Self-efficacy and risky eating behaviour in Chilean students: the mediating role of social anxiety

Introduction

Risky eating behaviours (RAB) represent a growing public health challenge, especially in adolescents and young adults (1). Adolescence is a particularly relevant developmental stage for examining RAB because rapid physical and psychosocial changes coincide with increasing autonomy over food choices, greater concern about body image, and heightened sensitivity to peer evaluation. These conditions may increase vulnerability to maladaptive eating behaviours and make self-efficacy and social anxiety especially relevant factors for early prevention (2). These behaviours are associated not only with physical consequences, such as malnutrition, electrolyte imbalances, and gastrointestinal problems, but also with psychological implications (3). People living with RAB often experience high levels of emotional distress, low self-esteem, depression and anxiety (4). These problems can lead to social isolation, as sufferers often avoid situations where food is present, limiting their interaction with friends and family and exacerbating feelings of loneliness (5). This combination of physical and psychological symptoms can have a significant impact on the quality of life of these people, affecting their performance in daily life, their interpersonal relationships and their emotional wellbeing (6). It is important to note that the long-term consequences can be even more severe, with an increased risk of chronic medical complications, hospitalisation and, in some cases, an increased risk of mortality(7).

Among the psychological factors that influence the development and maintenance of these behaviours, selfefficacy and social anxiety have been reported as under-explored variables of interest that deserve special attention (8-11).

Self-efficacy, defined as a person's belief in his or her ability to perform the actions necessary to achieve a goal (12), has been shown to be inversely related to several problem behaviours, including eatingrelated behaviours such as food restriction, binge eating and emotional eating (10, 13). In this context, the perception of control favours individual performance, allowing those who believe in their capacity to generate positive events in their lives to adopt a more active and autonomous attitude (14). The relevance of such beliefs lies in their impact on personal motivations, the level of commitment to task performance and the ability to tolerate frustration (15). People who are confident in their ability to take specific problem-solving actions tend to show greater commitment to those actions (16). Consequently, a high sense of competence facilitates cognitive processes and optimises performance in different domains, constituting a key motivational factor that drives more effective decision-making, even in the face of adverse conditions. Individuals with low self-efficacy may experience greater vulnerability to developing RAB, as they tend to feel that they cannot adequately control their behaviour in stressful situations (13). Quasi-experimental studies in children have reported that increased knowledge about healthy eating and increased selfefficacy are significant predictors of future eating behaviours (17).

On the other hand, social anxiety, characterised by a persistent fear of situations in which the person may be negatively evaluated by others, has frequently been linked to the development of RAB (18). Social embarrassment, fear of ridicule and negative selfevaluation can lead people to develop maladaptive behaviours, such as excessive control of eating, as a way of coping with anxiety (19).

In children, social anxiety together with punishment, parental nutritional knowledge, pressure to eat, family function and the child's perception of weight have been reported to have a large effect on the eating behaviour of overweight or obese children(20).

The aim of the present research is to analyse the mediating role of social anxiety on the effect of selfefficacy on risky eating behaviours in students from three cohorts. It is hypothesised that social anxiety significantly mediates the negative effect of selfefficacy on risky eating behaviours in students from the three cohorts, such that higher self-efficacy is associated with lower levels of social anxiety, which in turn is related to a reduction in risky eating behaviours. The use of three independent cohorts allows for the evaluation of the replicability of the proposed mediation model, rather than a multigroup comparison.

Materials and methods

Design

Three cross-sectional studies (21) were conducted in three different years with three samples of adolescent primary and secondary school students. In all three cases the sampling was purposive based on the accessibility of the participants.

Participants Study 1

The first sample consisted of 206 participants, of whom 115 (55.8%) were female and 91 (44.2%) males, aged between 12 and 18 years (M=14.39; SD=1.66). Of the participants, 87.4% (n=180) were Chilean and 12.6% (n=26) were foreigners. The sample consisted of participants from nine educational institutions. Of the total, 22.8% (n=47) attended public schools, while 77.2% (n=159) came from private subsidized (state-funded private) schools.

Study 2

The second sample consisted of 301 participants, of whom 167 (55.5%) were female and 134 (44.5%) males, aged between 12 and 18 years (M=15.46; SD=1.17). Of the participants, 90.4% (n=272) were Chilean and 9.6% (n=29) were foreigners. The sample consisted of participants from seven educational institutions. Of the total, 75.1% (n=226) attended public schools, 19.3% (n=58) attended subsidised private schools and 5.6% (n=17) attended private schools.

