Introduction

Preadolescence (ages 9–13) is a period of high risk for the development of eating disorders (EDs). EDs have been reported to emerge as early as the age of 9 [1] and continue rising in prevalence among individuals aged 10–13 [2,3,4]. As such, more research is needed at earlier ages to identify how EDs develop and emerge and possible prevention targets. Therefore, examining ED symptoms and its correlates is of critical importance during this developmental period.

Among anxiety disorders, evidence indicates that social anxiety disorder (SAD) has the highest prevalence out of all other anxiety disorders [5, 6] and that SAD can be a barrier to receiving effective care [7,8,9]. In younger ages, Schaumberg and colleagues [10] found that anxiety symptoms in middle childhood predict ED symptoms in early adolescence, suggesting that anxiety might predispose individuals to develop EDs. Importantly, the presentation of SAD symptoms remains constant between mid and late adolescent development [11]. Another study reported that adolescents with ED symptoms reported higher SAD symptoms [12, 13], which in turn, would contribute to binge eating [13]. Moreover, appearance-related SAD symptoms predicted ED symptoms in preadolescents [14] and adolescents [15]. In addition to Western societies, SAD symptoms were associated with ED symptoms in non-Western adolescents [16, 17]. Taken together, the literature has established a connection between social anxiety and ED symptoms among college students, adolescents, and ED treatment-seeking patients, and suggests that anxiety is an early predictor of later ED symptoms [10, 18,19,20,21].

In spite of clear links between social anxiety and ED symptoms, the causal mechanisms remain unclear in preadolescents. Most research on these links has been conceptualized from the perspective of traditional latent variable approaches (i.e., structural equation modeling) to psychopathology among college samples and adolescents. These approaches assume that mental disorder symptoms would result from a common latent variable [22]. For instance, traditional approaches postulate that an underlying latent disease (i.e., social anxiety) without any relation among symptoms would produce a variety of psychological symptoms (e.g., fear of evaluation, social avoidance, being observed in the public, doing something in front of others). In other words, traditional approaches to psychopathology often neglect to examine unique associations across symptoms of a disorder, limiting our understanding of the potentially reinforcing connections among those symptoms.

Network theory: a framework to understand ED psychopathology

Network theory is a framework which has received considerable attention and has offered a new perspective to psychopathology in recent years [23]. This theory addresses the above-mentioned limitations of traditional approaches to understanding the relation between social anxiety and ED symptoms. Specifically, network theory suggests that the existence of interactions among symptoms would cause and constitute psychopathology [24, 25]. In a network model, symptoms are represented as nodes, connected by edges that depict the strength (e.g., strong partial correlation) and valence (i.e., positive or negative partial correlation) between pairs of symptoms. Central symptoms, or symptoms with high centrality, are nodes that may demonstrate the strongest connections to other nodes, which are theorized to maintain the network [24,25,26].

To our knowledge, the majority of network research in EDs has focused on adults [27]. Overall this research has identified symptoms related to overvaluation of weight and shape (e.g., body dissatisfaction; desire to lose weight; fear of weight gain) to be central to ED psychopathology, with most ED behaviors remaining peripheral to the network structure. There are only three studies [28,29,30] examining network theory of ED symptoms among children and adolescents with EDs and none have tested social anxiety, though most of this work is in adolescents. The findings from Western societies reported that shape- and weight-related concerns and dieting are central symptoms of EDs in this age group [28,29,30]. However, the findings from these three studies would not be generalizable to nonclinical preadolescents. A network analytic approach to examining ED and social anxiety symptom interconnections in an at-risk sample can identify important pathways that may lead to the development of psychopathology and inform prevention targets. Additionally, these studies did not test network analyses across sex, respectively, as most of the participants were female, nor did they test the relation between ED symptoms and social anxiety symptoms. Furthermore, these findings are all from Western samples. The only network analysis study on EDs in Iran reported that desiring to lose weight was the most central symptom among a large sample of Iranian adolescents (ages 12–19) and college samples (ages 18–54) in both males and females [31]; however, the findings regarding preadolescents are understudied.

