Linking Trait Items of Self-Control to Broader Conceptualizations in Daily Life Using Machine Learning

Recent research suggests that self-control in everyday life is composed of different components (frequency, intensity, and successful resolution of self-control conflicts), and that individuals seem to differ in their abilities to resolve different types of self-control conflicts (initiation, persistence, inhibition). To date, however, it is unclear how established trait measures capture individual differences regarding these different components and types of self-control. We apply machine learning techniques to experience sampling data (N = 491; 9,634 measurement occasions) and identify which items from various trait scales best predict the different components and types of self-control in daily life. Commonly used trait items indexed all components to some extent, but most strongly successful conflict resolution. Contrary to our expectations, we were unable to predict the frequency, intensity, or successful resolution of inhibition conflicts, despite various trait measures being primarily developed for this conflict type. By comparing the most important predictors, we found meaningful theoretical differences between the components and types. We conclude that applying a broader conceptualization of self-control offers novel opportunities for exploring diverse routes to successful goal attainment and discuss theoretical and methodological implications of the findings for future research on trait self-control.