Study 3

The third sample consisted of 321 participants, of whom 189 (58.9%) were female and 132 (41.1%) males, aged between 12 and 19 years (M=14.87; SD=1.72). Of the participants, 85.9% (n=276) were Chilean, 8.7% (n=28) were foreigners, and 5.3% (n=17) did not provide this information. The sample consisted of participants from eight educational institutions. Of the total, 50.8% (n=163) attended public schools, 31.2% (n=100) private subsidised schools and 18.1% (n=58) private paid schools.

Instruments

Self-efficacy. To measure this variable, the General Selfefficacy Scale of Jerusalem and Schwarzer (22) was used in its Spanish adapted version of 10 items (23-24) which has presented good psychometric properties in Chilean population in both adolescents and adults (15). The scale is presented with four response options ranging from ‘Incorrect’ [1] to ‘True’ [4]. Higher scores reflect higher levels of self-efficacy (e.g., ‘I can find a way to get what I want, even if someone opposes me’). In all three studies, the scale scores had adequate Cronbach's Alpha coefficients (S1=.91; S2=.91; S3=.89).

Social Anxiety. To measure this variable, the Social Anxiety Scale of the questionnaire called Sistema de Evaluación de Niños y Adolescentes (SENA) (25) was used. The SENA self-report version has shown adequate psychometric properties in Chilean adolescents (30). This scale consists of eight items, with five response options ranging from   N Hun et al.

‘Never or almost never’ [1] to ‘Always or almost always’ [5]. Higher scores reflect higher levels of Social Anxiety (e.g., ‘I get nervous when there are a lot of people around’). In all three studies, the scale scores had adequate Cronbach's

Alpha coefficients (S1=.88; S2=.87; S3=.85).

Risky Eating Behaviours. To measure this variable, the Eating Behaviour Problems Scale of the SENA questionnaire (25) was used. This scale consists of eight items, with five response options ranging from ‘Never or almost never’ [1] to ‘Always or almost always’ [5]. Higher scores reflect higher Risky Eating Behaviours (e.g., ‘When I start eating, I feel like I can't stop’). In all three studies, the scale scores had adequate Cronbach's Alpha coefficients (S1=.89; S2=.85; S3=.84).

Procedure

This study focused exclusively on the psychosocial pathway linking self-efficacy, social anxiety and risky eating behaviours; therefore, anthropometric and nutritional variables were not collected.

This study is part of a larger project on mental health in educational institutions in Arica, northern Chile. The project collects data annually, and data for the years 2021, 2022 and 2023 were used for this work.

First, an invitation was sent to primary and secondary schools in the city to participate in the study. Once an institution confirms its participation, coordination takes place with representatives of each school to schedule the dates of implementation. The project is then presented to parents and guardians, explaining in detail the purpose and scope of the project. Those who decide to participate must sign an informed consent form. In addition, informed consent is requested from the students themselves.

When students confirm their participation, the questionnaires are applied in their classrooms, with the presence of at least two interviewers, who are available to answer any questions. A teacher from the school also accompanies the process to ensure a suitable environment.

During the application of the questionnaires, absolute silence is requested to avoid possible distractions that could affect the results. The evaluation lasts approximately 30 minutes.

The instruments and the procedure were reviewed and approved by the ethics committee of the university to which the first author belongs. The Scientific Ethical Committee issued its approval certificate N°17/2021 on 2 June 2021.

Inclusion criteria were: (a) being enrolled in primary or secondary education, (b) having signed informed consent from a parent or legal guardian, (c) providing personal assent, and (d) being present on the day of assessment. Students who did not meet all inclusion criteria or who did not complete the questionnaires were excluded.

Data Analysis

To assess the consistency of the proposed model, three independent studies (S1, S2 and S3) were conducted in which the mediating effect of Social Anxiety on the relationship between Self-efficacy and Risky Eating Behaviours in secondary school students was analysed. For this purpose, structural equation models were used, including age and gender as control variables. The goodness of fit of the models was assessed using the Chisquare (_χ²), root mean squared error of approximation (RMSEA), comparative fit index (CFI) and Tucker-Lewis index (TLI). According to standards suggested by the literature (26), values of RMSEA ≤ .08, CFI ≥ .95 and TLI ≥ .95 are considered indicators of a good fit. The weighted least squares robust least squares (WLSMV) estimation method, suitable for non-normal ordinal variables (27), was employed. Analyses were performed using the statistical packages SPSS v. 24 and Mplus v. 8.9. 

Results

Descriptive statistics

Table 1 presents the descriptive statistics and correlations between the variables considered in the models.

Mediation models

Study 1. The model of the first study presented excellent goodness-of-fit indicators according to the criteria recommended by the literature (see table 2) (26).

