Abstract
Purpose: Internal shame (IS) is a deeply distressing emotional state characterised by avoidance tendencies, diminished self-esteem, and negative self-perception, often accompanied by self-critical thoughts such as self-loathing. It significantly influences an individual's cognitive and emotional processes, shaping self-evaluation and overall functioning. Recent studies have highlighted its prominence as a construct with multifaceted psychological implications, impacted by various psychosocial and emotional factors. This study focuses on examining the determinants of internal shame, including emotional competency, childhood trauma, alexithymia, cognitive flexibility, and distress tolerance, utilising the Explainable Shrinkage Regression Model.
Method: A total of 906 participants (53.1% female and 46.9% male) residing in Tehran were recruited in 2023 through convenience sampling on online platforms. The survey encompassed demographic variables (age and gender) and psychological assessments, such as measures of childhood trauma, social-emotional competence, internal shame, cognitive flexibility, and distress tolerance. Data analysis employed regression techniques, including Linear Regression, Ridge Regression, Least Absolute Shrinkage and Selection Operator (LASSO), and Elastic Net Shrinkage Regression models. The independent variables included five features: childhood trauma, social-emotional competence, alexithymia, cognitive flexibility, and distress tolerance. The dependent variable was internal shame.
Results: The results indicated that while each model presents its own advantages, Elastic Net regression showed marginally the best performance in prediction of IS having the lowest MSE and highest R2.