Examining bridge symptoms which identify specific illness pathways would be of importance as these bridge symptoms may be important in treatments intended to reduce or prevent comorbid problems [32, 33]. To date, three studies in clinical and nonclinical samples reported that physical sensations (i.e., feelings of wobbliness, lack of interest in sex, changes in appetite), avoidance of social eating, low self-confidence, and feeling like a failure are bridge symptoms between the two diagnostic clusters of bulimia nervosa (BN) and ED symptoms and anxiety/depression symptoms [31, 34, 35]. Another study in a sample of adults with a primary diagnosis of EDs and SAD reported that difficulty with drinking beverages and eating in public were bridge symptoms between EDs and social anxiety symptoms [36]. To date, social anxiety has not been directly quantified in a network model with ED symptoms in non-clinical preadolescents, nor in a non-Western sample. Importantly, a recent review on network analysis in EDs [27] discussed the importance of including non-ED-specific symptoms in networks of ED pathology, such as SAD symptoms.

Eating disorder symptoms in Iran

Although once thought to be an exclusively Western phenomenon, research over the past two decades highlights the severity of ED symptoms and ED symptoms clinical impairment among children, adolescents and college samples in Iran [37,38,39,40,41,42,43,44]. Within the Iranian context, there are some similarities and differences in ED symptoms across sex [41, 42]. For example, adolescent girls had higher dietary restriction compared with boys; however, probable ED symptoms, binge and purge behaviors were comparable across sex [41]. Another study on children and adolescents, aged 6–18, indicated that with exception of other EDs, binge eating disorder (BED) was higher in boys than girls [37]. Given these differences, examining sex differences in an Iranian sample will help clarify unique network dynamics for each group. Moreover, examining bridge symptoms of ED symptoms and social anxiety would be of importance as EDs are highly comorbid with SAD (i.e., 10.13%) in individuals, aged 6–18 [37], suggesting that SAD contributes to EDs in non-Western societies. However, it is unknown how individual SAD and ED symptoms relate to one another in the network approach. Correspondingly, this question is examined for the first time in the current study among Iranian preadolescents.

Current study

The current study used network analysis to identify central ED symptoms among a large sample of Iranian preadolescents ages 9 to 13. We examined central symptoms of both an ED symptom network and a network of co-occurring ED and SAD symptoms. We examined central and bridge symptoms among ED and social anxiety symptoms. Finally, we compared networks between boys vs. girls in the ED and ED/SAD networks and in line with other studies comparing sex differences [31, 45], we postulated that networks structure would be similar across sex. In sum, research on central and bridge (i.e., social anxiety) symptoms of EDs has been conducted among adolescents and adults with EDs diagnosis [28, 29, 36]; however, these findings may not generalize to non-clinical preadolescents. It is important to model such networks in this age group as such research may help identify potential prevention targets. Correspondingly, identifying which specific items will be central and bridge symptoms remains exploratory.

Methods

Participants

Participants (N = 730, grades 3–7) were preadolescent boys (n = 405) and girls (n = 325) who were recruited from 27 classes among 120 classes in a total of eight schools from two cities with different ethnicities, including Tehran (Capital [Persian]) and Tabriz (North-Western [Turkish]). Participation rate was low (22.5%) due to teachers preferring to teach their subject or administer an exam instead of the survey. Preadolescent boys ranged in age from 9 to 13 (M = 10.79, SD = 1.09) and self-reported body mass index (zBMI) ranged from − 2.19 to 4.46 (M = 0.18, SD = 1.13). Adolescent girls’ age ranged from 9 to 13 (M = 11.30, SD = 1.26) and self-reported zBMI ranged from − 2.05 to 2.84 (M = − 0.23, SD = 0.73). School and regional administrators approved the research procedures and parental consent was obtained prior to their child’s participation. The preadolescents were approached during class and given information about the study. Participants provided written informed consent. Correspondingly, those who agreed to participate completed the paper-and-pencil version of the questionnaires in the presence of research staff without financial remuneration. Among boys, one preadolescent (0.2%) did not provide his weight and height to calculate zBMI. Additionally, seven preadolescent boys (1.7%) did not fill out the ChEAT and SASC. Among girls, one (0.3%) did not provide information on weight, two (0.6%) did not provide information on height, and six (1.8%) did not fill out the ChEAT and SASC. The project was approved by the university. Descriptive statistics are provided in Table 1.

Table 1 Means, ranges, and standard deviations for eating disorder and social anxiety disorder symptoms
Full size table

Measures

Demographic characteristics

Participants reported their age, sex, height, and weight (used to derive BMI; kg/m2; [46]). With the use of World Health Organization guidelines among children and adolescents aged 5–19 years, we converted BMI to zBMI [47].