As in the first study, Self-efficacy has a statistically significant negative direct effect of moderate magnitude on Social Anxiety, and of slight magnitude (β=−.203) (28) on Risky Eating Behaviours. Social Anxiety has a statistically significant positive direct effect of large magnitude (β=.470) (28) on Risky Eating Behaviours (Figure 2).

As for the indirect effect, as in the first study, Social Anxiety has a statistically significant mediating effect on the relationship between Self-efficacy and Risky Eating Behaviours (indirect effect=−.175, p=.000).

Study 3. Finally, the model was evaluated in a third study, whose goodness-of-fit indicators were close to those recommended in the literature (see Table 2). Although these results were not as favourable as those obtained in the first and second studies, they are still within acceptable criteria (29).

Consistent with previous studies, Self-efficacy has a statistically significant negative direct effect of moderate magnitude on Social Anxiety, and of slight magnitude on Risky Eating Behaviours. Likewise, Social Anxiety has a statistically significant positive direct effect of moderate magnitude on Risky Eating Behaviours (Figure 3).

Finally, as in the first and second studies, Socia Anxiety has a statistically significant mediating effect on the relationship between Selfefficacy and Risky Eating Behaviours (indirect effect=−.124, p=.000).

Based on the results of the three independent studies, the proposed mediation model appears to be an adequate representation of the observed relationships. This suggests that secondary school students with higher levels of Self-efficacy tend to show lower levels of Social Anxiety, which in turn is associated with a decrease in Risky Eating Behaviours. In this context, Social Anxiety would act as an explanatory factor in the relationship between Self-efficacy and Risky Eating Behaviours in secondary school students.

Discussion

The results obtained in the three studies conducted in cohorts of secondary school students in Chile reveal a significant relationship between self-efficacy, social anxiety and risky eating behaviours (RAB). Consistent with previous literature, it was found that higher selfefficacy is associated with lower levels of social anxiety, which in turn is related to a reduction in RABs. This finding reinforces the suggestions of Barbee & Timmerman (8), who highlight the importance of self-efficacy as a protective factor against RABs. Furthermore, the results are consistent with reports by Oikarinen et al. (13) and Lazarevich et al. (11), where it is observed that people with higher self-efficacy tend to cope better with stressful situations, thus reducing the likelihood of losing control over their eating behaviour, particularly in highpressure contexts.

In terms of social anxiety, the results are in line with the findings of studies such as those of Zhu et al. (20) and Sedova et al. (18), which indicate that social anxiety can exert a considerable influence on the development of RAB, particularly in contexts where young people experience social pressure or fear of judgement. In this regard, fear of ridicule or embarrassment in food-related social situations may lead adolescents to develop restrictive or uncontrolled behaviours as an attempt to manage such anxiety (19). Thus, our findings reinforce the idea that social anxiety acts as a key mediator between self-efficacy and RABs, as hypothesised.

The magnitude of the effects found in the three studies (_β between -.10 and -.30 for self-efficacy and social anxiety, respectively) is consistent with previous studies, such as those of Pigsborg et al. (10), which show that social anxiety has a significant impact on adolescents' daily lives, directly affecting their eating behaviour. Although the relationship between self-efficacy and RAB was more tenuous, it is still significant and consistent with previous studies showing that adolescents with low self-efficacy tend to lose control over their eating habits when faced with stressful situations (13).

From a practical perspective, these findings suggest that interventions aimed at improving students' self-efficacy may be effective not only in reducing social anxiety, but also in reducing the prevalence of RAB. Programmes designed to strengthen adolescents' confidence in their ability to manage stressful situations could have beneficial effects on both their emotional well-being and their eating behaviours (16). Specifically, as mentioned by Kulik et al. (17), school-based interventions that combine nutrition education with strategies to improve selfefficacy may be particularly effective in preventing RAB, especially in contexts where adolescents feel vulnerable to social judgement.

Although the results provide strong evidence for the mediating role of social anxiety in the relationship between self-efficacy and RABs, there are several limitations that need to be addressed in future research. First, the cross-sectional design used precludes establishing clear causal relationships between the variables. Therefore, we suggest that future studies employ longitudinal designs to assess how self-efficacy and social anxiety interact over time in the development of RABs.

Furthermore, the sample was composed exclusively of secondary school students from a specific region of Chile, which limits the generalisability of the findings. In addition, although nutritional status was not part of the original study objectives, anthropometric data were not collected; therefore, it was not possible to examine whether the observed relationships varied according to nutritional status. Future studies should replicate the model in other regions and countries and incorporate anthropometric indicators and contextual variables such as the family environment, which plays an important role in the formation of eating habits and the regulation of anxiety among young people (20).

