High self-esteem, a positive self-evaluation, is crucial for mental health, while low self-esteem is linked to various psychiatric conditions. Prior research showed that low self-esteem individuals differ in how they construct beliefs about self-worth from social feedback. However, it's unclear whether this extends to non-social contexts. This study examined self-performance estimates in individuals with high and low self-esteem, both with and without external feedback. The researchers hypothesized that low self-esteem individuals would underestimate their abilities even when their objective performance was comparable to those with high self-esteem, potentially reflecting a disconnect between local (task-specific) and global (self-worth) self-evaluations. The study used a novel behavioral paradigm that measured subjective self-performance estimates in the absence of feedback, reflecting many real-life situations, to assess the relationship between self-esteem (a global self-evaluation) and self-performance estimates formed over short timeframes. A large dataset of adolescents and emerging adults was used, selecting participants from the top and bottom 10% of the self-esteem distribution to maximize power. This design allows for investigating the neurocognitive mechanisms underlying self-esteem and potentially informing novel interventions for disorders linked to altered self-esteem, such as depression.

The introduction thoroughly reviews existing literature on the importance of self-esteem for mental health, its association with various psychiatric conditions, and the limited understanding of the cognitive mechanisms underlying self-belief formation. Previous work on the construction of momentary self-worth from social feedback in low self-esteem individuals is highlighted, showing slower updating of beliefs about others' liking and faster updating of self-worth in response to feedback. The study builds upon previous work that demonstrated the importance of decision confidence in self-performance estimates, especially in the absence of feedback. A hierarchical framework for metacognitive evaluation is proposed, where self-esteem acts as a global prior influencing self-performance estimates on specific tasks. This existing research sets the stage for the current investigation into the relationship between global self-esteem and local self-performance estimates.

Sixty-two participants (later reduced to 57 after exclusions) were selected from a larger community sample (N=2402) based on their Rosenberg Self-Esteem Scale (RSES) scores, representing the top and bottom 10% of the distribution. Participants (aged 18-25) completed a perceptual decision-making task involving two interleaved tasks with varying difficulty levels (easy or difficult) and feedback conditions (with or without feedback). The tasks involved judging which of two boxes contained more dots. After each block of trials (2-10 trials per task), participants provided two measures of self-performance estimates: (1) a task choice (selecting the task they believed they performed best on), and (2) subjective task ratings (rating their performance on a 50-100% scale). The study used a 2 x 2 x 2 repeated-measures ANOVA to analyze objective performance, task choices, and task ratings. Additional regression analyses examined the influence of block-to-block performance fluctuations and learning duration on self-performance estimates, considering self-esteem as a predictor. The methodology meticulously details the participant selection, task design, data collection procedures, and statistical analyses, ensuring replicability and rigor.

Objective task performance did not significantly differ between high and low self-esteem groups across all conditions. However, low self-esteem participants consistently provided lower subjective task ratings compared to high self-esteem participants (main effect of Self-Esteem, F(1, 56) = 5.92, p = 0.018). This difference was observed across both easy and difficult tasks and with and without feedback, highlighting a consistent underestimation of performance in the low self-esteem group. Replicating previous findings, task performance was better in easier tasks and with feedback (main effects of Difficulty and Feedback), with an interaction between these factors. Task choices reflected sensitivity to task difficulty and feedback but were not significantly influenced by self-esteem. Regression analyses revealed that task choices and ratings were sensitive to fluctuations in objective performance, irrespective of self-esteem level. Learning duration (number of trials per task) had minimal impact on self-performance estimates, suggesting rapid formation of performance expectations. Analyses using past self-esteem scores (at recruitment) yielded virtually identical results, indicating the robustness of the findings.

The findings support the hypothesis that low self-esteem is associated with a bias in self-performance estimation, independent of objective performance. The lack of performance differences between groups across all experimental conditions strengthens the conclusion that this underestimation is a metacognitive bias rather than a consequence of actual performance deficits. The insensitivity of self-performance estimates to learning duration suggests that participants quickly form an initial expectation of success that is not easily updated based on subsequent experience. The limited influence of feedback and task difficulty on the interaction between self-esteem and performance estimation further underscores the domain-general nature of this bias. This contrasts with findings suggesting low self-esteem individuals are more sensitive to feedback in other contexts. The study's limitations, such as the limited age range and potential influence of co-morbid anxiety or depression, are acknowledged. The study's strength is its use of a novel task minimizing prior beliefs and social influences, allowing for clearer isolation of self-esteem's effects.

This study demonstrates a consistent underestimation of performance in individuals with low self-esteem, independent of objective performance. This bias appears to be domain-general, affecting self-evaluation across various task conditions. Future research could explore the temporal dynamics of global self-performance estimates and the interplay between local and global metacognitive processes across the lifespan. Investigating potential non-linear relationships between self-esteem levels and self-performance estimates could further refine our understanding.

The study's limitations include the relatively narrow age range of participants and the potential confounding effects of co-morbid conditions like anxiety and depression, which are frequently associated with low self-esteem. The use of a specific perceptual task might not generalize perfectly to all contexts of self-evaluation, and the limited number of participants in each self-esteem group could reduce statistical power. Further research with a larger and more diverse sample and additional measures of metacognitive processes would strengthen the conclusions.