Longitudinal trait and state-like differences in the components model of addiction: An illustration through social media addiction and work addiction.

Background and aims: Conflicting findings have been reported for the longitudinal course of behavioral addictions, especially for social media addiction (SMA) and work addiction (WA). Therefore, evaluating whether these constructs are more trait-like or state-like might be informative. The aim of the present study was to examine the proportion of variance of SMA and WA symptoms (as defined by the components model of addiction) explained by trait and occasion-specific factors in addition to exploring cross-lagged relationships between SMA and WA. Methods: Young adults from a representative sample who continuously used social media and worked at least 40 hours a week during the first three waves of the Budapest Longitudinal Study were included (N = 1,551; Females: 50.6%; Age: M = 27.7 years [SD = 4.40]). The Bergen Social Media Addiction Scale and the Bergen Work Addiction Scale were administered in all three waves. Results: A latent state-trait model with a general trait factor was considered for both SMA and WA. Symptomatic variability in SMA was explained approximately equally by trait and state-like factors, while WA-related symptom variability was mostly attributed to state-like factors. SMA negatively predicted WA over time, while WA showed a positive cross-lagged effect on SMA. Discussion and Conclusions: While the symptoms of WA were more state-like, the trait-like effects were stronger in SMA. Situational influences and previous symptom severities might have to be considered in the screening process.