Anxiety, gender, and social media consumption predict COVID-19 emotional distress

The COVID-19 pandemic caused a global mental health crisis, with significant increases in anxiety and depression reported worldwide. While considerable data exists on various factors potentially linked to COVID-19 distress, it remains unclear which variables are most crucial in predicting who will experience the most severe emotional impact. This study aimed to address this gap by investigating a broad range of socio-psychological factors, including COVID-19 media consumption, demographics, mental health indicators, personality traits, emotion regulation abilities, and COVID-19 knowledge and behaviors. The researchers used a data-driven approach to determine the independent predictive power of each factor in relation to COVID-19 emotional distress.

Prior research on pandemics and mental health reveals some consistent associations. For example, women consistently report more mental health issues and perceived stress related to pandemics than men. Political ideology also appears to play a role, with some studies suggesting that individuals with conservative political perspectives exhibit less fear of the virus. However, the relationship between age and COVID-19 fear is inconsistent across studies. The existing literature highlights the need for a comprehensive analysis accounting for the interdependencies between these various factors to identify the most critical predictors of COVID-19 emotional distress.

Researchers recruited 1000 participants (506 female; mean age = 44.75) via Prolific, an online platform employing representative sampling. After excluding participants based on pre-registered criteria, the final sample comprised 948 participants. Participants completed questionnaires assessing 30 variables across seven categories: COVID-19 media consumption, demographics, mental health (depression, anxiety, alexithymia), personality traits (extraversion, neuroticism, intolerance of uncertainty), emotion regulation, COVID-19 knowledge, and COVID-19 behaviors. Emotional distress was measured using a 15-item scale assessing worry, stress, and time spent ruminating on the coronavirus. A hybrid stepwise regression procedure with 10-fold cross-validation was used to identify the most predictive variables, minimizing cross-validation error and ensuring model robustness. The model was trained on 75% of the data and tested on the remaining 25%.

The cross-validated model revealed that trait anxiety, gender, and social media consumption (but not government media consumption) were the strongest predictors of COVID-19 emotional distress. Anxiety was the most significant predictor, followed by gender (women reporting significantly higher distress), and social media usage from platforms like Facebook and Twitter. While government sources did not independently contribute to distress, social media and television consumption were significantly associated with higher distress levels. Other factors like age, political ideology, neuroticism, and depression showed less unique explanatory power once anxiety was accounted for. The model explained 46% of the variance in COVID-19 emotional distress. The model's calibration (slope close to 1) confirmed its accuracy in predicting real distress ratings.

The findings highlight the crucial role of anxiety in predicting COVID-19 emotional distress, surpassing the influence of other commonly studied variables like neuroticism and depression. The study suggests that the influence of gender and social media may be mediated by anxiety levels, warranting further investigation into the mechanisms underlying these relationships. The differential impact of social media versus government sources of information may point to the importance of information framing and the role of social contagion in shaping emotional responses. The results underscore the need for targeted interventions addressing anxiety and managing social media consumption, particularly in vulnerable populations.

This study identifies anxiety, gender, and social media consumption as critical predictors of COVID-19 emotional distress. Future research should explore the interplay between these factors and investigate the effectiveness of tailored interventions to mitigate pandemic-related distress. Further exploration into information framing and social influence on emotional responses is also warranted.

The study's reliance on self-reported data may introduce bias. The cross-sectional design limits causal inferences, and the sample, although representative of the US population, may not generalize to other contexts. The time frame of data collection (early stages of the pandemic) may limit the generalizability to later phases of the pandemic.