Finally, we suggest that future research should explore how self-efficacy may differentially influence subgroups of adolescents, considering factors such as socioeconomic status, access to healthy eating programmes, and the presence of comorbid psychological disorders. These studies could provide valuable information for the design of more personalised and effective interventions to prevent risky eating behaviours.

Conclusions

Across three independent cohorts of Chilean students, the proposed mediation model showed a consistent pattern: higher self-efficacy was associated with lower social anxiety, which, in turn, was related to fewer risky eating behaviours. The convergence of these results across samples collected in different years provides empirical support for the replicability of the model in this population and constitutes the main contribution of the study. Although the cross-sectional design precludes causal interpretations, these findings provide a basis for further longitudinal and crosscultural evaluations of the model.

Acknowledgments

The authors would like to thank all participants for their valuable time and willingness to contribute to this study. We also express our sincere appreciation to the participating institutions for their support and collaboration throughout the research process. This research was made possible thanks to funding from ANID PIA CIE160007.

Conflict of Interest Statement

The authors declare that there are no conflicts of interest related to the conduct, analysis, or publication of this study.

CRediT authorship contribution statement

NH: conceptualisation, methodology, writing - original draft, writing – review and editing, corresponding author.

DH: conceptualization, methodology, formal analysis, data curation, writing - original draft, writing - review and editing.

ACU: conceptualization, funding acquisition, supervision, writing – review and editing, project administration, resources.

