Dimensions and modulators of behavioural and mental-health change during the Covid-19 pandemic

How has the Covid-19 pandemic affected mental health? What are the most common positives and negatives? How do population variables mediate the impact on mood and behaviour? Who is most at risk of adverse consequences? Which pragmatic measures can help? We address these questions in a data-driven manner by applying multivariate, machine-learning and natural-language processing methods to a survey database collected from 376,987 members of the general public. We report that small average changes in mood from pre-to mid-pandemic obfuscate substantial consequences, both positive and negative, for people from particular sub-populations, vocations, circumstances and personality profiles. The coping strategies that people find helpful during the pandemic are correspondingly diverse yet predictable. We propose that by combining psychological, and demographic variables, it is possible to identify individuals who are at most risk of adverse consequences and to extract individually tailored advice from the collective lived experiences of the general population.