Children eating attitudes test (ChEAT)

The ChEAT [48] is a 26-item questionnaire that measures an individual’s body size concerns, eating habits, and weight control behaviors. Anton and colleagues [49] suggested that the original scoring procedure (i.e., responses to each 6-point item are transformed to a 4-point scale, with the three most pathological responses receiving a score between 1 and 3, and the three least pathological responses receiving a score of 0) may produce low variability in scores for each item as well as the measure, as has been demonstrated among Western and non-Western clinical and non-clinical adolescent samples in the literature [50, 51]. As such, we used Anton and colleagues’ [49] new suggested scoring procedure, where 20 items are scored on the full 1–6 Likert scale with 1 (Never) representing the least and 6 (Very often) representing the most ED pathology. The ChEAT with this new scoring produced six subscales including over concern with body size, dieting, food preoccupation, social pressure to gain weight, vomiting, and caloric awareness and control which was appropriate in non-clinical children in the 2nd through 6th grade, ages 7–13 years [49].

Social anxiety scale for children (SASC)

The SASC [52] has widely been used to assess social phobia in children in the 2nd through 6th grade. The scale has two subscales (i.e., Fear of negative evaluation, Social avoidance and distress). The scale has 10 items ranging from 0 (Never true) to 2 (Always true) and higher scores indicate greater pathology [52].

The current study used a Farsi version of the ChEAT (F-ChEAT) and SASC (F-SASC) for the first time (See Supplementary Materials for translation procedures, as well as factor structure in boys and girls, respectively). Cronbach’s alphas of the F-ChEAT (boys: αs = 0.85; girls: αs = 0.88) and F-SASC (boys: αs = 0.76; girls: αs = 0.77) were adequate to good. Example items are: I am scared about being overweight (i.e., F-ChEAT) and I worry about being teased (i.e., F-SASC).

Data analytic procedure

Item selection

First, we estimated networks in the full sample using all the items from the ChEAT and SASC, however, CS-coefficients for strength centrality and bridge strength indicated the networks were not stable (strength [S] < 0.50; [53]). Networks should include items that represent distinct constructs as to not inflate centrality. In addition, it is important to have a balance of parameters and sample size to ensure network stability [53]. Therefore, we used the goldbricker function in the networktools package in R [54] to identify pairs of nodes that were strongly intercorrelated with one another (r > 0.25). Results from the goldbricker analysis indicated one item from the SASC that should be removed due to high correlations with another item. However, removing this item still resulted in low stability (S < 0.50) for strength centrality and bridge strength. Therefore, we proceeded to reduce network items by combining (averaging) some items based on similar item content. For example, from the ChEAT, item 7 (“I eat diet foods”) and item 9 (“I have been dieting”) were combined, as they both measured dieting behavior. From the SASC, item 5 (I worry about what other kids think of me) and item 8 (“I worry about what other children say about me”) were combined, as they both measured concerns over being judged. For additional information on item combination, please see Table 2.

Table 2 Node abbreviation index
Full size table

Overall models

We estimated two network models: (1) ED items (2) ED and SAD items. We also tested if there were sex differences across these models. We planned to model individual boy and girl networks if we identified sex differences. Therefore, we planned to estimate six networks in total. To estimate the networks, we used the EBICglasso function in the qgraph package in R [55]. The graphical LASSO method was used to create a network of regularized partial correlations between nodes. We tested the stability of the network using the bootnet package in R [53].

Centrality and bridge symptoms

We calculated strength centrality (i.e., the sum of the absolute value of all of a node’s edges) using the centralityplot function in the qgraph package in R [55]. We chose to calculate strength centrality to identify the most central symptoms as it has been found to be the most reliable measurement of centrality [53, 56]. We used centrality difference tests to calculate whether the most central symptoms were significantly different than other symptoms in the network using the bootnet package in R [53]. For the three networks with both SAD and ED symptoms, we also calculated bridge symptoms (i.e., symptoms that are most strongly connected to all the symptoms of a different symptom cluster) using the bridge function in the networktools package in R, allowing us to identify which items are most interconnected across SAD and ED symptoms [33].

