Revival of positive nostalgic music during the first Covid-19 lockdown in the UK: evidence from Spotify streaming data

Abstract

This work shows that positive old music listening surged during the early phase of the Covid-19 pandemic, suggesting a rise in preference for nostalgia and positivity in music. Using the music streaming data of Spotify users in the UK and employing multivariate regression analysis, this work documents that users were more likely to listen to songs older than 5 years during the national lockdown that began in late March 2020 compared with the pre-lockdown period. Such a change in preference was not observed in the same period in 2019. Meanwhile, more frequent listening to old music is found in samples of positive songs and also negative songs. This suggests that the preference for nostalgic music is to a certain extent independent of the positivity bias during the pandemic found in the literature. Yet, this work also provides evidence that the nostalgia-seeking behaviour and the preference for positive songs reinforced each other during the lockdown as the surge in positive old music was more persistent than that in positive recent music.

Keywords: Cultural and media studies; Psychology.

Fig. 1. UK daily number of plays (top 200 only), January–July 2020.

Note: The vertical reference line indicates the beginning of the lockdown on 26 March 2020. Source: Spotify.

Fig. 2. UK daily plays expressed as a proportion of all plays of top 200 songs, January–July 2020.

Note: The vertical reference line indicates the beginning of lockdown on 26 March 2020. Source: Spotify.

Fig. 3. Oxford Stringency Index (UK), January–August 2020.

Source: COVID-19 Government Response Tracker (Hale et al., 2021).

Fig. 4. Bin-scatter plots of regression residuals, January–August 2020.

Notes: Quadratic trends are shown in red curves. The behavioural tendency to listen to old music followed a downward trend before the lockdown was imposed, which is indicated by the dashed vertical line. The preference took an upward turn after the first lockdown day and gradually went up over time. Source: Author’s analysis and illustration of Spotify data.

Fig. 5. Estimated coefficients of the change in trend at 21 supposed break points.

Notes: The blue line plots 21 estimates of the coefficient of the change in preference at a supposed breakpoint, while the green lines refer to the corresponding confidence intervals. The blue line is hump-shaped with the peak, roughly around t = 0. It shows that the change in preference was most marked at around the actual breakpoint (to be precise at t = −1). This result suggests that the lockdown sparked off an upward change in the preference for nostalgic music. Source: Author’s analysis and illustration of Spotify data.

Fig. 6. Estimated coefficients of the change in trend at 21 break points in 2019.

Note: There was no sharp change in preference around t = 0. Source: Author’s analysis and illustration of Spotify data.

Fig. 7. Number of times 10 selected revived songs hit the UK top 200, January–July 2020.

Note: Only 10 of 28 are shown in this graph. Source: Author’s illustration of Spotify data.

Fig. 8. Changes in the total number of plays of four types of songs.

Source: Author’s illustration of Spotify data.

Fig. 9. Log difference in the total number of plays between positive and negative songs.

Source: Author’s illustration of Spotify data.

Fig. 10. Log difference in the total number of plays between old and recent songs.

Source: Author’s illustration of Spotify data.