21st-century skills are a new category of competencies recommended for people to adapt to the digital era. Digital communication skills, particularly, are regarded as an important facet in a progressively moving online society. Empirical evidence of their actual value, however, is largely missing. Recently, the COVID-19 pandemic offered an appropriate testing ground for the possible beneficial role that such skills might play. Our results show that digital communication skills correlate with loneliness. However, social media usage, online contacts, and offline contacts only partially mediate the relationship between digital communication skills and the loneliness levels of students. In addition, we found substantial differences between the two waves of data collection and the types of contacts that influenced loneliness. While skills may have reduced loneliness, the precise mechanism of this effect needs to be clarified more. We offer suggestions for future research to examine the potential benefits of 21st-century skills.

Self-regulated learning has seen a large increase in research interest due to its importance for online learning of higher education students. Several ways to measure self-regulated learning have been suggested. However, most measurements are either obtrusive, necessitating time and effort from students and potentially influencing the learning process, or only partially portable across courses. In the current study, we develop clickstream-based scales of four self-regulated learning phases that we show are portable across courses. The final scales are based on the COPES model and include two strong and reliable dimensions, enactment and adaptation, one dimension that performs reasonably, task definition, and a weaker one, goal-setting. By considering portability as the main criterion in the scale construction process, we ensured reliable transfer to both similar and dissimilar courses. When considering convergent validity, the created scale has higher bivariate and partial correlations with final student grades than the often-used self-reported MSLQ SRL scale. We discuss limitations and future research to improve the scale further and facilitate adoption.