Referenes

1. Jebeile H, Lister NB, Baur LA, Garnett SP, Paxton SJ. Eating disorder risk in adolescents with obesity. Obes Rev. 2021;22(5):e13173. https://doi.org/10.1111/obr.13173
2. Neufeld LM, Andrade EB, Ballonoff Suleiman A, Barker M, Beal T, Blum LS, et al. Food choice in transition: adolescent autonomy, agency, and the food environment. Lancet. 2022;399(10320):185-97. https://doi.org/10.1016/S0140-6736(21)01687-1
3. Lin JA, Jhe G, Vitagliano J, Milliren C, Spigel R, Woods E, et al. The association of malnutrition, illness duration, and pre-morbid weight status with anxiety and depression symptoms in adolescents and young adults with restrictive eating disorders: a cross-sectional study. J Eat Disord. 2021;9(1):60. https://doi.org/10.1186/s40337-021-00415-7
4. Aloi M, Segura-García C. Anxiety and depression mediate the role of low self-esteem and low self-directedness in the development of eating disorders. Int J Adolesc Med Health. 2019;31(3):20160084. https://doi.org/10.1515/ijamh-2016-0084
5. Hovrud L, Simons R, Simons J. Cognitive schemas and eating disorder risk: the role of distress tolerance. Int J Cogn Ther. 2020;13:5466. https://doi.org/10.1007/s41811-019-00055-5
6. Okonishnikova E, Bryukhin A, Lineva T, Belokrylov I. Dysmorphophobia as a factor that worsens the affective state and the life quality of patients with eating disorders: the final data of the study. Eur Psychiatry. 2021;64(S1):S356. https://doi.org/10.1192/j.eurpsy.2021.954
7. van Hoeken D, Hoek HW. Review of the burden of eating disorders: mortality, disability, costs, quality of life, and family burden. Curr Opin Psychiatry. 2020;33(6):521-7. https://doi.org/10.1097/YCO.0000000000000641
8. Barbee KG, Timmerman GM. Emotional eating, nonpurge binge eating, and selfefficacy in healthy perimenopausal women. J Holist Nurs. 2015;33(4):298-307. https://doi.org/10.1177/0898010115569574
9. Gómez-Peresmitré G, Platas-Acevedo RS, León-Hernández R, Pineda-García G, GuzmánSaldaña R. Psychometric assessment of an online self-test measuring risky eating behavior, depression, social anxiety, and selfinjury in Mexican adolescents. Int J Environ Res Public Health. 2022;20(1):399. https://doi.org/10.3390/ijerph20010399
10. Pigsborg K, Kalea AZ, De Dominicis S, Magkos F. Behavioral and psychological factors affecting weight loss success. Curr Obes Rep. 2023;12(3):223-230. https://doi.org/10.1007/s13679-023-00511-6
11. Lazarevich I, Irigoyen Camacho ME, VelázquezAlva MDC, Zepeda Zepeda M. Relationship among obesity, depression, and emotional eating in young adults. Appetite. 2016;107:639-644. https://doi.org/10.1016/j.appet.2016.09.011
12. Bandura A. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev. 1977;84(2):191-215. https://doi.org/10.1037/0033-295X.84.2.191
13. Oikarinen N, Jokelainen T, Heikkilä L, Nurkkala M, Hukkanen J, Salonurmi T, et al. Low eating self-efficacy is associated with unfavorable eating behavior tendencies among individuals with overweight and obesity. Sci Rep. 2023;13(1):7730. https://doi.org/10.1038/s41598-023-34513-0
14. Ruiz-González L, Videra A, Moreno-Murcia JA. Predictive power of task orientation, general self-efficacy and self-determined motivation on fun and boredom. Motriz Rev Educ Fis. 2015;21(4):361-369. https://doi.org/10.1590/S1980-65742015000400004
15. Clavijo M, Yévenes F, Gallardo I, Contreras AM, Santos C. Escala de autoeficacia general: reevaluación de su evidencia de confiabilidad y validez en Chile. Rev Med Chil. 2020;148(10):1452-1460. https://doi.org/10.4067/S0034-98872020001001452
16. Usher EL, Butz AR, Chen XY, Ford CJ, Han J, Mamaril NA, et al. Supporting self-efficacy development from primary school to the professions: a guide for educators. Theory Pract. 2023;62(3):266-278. https://doi.org/10.1080/00405841.2023.2226559
17. Kulik NL, Moore EW, Centeio EE, Garn AC, Martin JJ, Shen B, et al. Knowledge, attitudes, self-efficacy, and healthy eating behavior among children: results from the Building Healthy Communities Trial. Health Educ Behav. 2019;46(4):602-611. https://doi.org/10.1177/1090198119826298
18. Sedova E, Kalina S, Gardanova Z. Negative attitude towards the appearance: connection with eating behavior and social anxiety. Eur Psychiatry. 2021;64(S1):S602. https://doi.org/10.1192/j.eurpsy.2021.1606
19. Anderson L, Wong N, Lanciers S, Lim C. The relative importance of social anxiety facets on disordered eating in pediatric obesity. Eat Weight Disord. 2020;25:117-126. https://doi.org/10.1007/s40519-018-0526-x
20. Zhu H, Zhao K, Huang L, Shi W, Tang C, Xu T, et al. Individual, family and social-related factors of eating behavior among Chinese children with overweight or obesity from the perspective of family system. Front Pediatr. 2024; 12:1305770. https://doi.org/10.3389/fped.2024.1305770
21. Ato M, López-García JJ, Benavente A. Un sistema de clasificación de los diseños de investigación en psicología. An Psicol. 2013;29(3):1038-1059. https://doi.org/10.6018/analesps.29.3.178511
22. Jerusalem M, Schwarzer R. Self-efficacy as a resource factor in stress appraisal processes. In: Schwarzer R, editor. Self-efficacy: thought control of action. 1992 (195213). Washington (DC): Hemisphere. Baessler J, Schwarzer R. Evaluación de la autoeficacia: adaptación española de la Escala de Autoeficacia General. Ansiedad Estrés. 1996;2(1):1-8.
23. Schwarzer R, Bäßler J, Kwiatek P, Schröder K, Zhang JX. The assessment of optimistic self-beliefs: comparison of the German, Spanish, and Chinese versions of the General Self-Efficacy Scale. Appl Psychol. 1997;46(1):69-88. https://doi.org/10.1111/j.1464-0597.1997.tb01096.x
24. Fernández-Pinto I, Santamaría P, Sánchez-Sánchez F, Carrasco M, Del Barrio V. Sistema de Evaluación de Niños y Adolescentes (SENA): manual técnico. Madrid: TEA Ediciones; 2016.
25. Schreiber JB. Update to core reporting practices in structural equation modeling. Res Social Adm Pharm. 2017;13(3):634-643. https://doi.org/10.1016/j.sapharm.2016.06.006
26. Beauducel A, Herzberg PY. On the performance of maximum likelihood versus means and variance adjusted weighted least squares estimation in CFA.
27. Struct Equ Modeling. 2006;13(2):186-203. https://doi.org/10.1207/s15328007sem1302_2
28. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale (NJ): Lawrence Erlbaum Associates; 1988.
29. Hu LT, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling. 1999;6(1):155. https://doi.org/10.1080/10705519909540118
30. Flores J, Caqueo-Urízar A, Ramírez C. Propiedades psicométricas de las versiones de Autoinforme del Sistema de Evaluación de Niños y Adolescentes (SENA) en una muestra de Chile. Rev Iberoam Diagn Eval Psicol. 2022;64(3):171-183. https://doi.org/10.21865/RIDEP64.3.13