Exploring the Associations between Media and Instagram Interaction Patterns with Weight Bias among Undergraduate Nutrition Students in the Brazilian Nutritionists' Health Study

Abstract

This study examined the association between media and Instagram interaction patterns with weight bias among undergraduate nutrition students in the Brazilian Nutritionists' Health Study. We also explored the potential mediating role of students' own body image perception in these relationships. A total of 406 students (78% women) participated in this cross-sectional analysis. Sociodemographic data, media influence, Instagram interaction patterns, body image perception, and weight bias were assessed using semi-structured questionnaires. Findings indicated that exposure to fitness content on Instagram (β = 0.17, p < 0.001) and the pursuit of an ideal athletic body (β = 0.12, p = 0.034) were associated with increased weight bias. In contrast, engagement with body diversity content (β = -0.23, p < 0.001) and perceived pressure from media to conform to appearance ideals (β = -0.24, p < 0.001) had a mitigating effect on weight bias. Notably, body image perception did not mediate these relationships (p > 0.05). In conclusion, this study revealed a link between media exposure and weight bias among undergraduate nutrition students, independent of their body image perception. Developing social media literacy programs that encourage students to critically evaluate media content is imperative to reduce weight bias. Additionally, a deeper examination of the media content that contributes to weight bias and the potential need for targeted regulatory measures is warranted.

Keywords: Brazil; SATAQ-3; body image; nutrition students; social media; weight bias.

Figure 1

Structural equation model using the internalization of media appearance ideals (i.e., Internalization-Athlete and Pressures) and Instagram interaction patterns (i.e., Body Diversity, Fitness, and High Engagement) as latent variable and predictors of weight bias, and body image perception as mediating variable. Note: The model was adjusted for sex and age. Results are presented as standard estimates. * p < 0.05; ** p < 0.001. Only statistically significant associations are displayed in the figure. The goodness-of-fit indices indicated an acceptable fit of the SEM (Comparative Fit Index of 1.00, Tucker–Lewis Index of 0.83 and root-mean-square error of approximation, RMSEA of zero (CI 90%: 0, 0.078)).