The bidirectional associations between self‐esteem and problematic eating behaviors in adolescents

Abstract

ObjectiveTheories propose that low self‐esteem and problematic eating behaviors (PEBs) negatively impact each other. While previous studies suggested bidirectional associations between self‐esteem and PEBs, they did not separate within‐person from between‐person associations. Therefore, this prospective study investigated the within‐person bidirectional associations between self‐esteem and four PEBs in adolescence, while accounting for between‐person differences.MethodWe used two independent longitudinal samples of Dutch adolescents, each including three annually collected waves of data. Sample 1 consisted of 1856 adolescents (Baseline: 50.4% males; Mage = 13.79 years, SDage = 0.72), with measures of self‐esteem, emotional eating, restrained eating, and loss of control (LOC) while overeating. Sample 2 consisted of 555 adolescents (Baseline: 49.7% males; Mage = 13.13 years, SDage = 0.68), with measures of self‐esteem and LOC eating. The data were analyzed using random intercept cross‐lagged panel models (CLPMs).ResultsWithin persons, lower self‐esteem was associated with higher emotional and restrained eating (both Sample 1) one year later, and vice versa. Self‐esteem did not predict, nor was predicted by, LOC while overeating (Sample 1) or LOC eating (Sample 2). Between persons, self‐esteem was negatively correlated with all PEBs (Samples 1 and 2).DiscussionWe found within‐person bidirectional associations between low self‐esteem and emotional and restrained eating (but not LOC while overeating/LOC eating), and between‐person correlations between low self‐esteem and all PEBs. These results have theoretical and practical implications. Within‐person processes clarify underlying mechanisms that explain the occurrence of PEBs; between‐person associations are important to identify adolescents at risk of PEBs.Public SignificanceWhile theories indicate that low self‐esteem and PEBs are inversely associated within individuals, empirical studies have not disentangled within‐person processes from between‐person differences. This study addressed this disparity, finding that lower self‐esteem was bidirectionally associated with higher emotional and restrained eating (but not LOC eating) within persons. These findings suggest that enhancing self‐esteem is a viable option for prevention and intervention.

Spurious effects in random-intercept cross-lagged panel models: Results from simulations and reanalyses of data on self-esteem and problematic eating behaviors used by Beckers et al. (2023)

“…This paradoxical association was accentuated by a strong auto-correlation between the state factors (Figure 3, Panel A). The same paradoxical association was seen between the within-individual adjusted cross-lagged effect (as Reanalyses of Beckers et al (2023) Our reanalyses with RI-CLPM confirmed the within-individual adjusted cross-lagged effects reported by Beckers et al (corresponding to effects cx and cy in Figure 1, Panel B) between self-esteem and problematic eating behaviors (Table 1, Row 1). However, similarly as in the simulations above, without any direct effects between X and Y scores, cross-lagged correlations between prior and subsequent person-mean centered levels of selfesteem and problematic eating behaviors had the opposite sign compared with the cross-lagged effects in RI-CLPM, i.e.…”

“…and the two random intercepts (RI-X and RI-Y) were freely estimated. Reanalyses of Beckers et al (2023) We refer to Beckers et al (2023) for more comprehensive information on the study sample, used instruments, methodology, etc. In short, data are from the longitudinal Mental Health and Health Habits project, conducted in the Netherlands.…”

“…Withinindividual effects are often assumed to allow stronger causal inference than between-individual effects. Beckers et al (2023) used RI-CLPM and found negative bidirectional cross-lagged effects between within-individual levels of self-esteem and emotional and restrained eating. Beckers et al concluded, using causal language, that their findings suggested that enhancing self-esteem is a viable option for prevention and intervention of problematic eating behaviors.…”

“…Within-individual adjusted crosslagged effects in RI-CLPM may be due to withinindividual correlations with residuals and withinindividual regression to the mean similarly as between-individual adjusted cross-lagged effects in traditional CLPM may be due to between-individual correlations with residuals and between-individual regression to the mean (Sorjonen, Nilsonne, et al, 2023). With these possible biases in mind, the present study had two objectives: (1) To assess bias in RI-CLPM through simulations; (2) To reanalyze data used by Beckers et al (2023) in order to evaluate if the within-individual bidirectional crosslagged effects between self-esteem and emotional and restrained eating may have been spurious rather than genuine.…”

Inconclusive effects between executive functions and symptoms of psychiatric disorders in random-intercept cross-lagged panel models: A simulated reanalysis and comment on Halse et al. (2022)

“…Without this combination of expected effects, observed effects would appear to be spurious and conclusions of causality premature. We Sorjonen, Nilsonne, et al, 2024; have previously used this type of triangulation and challenged conclusions in studies using RI-CLPM and claiming effects of need of cognition on anxiety and depression symptoms (Zainal & Newman, 2022), of income on self-esteem (Bleidorn et al, 2023), of curiosity on creativity and vice versa (Ma & Wei, 2023), and of self-esteem on eating pathology and vice versa (Beckers et al, 2023). We recommend researchers to use similar triangulations to scrutinize findings from analyses of observational data.…”