A Comparative Content Analysis of Anti- and Prosocial Rap Lyrical Themes Found on Traditional and New Media Outlets

Abstract

Previous studies investigated the content of rap music within the context of traditional media and found that rap often contains antisocial themes associated with negative effects. The current content analysis investigates whether rap’s lyrical themes consumed and shared online are more diverse and less anti-social than rap aired on traditional outlets. The analysis revealed that songs shared on Facebook were less antisocial and more prosocial than the songs that made the Billboard charts. Additionally, online lyrical themes were more diverse than traditionally distributed lyrics. Rap consumption and sharing behaviors will be discussed in light of the theory of selective exposure

Funding

The research was supported in part by funding from the Ronald E. McNair Research Scholars Program.

Notes

^1. We have reported three reliability indicators for our study. These include Krippendorf’s alpha, Cohen’s kappa, and percent of agreement. Our goal here was to use multiple indicators of reliability given that we were assessing categories that may occur infrequently within the population of songs being assessed. Although Krippendorf’s alpha has become a standard reliability coefficient in media scholarship, several scholars have noted that it tends to punish investigators for infrequently varied judgments among coders, compared to other measures (Artstein & Poesio, 2008; Zhao, Jun S. Liu, & Deng, 2012). These infrequently varied judgments might be more an indication of the population distribution under consideration than a problem with actual coder reliability. In order to determine whether lower alphas were due to issues with reliability rather than infrequently occurring categories, we also examined percent agreement and kappa. When kappa and percent of agreement were also low, we took this as a problem with attaining reliability for that variable and it was not included in our analysis. In cases where kappa and percent agreement remained above .70, we retained the variables for analysis even if alpha dipped below .70.

Additional information

Funding

The research was supported in part by funding from the Ronald E. McNair Research Scholars Program.

Notes on contributors

Avriel C. Epps

Avriel C. Epps (B.A., University of California, Los Angeles) conducted this research as part of her honor’s thesis for the Department of Communication Studies at UCLA.

Travis L. Dixon

Travis L. Dixon (Ph.D., University of California, Santa Barbara, 1998) is the communication alumni professorial scholar and associate professor of Communication at the University of Illinois at Urbana-Champaign. His research investigates the prevalence of stereotypes in the mass media and the impact of stereotypical imagery on audience members.