Network comparison test

To examine sex differences in the structure of networks, we ran Network Comparison Tests (NCTs) to examine sex differences between girl and boy subsamples for the ED and ED/SAD networks using the NCT package in R [57]. We used the network invariance test (i.e., how the connections between nodes within a network differs across samples) and the global strength invariance test (i.e., how the density of the network – sum of all edge strengths – differs across samples) to quantify network differences.

Results

Networks and stability tests

For strength centrality, CS-coefficients were good (S > 0.50) for all ED and ED/SAD networks except for the boy ED (S = 0.44), the boy ED/SAD (S = 0.36), and the girl ED/SAD (S = 0.28) networks, which were adequate [53]. For bridge strength stability in the ED/SAD networks, CS-coefficients were adequate for the full sample (S = 0.44) and poor for the girl (S = 0.21) and boy (S = 0.13) networks. Since the CS-coefficients for bridge strength were poor (< 0.25) for the girl and boy networks, we do not interpret these results (See Supplementary Materials for bridge symptoms in boys and girls, respectively) and only interpret bridge symptoms in the full sample. See Fig. 1 for all estimated networks and Table 2 for node abbreviations.

Fig. 1
figure 1

A ED Network in the Overall Sample, B ED/SAD Network in the Overall Sample, and C Bridge Strength Coefficients for ED/SAD network in the Overall Sample

Full size image

Central symptoms

See Table 3 for the two most central symptoms and Supplemental Fig. 3 for the strength centrality of all symptoms in each of the networks.

Table 3 Summary of the top two symptoms with highest centrality in the overall sample
Full size table

ED networks

Discomfort eating sweets was the most central symptom in the overall networks, with centrality difference tests indicating this symptom was more central than 64–71% of other symptoms in the network. Driven exercise was the second most central symptom in all of the ED only networks, with greater centrality than 50–64% of other symptoms.

ED/SAD networks

Concern over being judged was the most central symptom across all ED/SAD networks (1.31–1.54), with greater centrality than 43–71% of other symptoms. Discomfort eating sweets was the second most central symptom in the overall ED/SAD networks, with greater centrality than 57–71% of other symptoms.

Bridge symptoms

See Fig. 1 for the bridge strength centrality of all symptoms in the ED/SAD network. In the ED/SAD network, concern over being judged had the strongest bridge strength, followed by engaging in novel activity. See Fig. 2 for the bridge pathways for concern over being judged and engaging in novel activity. Concern over being judged bridged to loss of control over eating (r = 0.06), binge eating (r = 0.05), pressure to eat (r = 0.04), social concern with thinness (r = 0.04), preoccupation with food (r = 0.03), driven exercise (r = 0.03), calorie counting (r = 0.02), desire for empty stomach (r = 0.01), discomfort eating sweets (r = 0.01), and dieting (r = 0.01). Engaging in novel activity bridged to preoccupation with food (r = 0.09), desire for empty stomach (r = 0.09), binge eating (r = 0.06), guilty after eating (r = 0.02), fear of weight gain (r = 0.01), loss of control over eating (r = 0.01), and pressure to eat (r = 0.01).

Fig. 2
figure 2

Bridge Pathways for Two Strongest Bridge Symptoms in ED/SAD network

Full size image

NCTs

The results of the NCT indicated no significant sex differences in ED and ED/SAD networks based on network invariance or global strength, ps > 0.05. Accordingly, we only interpret central symptoms in the full sample. See Supplementary Materials for central symptoms in boys and girls, respectively, though we note that differences should not necessarily be interpreted as significant.

Discussion

The current study used network analysis to conceptualize ED symptoms, as well as ED and SAD symptoms in preadolescent boys and girls ages 9–13 from Iran. We also tested for sex differences in each of our models. Discomfort eating sweets was the most central symptom in the ED symptom network. In the ED/SAD network, concern over being judged was the most central symptom. Concern over being judged was also the strongest bridge symptoms. Finally, no significant sex differences were found in the ED and ED/SAD networks based on network invariance or global strength.

