Conclusion
​
Ultimately, we did not find any significant, consistent differences between the lyrical sentiment of male and female artists, therefore we have decided to accept our null hypothesis.
Research into linguistic and speech patterns has found differences found between men and women in everyday speech, which partly informed our null hypothesis: male and female songs do not differ in terms of their lyrics’ sentiment. As per our aims and objectives, we quantified some of the differences in lyrical sentiments expressed by men and women, but such differences were too nuanced to warrant a rejection of the null hypothesis. Using an English lexicon led to the misclassification of songs written partly or fully in another language, and words could be misclassified if used in a different context to the definition within the lexicon. These two issues illustrate some limitations of the analysis. Using the NRC sentiment analysis, we found that patterns generally exist irrespective of gender, and there is little difference between the distribution of sentiment scores of male and female songs. It is important to remember the effect of men dominating the music billboards. There is less data available on female artists, lessening the precision of conclusions drawn about female artists.
Further Research
Despite the limitations, our data set and findings could provide a good springboard for further research.
Since we found that gender does not heavily influence the sentiment score of lyrics, but we still observed a range of sentiment scores in our data set, the sentiment of popular songs must be influenced by other factors, perhaps a combination of genre, topic, language, or something else entirely. The main difference we found between male and female songs was in their genres, so perhaps this is a good starting point to take our research further. Further research could include whether these differences in genre frequency between male and female artists is consistent in less mainstream music, whether any stereotypes about male/female musicians emerge from or cause these genre differences, and whether genre plays a significant role in impacting lyrical sentiment regardless of gender.
​
As mentioned, in our initial research summarised in our literature review, we found evidence that men and women do use different language, but our investigation found few differences in lyrical sentimentality. It could be the case that although men and women tend to have a different vernacular in everyday speech, the underlying sentiment and intention of the language is the same. It could also be the case that both genders conform to a particular type of language use when writing songs marketed as mainstream music, keeping in mind that our dataset is of Billboard’s most popular songs which have had to appeal to people of all backgrounds, genders, etc. Further research into the differences between lyrical sentiment and the sentiment of daily speech for each gender can help clarify this. Our data also focuses on popular music, but it would also be interesting to see if in less mainstream music any gendered sentiment differences emerge, and see how representative the Billboard charts are of the music industry as a whole.