on eating attitudes, as the 95% confidence interval included zero in all three models.

The previous individual mediation models established that the Superwoman Ideal but not body dissatisfaction is a significant individual mediator of the association between ERO and DEP. However, because body dissatisfaction has been shown to be the most proximal risk factor associated with DEP, we examined whether body dissatisfaction is a significant mediator, when mediated first by the Superwoman Ideal. The serial mediation model (where body dissatisfaction is serially preceded by adherence to the Superwoman Ideal) was significant with all three DEP outcome variables (95% confidence interval ranged from did not include zero), indicating that the indirect effect for serial mediation was significantly different from zero at the p < .05 level. Reversing the order of the serial mediators resulted in nonsignificant indirect effects. These results suggest that extrinsic religious orientation is associated with increased adherence to the Superwoman Ideal which in turn leads to increased body dissatisfaction, leading to increases in DET. In other words, the association between an extrinsic religious orientation and greater levels of DEP which is serially generated by adherence to the Superwoman Ideal subsequently leads to greater body dissatisfaction.

Discussion

To our knowledge, cross-cultural studies of the relationship between extrinsic religious orientation and DEP have not been conducted. Furthermore, to the best of our knowledge the mediating mechanisms of this relationship have not been adequately explored. Therefore, the aim of this study was to assess the association between extrinsic religious orientation and DEP among Israeli Modern Orthodox adolescents and to explore whether the Superwoman Ideal mediates this relationship. In addition, we further explored whether greater levels of adherence to the Superwoman Ideal impact body dissatisfaction which then functions as a serial mediator of the association between extrinsic religious orientation and DEP.

Consistent with our hypothesis, extrinsic religious orientation initially was found to be associated with the EDI total score as well as the EDI Drive for Thinness Subscale. These results are consistent with previous results among non-Jews as well as Orthodox and non-Orthodox Jews in the USA (Smith et al. 2004; Weinberger-Litman et al. 2008). This finding further highlights the importance of religious variables and in particular religious orientation with regard to DEP, by spanning geographical and cultural differences.

Furthermore, the Superwoman Ideal was found to significantly mediate this relationship. This is of even greater relevance considering the current sample focused on the specific subgroup of Modern Orthodox Jewish adolescents in Israel. As mentioned earlier, this sample represents a unique group within Israel’s Orthodox population and is also unique compared to Modern Orthodox adolescents in the USA given the cultural differences and expectation between the two countries. This subgroup is not isolated from mainstream cultural values, and in addition to fulfilling expectations with regard to religious obligations, these individuals are exposed to media, participate in compulsory military service and engage with the larger cultural activities. Therefore, they are often forced to navigate between contradictory and competing cultural expectations. As a result, young Modern Orthodox women are expected to fulfill traditional roles such as wife and mother while also fulfilling culturally sanctioned expectations with regard to appearance, education and the workplace, thereby engendering the Superwoman role. A greater level of extrinsic orientation may predispose one to greater internalization of the importance of the Superwoman Ideal which has consistently been shown to be associated with DEP.

We even further explored this relationship by investigating whether the Superwoman Ideal is serially related to greater levels of body dissatisfaction. Body dissatisfaction is one of the most strongly identified factors related to DEP (Rohde et al. 2015; Treasure et al. 2010). Therefore, we hypothesized that greater adherence the Superwoman Ideal would also be associated with greater levels of body dissatisfaction which further mediates the effect of extrinsic religious orientation. Consistent with our hypothesis, body dissatisfaction emerged as a significant serial mediator. The addition of body dissatisfaction to the mediation model greatly contributes to our understanding of how psychosocial variables such as the Superwoman Ideal may play a causal role in the development of DEP. These results are further in line with the sociocultural theory of DEP development, in particular research relating to objectification of women (Dittmar et al. 2009), the internalization of unrealistic appearance ideals and general models of social pressures (Rohde et al. 2015; Thompson and Stice 2001). Adherence to the Superwoman Ideal is consistent with all of these variables and further contributes to risk factors such as body dissatisfaction which have been shown to contribute to DEP.

It is important to note that although we found strong support for the above models, the general levels of DEP in the current sample are relatively low compared to secular samples both in Israel and in the USA (Latzer et al. 2007; Shore and Porter 1990) and are consistent with other religious samples in Israel and the USA (Latzer et al. 2007, 2015). This may suggest that overall higher levels of religiosity may serve a protective factor toward psychopathology including DEP. However, our results emphasize that an extrinsic orientation and adherence the Superwoman Ideal may confer a greater level of vulnerability and risk to DEP through the development of body dissatisfaction even among a highly religious population.

Limitations

There are several limitations that need to be addressed. The sample is of a relatively small size with a large age range. Given that the age range of the sample is 12–18, there may be several developmental issues that may relate to adherence to the Superwoman Ideal and body dissatisfaction; however, the small sample size did not allow for these comparisons. Specifically older individuals, that are closer to marriageable age, may be more likely to internalize messages about the importance of the Superwoman Ideal and therefore experience an increase in body dissatisfaction. A larger sample is necessary to identify whether age further contributes to the observed relationship.

Sample characteristics may also be contributing to the low level of Cronbach’s alpha reliability observed with the ROS. To our knowledge, this is the first time this questionnaire has been translated into Hebrew. In addition, as mentioned above, many of the participants were very young and the questions may have been difficult for some of them to respond to or understand. Despite lower than typical reliability, we still consider this to be a successful first step in introducing the concept of religious orientation into this specific subset of the population. Future studies will address the potential translation or methodological issues associated with the Religious Orientation Scale.

Additionally, data were collected through self-report questionnaires. Some of the questions related to disordered eating and body dissatisfaction are of a personal and sensitive nature, and younger adolescents, particularly in this population, may have difficulty relating to or answering these questions. Modesty is highly valued in this population, and questions relating to specific parts of one body may be seen an intrusive and inappropriate. This may have impacted the responses to the various questionnaires.

The current sample represents a very specific subgroup of the Orthodox population in Israel. It is difficult to generalize our results to the entire Orthodox Jewish community specifically given that more Ultra-Orthodox individuals have less exposure to mainstream cultural ideals and media and may not experience and internalize these values to the same extent. Further, it is not known whether this model relates to the secular population in Israel. Therefore, further research is necessary to clarify whether extrinsic orientation plays a role in other segments of society as well as whether the Superwoman Ideal remains a robust contributor to body dissatisfaction and DEP. Therefore, it is not yet clear whether this segment of the Modern Orthodox population has higher levels of adherence to the Superwoman Ideal relative to their secular or Ultra-Orthodox counterparts. However, previous research in the USA indicates that Modern Orthodox women had higher levels of adherence to the Superwoman than secular Jewish women (Weinberger-Litman 2007). Therefore, further research is needed to clarify the specific cultural factors that may play a role in this particular risk factor.

Conclusions

To our knowledge, this study is the first to explore the association between extrinsic religious orientation, adherence to the Superwoman Ideal and DEP in Modern Orthodox adolescents in Israel. The current findings represent a consistent cross-cultural pattern in line with previous research conducted in the USA. The current findings clarify how an extrinsic orientation may increase risk by being associated with particular psychological variables such as internalization of social pressure and body dissatisfaction. In addition, an understanding of the current model of DEP as it relates to externally focused values (i.e., the Superwoman Ideal) may be of clinical value as well.