This study, which focused on DE and family dinners among Arab and Jewish middle and high-school children in Israel by using nationally representative data, brings to light several unique findings. First, about one out of every five Jewish adolescents and one out of every four Arab adolescents in Israel reported having DE, according to the SCOFF scale. Second, while Arab and Jewish girls reported comparable rates of DE (Jews: 29%; Arabs: 32%), Arab boys reported a significantly higher prevalence of DE compared to Jewish boys (Jews: 11%; Arabs: 16%). Third, eating yesterday’s dinner at home with parents/family members was found to be slightly more common among Arab than Jewish adolescents. Fourth, eating dinner with parents/family members seems to act as a protective factor against DE among Arab and Jewish adolescents in Israel.
To our knowledge, the current study is the first nationally representative assessment of DE among Arab and Jewish adolescents in Israel. The rates reported previously by Kaluski et al. [18] (using data from the first MABAT youth study) included only girls and were based on only four out of the five SCOFF items. Kaluski et al. [18] defined DE as consisting of more than one affirmative SCOFF response, and not as consisting of more than two affirmative responses, as suggested by Morgan et al. [31] Furthermore, Latzer et al. [21], who also reported rates of DE in Israel, concentrated solely on girls, using the Eating Disorder Inventory (EDI) [37] with a non-representative sample and without anthropometric measurements. It is important to note that the rates of DE reported in our study are similar to those reported by other Western countries using the SCOFF questionnaire. For example, German and Finnish studies [12, 38] among large adolescent samples, reported that 24–29% of girls and 15–16% of boys showed at least one symptom of eating disorders.
Our findings suggest slightly higher rates of DE among Arab adolescents compared to Jewish adolescents in Israel, with Arab boys reporting higher rates than Jewish boys, and Arab girls reporting similar rates to Jewish girls. This finding is not surprising, as previous reports had already suggested high rates of DE among Arab youth in Israel [18, 21]. This result is in line with some prior research about DE among ethnic minority youth in the USA, including Hispanic and American Indian youth who reported higher rates of DE relative to white youth [10, 11]. One explanation for this finding could be that youth of ethnic minorities, living in a dominant Western culture, are likely to adopt ideas that stress the importance of body shape and weight (e.g., the thin ideal standard of beauty) [9], as they perceive these ideas as being a way of acclimating to the dominant culture. Relatedly, a study about Muslim-Australian women suggested a positive link between adopting Western values and DE [39]. Nevertheless, some prior results have demonstrated that ethnicity may protect against the development of eating disorder symptoms [40]. Future studies are warranted to understand the exact mechanism which encourages youth of ethnic minority groups to adopt these ideas about the importance of body weight and shape.
Participants were asked to indicate where and with whom they had eaten dinner the day before. The prevalence of home family dinners was significantly higher among Arabs (69%) relative to Jews (65%), but the difference was negligible, suggesting that patterns of family dinners are quite similar among Arabs and Jews. However, interesting differences between Arabs and Jews were found in the alternatives to home family dinners, as when not eating dinner together with parents and family, Arabs were more likely not to eat dinner at all (Arabs:12% Jews: 5%), and Jews were more likely to eat out of the home (Arabs: 8%, Jews: 17%). More research is warranted to specify the reasons behind these differences in dinner patterns among Arab and Jewish adolescents (e.g., differences in SES, family structure, traditions, values).
Multivariate analyses stressed that among both groups (i.e., Arabs and Jews) female gender, older age, and higher BMI were linked with a higher likelihood of DE. Furthermore, relative to home family dinners, all other dinner options (i.e., not eating dinner or eating out of the home among Arabs and Jews, and eating at home alone among Arabs) acted as risk factors for DE. These findings are in line with the literature, as several studies have already demonstrated that male gender, younger age, lower BMI, and frequent family meals can be protective against DE among youth [28]. In addition, Fiese et al. [41] stress that family mealtime is a household routine that provides stability and predictability for children and is related to several important positive child health outcomes, including a lower prevalence of DE. Family meals can serve as opportunities for parents to demonstrate healthy eating patterns, potentially influencing children’s DE attitudes and behaviors [42]. Furthermore, family dinners may provide opportunities for parents to set examples for healthy eating practices and to expose children to a variety of foods. In addition, the time when all family members are sitting around the table can be used in order to engage in a group discussion during which parents gain exposure to adolescents’ life, while also establishing routines and strengthening family connections [24].
