A new analysis of popular music from the last half-century reveals that while the lyrics of chart-topping American songs have generally become angrier and simpler over time, this trend shifts unexpectedly during major societal crises. The findings suggest that during events like the COVID-19 pandemic, listeners may seek out songs that offer an emotional escape rather than those that mirror their anxieties. This research was published in Scientific Reports.
Art has historically been viewed as a reflection of the society that creates it. Philosophers and scientists alike argue that cultural artifacts can reveal collective psychological states that are difficult to measure through individual surveys alone.
Music is particularly suited for this type of historical analysis because it is widely consumed and contains explicit emotional messages within its lyrics. The consumption of music provides a unique window into the emotional needs of the population at a specific moment in time.
Markus Foramitti and Maurício Martins led the research team at the University of Vienna. They collaborated with Urs Markus Nater and Claus Lamm. The team sought to understand how major societal shocks influence the music people choose to consume.
They specifically investigated whether listeners gravitate toward songs that reflect their stress or if they prefer music that helps them regulate their emotions. This distinction is significant for understanding how populations cope with collective trauma.
The researchers analyzed the weekly top 100 songs from the Billboard Hot 100 chart between 1973 and 2023. This dataset included over 20,000 unique English-language songs. The team chose the Billboard charts because they reflect what people are actively listening to rather than just what is being produced by artists. This approach captures the preferences of the audience.
The team used natural language processing to evaluate the text of each song. They employed a variety of computational tools to assess different linguistic features. To measure the frequency of stress-related words, they used a dictionary-based program called Linguistic Inquiry and Word Count. This tool scans text for specific words associated with anxiety, pressure, and struggle.
They estimated the emotional tone of the lyrics using a sentiment analysis tool known as VADER. This program is designed to handle the nuances of informal language often found in social media and lyrics. It assigns a score based on how positive or negative the language appears.
To measure lyrical complexity, the study employed a data compression algorithm. This method relies on the principle that repetitive or simple text can be compressed more efficiently than complex or varied text. A higher compression rate indicated simpler lyrics.
The researchers found that the frequency of stress-related language in popular lyrics increased over the 50-year period. At the same time, the sentiment of the songs became increasingly negative. The structural complexity of the lyrics also declined over time. This indicates that the music rising to the top of the charts has become simpler, more repetitive, and more emotionally negative since the 1970s. These findings align with broader observations of rising stress and mental health issues in the general population.
The team then examined how these trends changed around two major events. These events were the terrorist attacks on September 11, 2001, and the onset of the COVID-19 pandemic in early 2020. The data did not support the hypothesis that stress in lyrics would spike during these times. The trends observed during these crises defied the expectation that art consumption would simply mirror societal distress.
The onset of the COVID-19 pandemic coincided with a decline in stress-related language in the most popular songs. It also coincided with an increase in positive sentiment in the charts. This effect suggests that listeners preferred music that did not match the prevailing societal mood. This behavior points to a desire for emotional relief or distraction. The analysis of the period surrounding 9/11 showed a similar but less robust pattern, where the rise of negative language appeared to be attenuated rather than amplified.
The researchers also looked for links between economic factors and musical preferences. They compared the lyrical data to the annual growth of real median household income. They found no significant relationship between income growth and the level of stress or sentiment in the songs. Fluctuations in economic prosperity did not appear to predict whether people listened to happier or sadder music.
An exploratory analysis looked at the relationship between stress and complexity. Periods of higher stress language generally corresponded with increased lyrical complexity. The researchers also noted a reversal in the trend of simplifying lyrics starting around 2016. Lyrical complexity increased during the first presidential term of Donald Trump compared to the second term of Barack Obama. This finding suggests that the long-term trend toward simplicity is not irreversible.
“During Trump’s presidency, lyrical complexity increased, marking a distinct deviation from earlier trends,” the researchers wrote.
This study adds to a growing body of work linking cultural products to psychological states. Research published in 2024 by Emilia Parada-Cabaleiro and colleagues analyzed over 350,000 songs and found similar patterns. That team observed that lyrics have become more repetitive and personal over the last five decades. They noted a decrease in vocabulary richness and an increase in anger. The current study supports these observations while adding a focus on specific historical events.
A 2021 study by Michael E. W. Varnum and colleagues also identified a steady simplification of song lyrics. They attributed this shift largely to the increasing number of new songs produced each year. They argued that when listeners face an abundance of choice, they tend to prefer simpler, more easily processed content. The new findings by Foramitti and his team confirm the long-term decline in complexity. However, the new study identifies a potential reversal of this trend in recent years that previous models did not predict.
Other researchers have looked at economic indicators to explain lyrical trends. A 2020 study by Lin Qiu and associates found that unemployment rates predicted anger in popular music lyrics. They analyzed songs in both the United States and Germany. Their work suggested that negative economic conditions led to angrier music.
The current study differentiates itself by focusing on income growth rather than unemployment. The lack of a link between income and lyrical sentiment suggests that different economic metrics may relate to cultural expression in different ways. It is possible that subjective perceptions of the economy matter more than objective income data.
The findings align with the concept of mood management theory. This theory posits that individuals select media to regulate their emotional states. People often seek stimuli that can improve their mood or reduce tension. The preference for less stressful music during the COVID-19 pandemic provides evidence for this behavior on a collective level. Instead of wallowing in the collective anxiety, American listeners appeared to use music as a form of escapism.
Recent research published in 2025 by Pablo Marín-Liébana suggests that this choice of music matters significantly. His meta-analysis found that song lyrics can actively shape the thoughts and behaviors of listeners. Listening to prosocial lyrics can increase empathy, while aggressive lyrics can increase hostility. If the population gravitates toward less stressful music during a crisis, this choice could serve a functional purpose in maintaining social cohesion and individual well-being.
The authors of the new research acknowledge limitations to their work. The study relied exclusively on the text of the lyrics. It did not account for musical elements like melody, harmony, or tempo, which also convey emotion. A sad lyric set to an upbeat tempo might be interpreted differently by a listener. Future research could combine text analysis with audio features for a more complete picture of musical sentiment.
The dataset was also limited to the United States. This restricts the ability to generalize the findings to other cultures or countries. Musical preferences can vary significantly across different societies. The researchers suggest that future studies should examine charts from other countries to see if these patterns hold globally.
In addition, the analyses are observational in nature. This means they cannot prove that societal events caused the changes in music consumption. The trends represent temporal associations rather than definitive causal effects. Additionally, by using the Billboard Hot 100, the data inherently reflects mainstream consumption and may underrepresent underground production or genres that are distributed outside traditional channels.
“Our post hoc analysis suggests that lyrical complexity increased during Trump’s first term compared to Obama’s second term,” the researchers wrote. “This result is interesting but should be taken cautiously, as this work was not designed to systematically test the effects of political cycles. While Trump’s election sparked several cultural reactions, future research should explore the potential drivers of this complexity reversal and whether similar patterns emerge in other media and political cycles. Furthermore, this temporal association should not be interpreted as evidence of a causal effect, particularly given the overlap with other events such as the onset of COVID-19.”
The study, “Societal crises disrupt long-term increases in stress, negativity, and simplicity in US Billboard song lyrics from 1973 to 2023,” was authored by Markus Foramitti, Urs Markus Nater, Claus Lamm, and Maurício Martins.
Leave a comment
You must be logged in to post a comment.