Central ED symptoms

Discomfort eating sweets was the most central symptom and was strongly connected to other ED symptoms in the full sample of preadolescent boys and girls. This finding is in line with prior clinical and non-clinical samples, which also support that there are no significant differences in the structure of the networks between sexes [31, 45], and suggest that the structure of ED symptoms does not differ in Iranian preadolescents. Unlike past ED network studies with clinical samples in children and/or adolescents [28,29,30] and non-clinical samples in adolescents [31], the items assessing overeating, having an empty stomach, asceticism, low-self-esteem, drive for thinness, desire to lose weight, and dissatisfaction with shape and weight were not highly central in this study. Potential reasons for this could be due to differences in methodology, including that all the prior studies examining ED networks used the Eating Disorder Examination Questionnaire (EDE-Q, [58]), Eating Pathology Severity Inventory (EPSI, [59]), and Eating Disorder Inventory-3 (EDI-3, [60]). Further, this is a nonclinical sample of preadolescents.

Discomfort eating sweets may be a risk factor for ED symptoms in preadolescents, and this association may occur through several theoretical pathways. First, in a sample of Iranian children and adolescents, sweets are associated with ED and weight-related problems [61, 62]. For example, eating sweets has been linked to a higher weight in Iranian children [61]. The connection between sweets and weight gain may influence dieting and food avoidance (i.e., of sweets) to reduce this feared connection. Both high BMI and weight are well-established risk factors for ED symptoms in adolescence [63]. Indeed, research supports discomfort eating sweets is a part of the dieting construct among preadolescents, and adolescents prefer to consume less and/or eliminate sweets [51, 64, 65]. Taken together, it can be noted that discomfort eating sweets might stem from diet messages that sweets are not good for preadolescents, and may contribute to ED symptoms [37], such as food avoidance and restriction/dieting, which in turn can lead to binge eating on sweets. Importantly, discomfort eating sweets is easy to assess for target, as well as easily addressable with treatments such as systematic exposure and chaining to increase comfort with eating sweets in appropriate amounts [66, 67]. In line with Christian and colleagues’ findings [28], exercise was another central ED symptom. As expected, having an athletic/muscular body is common in Iran individuals as they have higher or comparable muscularity compared to their counterparts in Western societies [39, 40, 68, 69]. In addition, ED symptoms were associated with athletic/muscular bodies [39]. Another study reported that exercise-related ED was associated with ED symptoms in Iranian individuals [40, 70]. As such, although exercise is not examined in Iranian preadolescents, it makes sense that exercise would contribute to EDs among this group as well. Overall, this finding suggests that screening for and addressing discomfort with eating sweets and exercise may be important for preadolescents in Iran.

Central SAD symptoms in the ED-SAD network

Concern over being judged was the most central symptom, which is aligned with diagnostic criteria for SAD suggesting that fear of negative evaluation is central to the maintenance of SAD [71] and considered as an important component among preadolescents in educational settings [72]. Similarly, the current finding was partially in line with previous studies [73, 74] in which shape-related judgment and fear of judgment due to weight gain were one of the most central ED fears. Additionally, outside of network studies, concern over being judged and/or evaluated would be conceptualized as a cognitive vulnerability or risk factor for SAD [20, 21, 75, 76]. Notably, concern over being judged and/or evaluated is considered a core component of SAD and a feared consequence of social situations [77]. Similarly, Heimberg and colleagues [78] outlined a model of social anxiety in which social anxiety stems from highlighted concern over being judged and/or evaluated in social situations in which the individual may be evaluated. The current study supports this past research and extends these findings in a sample of Iranian preadolescents, suggesting that concern over judgment may confer risk for both SAD and ED development from an early age.

Bridge symptoms

Across ED/SAD networks, concern over being judged had the strongest bridge strength. To our knowledge, only one network analysis study investigated symptoms that may represent illness pathways (i.e., bridge symptoms) from ED symptoms to SAD symptoms in adults with ED/SAD comorbidity [36]. The current study sample and results differed from those of the previous study [36], which could reflect the methodological differences, as Levinson and colleagues’ study [36] utilized different measures in a primarily adult sample. However, our findings were in line with past studies, demonstrating that concern over being judged and/or evaluated could link SAD to ED symptoms among Western societies [20, 79]. Similarly, in cross-sectional, prospective, and intervention studies, concern over being judged and/or evaluated was associated with ED symptoms [15, 19, 80, 81]. The current theoretical models of ED development propose that stress from concern over being judged and/or evaluated may play a pivotal role as a cause of ED symptoms such that concern over being judged and/or evaluated may explain the high occurrence of SAD with ED symptoms [82]. Importantly, a few non-network findings [83,84,85] reported that interpersonal problems, additional bias to rejecting faces, and socio-emotional distress contribute to ED symptoms. These findings may reflect the importance of concern over being judged as the most important node which would be implicated in ED symptoms among preadolescents [14] and adolescents [86]. Thus, given that SAD is characterized by high concerns of being judged (i.e., worrying about what other kids think and say about me), it makes sense that this concern is directly associated with ED symptoms and vice versa. Relatedly, concern over being judged and/or evaluated has been included in a revised model of the dual-pathway model of bulimia nervosa [87], indicating its importance as a transdiagnostic feature in the development and/or maintaining of ED symptoms in adolescents [86]. Theoretically, it may be that because preadolescents are socially anxious about what others think or say these concerns would lead to additional ED symptoms used to attempt to alleviate such concerns. However, it should also be noted that the relationships across SAD and ED symptoms in these networks were weak, compared to within-disorder relationships. This pattern of connections is expected; however, it remains important to examine the clinical utility of the few connections identified between SAD and ED symptoms.