Interestingly, eating dinner at home alone (vs. the home family dinner) acted as a risk factor for DE only among Arab, but not among Jewish, adolescents. It is possible to speculate that among some families, parents may have been involved in the child’s meal planning (e.g., by providing him/her with a plate of food), even when the child ate dinner at home alone. Therefore, it could be that among some participants (e.g., Jewish adolescents), eating at home alone was not necessarily linked with a higher risk for DE. Additionally, among both ethnic groups, not eating dinner at all, as well as eating dinner out of the home (vs. the home family dinner), were both related to a higher risk for DE. It is possible that when children eat dinner outside of the home or do not eat dinner at all, parents have fewer opportunities to be vigilant about their children’s eating patterns, as well as a lessened capability to have an impact on their food-related attitudes and behaviors [42]. Indeed, prior studies have suggested that both skipping meals [43] and eating out of the home [44, 45] are linked with weight gain and unhealthful dietary intake, probably due to a lack of parental involvement in the child’s eating habits on those occasions. In contrast, when the child eats at home alone, parents may potentially still be involved in the child’s meal (e.g., via the preparation of the food), and therefore have an impact on his/her eating patterns. It is thus important to look not only at the frequency of family meals, but also at the alternative routines that parents adopt in order to remain vigilant over their child’s eating patterns.
The present study used the SCOFF questionnaire to assess DE prevalence. Several other questionnaires are available for the screening of eating disorders and DE (e.g., EAT-26 [46], EDI [21]), and the prevalence rates of DE among youth might vary considerably depending on the questionnaire and the methods used to obtain the data. The SCOFF is probably the most promising short questionnaire to use in community samples, mainly because it is a brief self-report scale with only five items [47]. Confirmatory factor analyses [48] that assessed the SCOFF in comparison with health examinations among adolescents indicated that the SCOFF is a useful tool for the detection of DE. Relatedly, the third item of the SCOFF assesses for losing more than one stone (=6.35 kg) in weight over a three-month period, while in the design of the current study, a decision was made to be consistent with the first MABAT study [18] by asking participants about losing more than half a stone (=3 kg) in weight. The literature suggests that among adolescents, the third SCOFF item, which asks about weight loss of one stone, fails to provide accurate information in terms of risks for eating disorders [32, 49]. For example, a recent study [49], examining the SCOFF questionnaire among high school population in Ohio, reports that when examining the correlation of each of the five SCOFF items with the summed score scale, the original third item of the SCOFF has the lowest correlation coefficient. In addition, this same study suggests that the original third item of the SCOFF has the worst ability to discriminate risk for eating disorders (relative to the other SCOFF items).
Findings may have several important implications for health policy. Importantly, given that our data indicated high rates of DE among both Arab and Jewish adolescents in Israel, there is a need to devote funding to DE prevention efforts (e.g., through the dedicated training of healthcare providers working with adolescents) [50]. Furthermore, our findings about the link between family dinners and DE may suggest that possible programs to prevent and reduce DE should be conducted not only among adolescents but also among their family members, with a particular emphasis on educating parents of both Arab and Jewish youth regarding the risks associated with low frequency of family meals.
Strengths and limitations
This study marks a first step in addressing an important issue for the health and well-being of Arab and Jewish adolescents. To our knowledge, this is the first study to investigate DE and dinner options among Arab and Jewish adolescents in Israel. The strengths of the current study lie in the high response rate and the large representative sample that included adolescents from both Arab and Jewish schools. Additionally, the BMI was calculated based on measured height and weight by trained personnel, providing a reliable and valid estimate of BMI status. While this study has multiple strengths, it also has several limitations. First, the cross-sectional nature of this study makes it difficult to determine the direction of influence and whether a higher frequency of eating alone contributes to, or results from, more severe DE symptoms. Second, we assessed where and with whom dinner was eaten, but we did not have information about breakfast and lunch [23]. Third, the majority of the data in this study were self-reported and therefore subject to reporting bias, including a social desirability response bias. Fourth, the MABAT Youth Survey is a school-based study and does not include children who are not in school or children who are in private, independent, or boarding schools, including the ultra-Orthodox (Haredi) sector. Fifth, the SES variable was defined ecologically according to the Ministry of Education’s classification of the welfare level of the school. Last, dinner options were reported only in regard to the previous day. Nevertheless, studies suggest that data reported about the previous day may be quite accurate relative to other reporting methods (e.g., reporting about weekly frequency of family meals), because they (the former) are less subject to recall bias [51]. Last, positive and negative predictive values of the SCOFF questionnaire were not examined in the current study. However, a recent systematic review and meta-analysis [32] suggests that the SCOFF is moderately helpful in detecting and ruling out the presence of an eating disorder.