*2.15. Statistical Analysis*

A logistic regression model was fitted predicting behavior from before (0) or during the lockdown (1) with the factors of active engagement, functions of music listening, and attributes of chosen music. For further analyses, a change score for each of these factors was created by subtracting the values from before the lockdown from the ones during the lockdown. Then, eight linear models were fitted with the factors of everyday life, worries, personality, and stress reactivity, predicting each of these change scores. The results are to be interpreted as changes in behavior from before to during the lockdown: In case of a positive estimate, the behavior is currently higher than normal. In case of a negative estimate, the behavior is currently lower than normal.

Similarly, three further linear models were fitted predicting the factors of perceived changes in musical behavior (music value, music use, live music) with the factors of everyday life, worries, personality, and stress reactivity.

#### **3. Results**

#### *3.1. Descriptive Statistics*

In different sections of the questionnaire, participants had to describe effects of the lockdown on their daily lives (job and living situation, changes in everyday life, worries) and on their music-listening behavior.