Clinical implications

Network theory predicts that clinical interventions targeted to central symptoms should lead to reductions in other symptoms [23, 24]. Similarly, clinical interventions targeting bridge symptoms should theoretically improve symptoms transdiagnostically [33]. While some evidence supports that centrality corresponds to treatment outcomes [88, 89], item variance is also associated with treatment outcomes [90, 91]. To fully understand the utility of centrality specifically in relation to treatment outcomes, experimental and longitudinal research are needed.

With respect to bridge symptoms, examination of concern over being judged and/or evaluated is a useful target for reducing ED symptoms in Western societies [80] and could be adapted for use in Iran. Thus, experimental studies need to focus on concern over being judged and/or evaluated-related social cues and threats among nonclinical preadolescents from Iran. These results pinpoint that exposure therapy as an efficacious intervention in both SAD [92] and ED symptoms [93, 94] may disrupt the associations between ED and SAD symptoms by focusing on preadolescents’ worries about what other kids think and say. It is also possible that the implementation of treatment aimed at central symptoms with Iranian preadolescents may be efficacious for the development of comorbid SAD and ED symptoms in an at-risk sample.

Strengths and limitations

A notable strength of the current study includes the investigation of how ED and other psychological symptoms interconnect among Middle Eastern preadolescents. This aspect is important considering that most ED theories and research are based on findings among women, and the extent to which these generalize to men and to preadolescents remains understudied.

Several limitations deserve mention. First, we included a non-clinical sample, suggesting that results may not be generalizable to individuals with clinical-threshold ED symptoms. However, other studies have shown that network structures among non-clinical samples and patients with EDs [95, 96] are similar. Therefore this should be tested in future research. Second, like many other network studies, results were cross-sectional. To fully conceptualize how ED symptoms and comorbid symptoms dynamically interact with one another, experimental and longitudinal data are needed. However, other studies have shown that network structures among cross-sectional and longitudinal studies did not differ [29, 35, 97]. Further, we examined two cities with different ethnicities; however, future studies need to examine other ethnicities (i.e., Arab in Ahvaz, Lori in Lorestan) in Iran. Lastly, our assessment of ED symptoms did not include items specifically tailored to males, which may be needed to fully capture the extent of males’ symptoms [98].

Conclusions

This is the first study on ED and SAD networks in a non-clinical sample of preadolescents. In conclusion, we found that discomfort eating sweets were the most central symptom in the ED networks. Additionally, concern over being judged was the strongest bridge symptom. These findings are consistent with non-network studies in Western conceptualizations of ED and identify potential targets for the prevention of ED and SAD symptoms.

What is already known on this subject?

Few studies have assessed ED symptoms and their comorbidities with SAD in any sample and none have been tested in either preadolescents or in Iranian samples. As such, it is unclear to know how ED and SAD symptoms relate to one another in the network approach among preadolescents.

What this study adds?

This study examined core central symptoms of ED and SAD symptoms as well as an examination of bridge symptoms among preadolescents. We found that discomfort eating sweets were the most central symptom in the ED symptom network. In the ED/SAD network, concern over being judged was the most central symptom. Concern over being judged was also the strongest bridge symptoms. This study points us to the idea that future intervention research is needed on discomfort eating sweets and concern over being judged. Such interventions may be efficacious in disrupting co-occurring SAD and ED symptoms in preadolescents at risk for developing with SAD and ED symptoms in Iran.